1.1.1. Taxonomy

Schistosomes are trematode worms (‘flukes’) belonging to the phylum Platyhelminthes. The adult worms live in the vascular system of birds and mammals (‘blood flukes’). Other pathologically important Platyhelminthes include the digenetic trematodes Opisthorchis, Clonorchis, Paragonimus, Fasciolopsis and Fasciola and the cestodes (tapeworms).

All the schistosomes that mature in man belong to the genus Schistosoma of the family Schistosomatidae, which contains 11 other genera, some of which cause cercarial dermatitis (Rollinson & Southgate, 1987). The genus Schistosoma contains 19 species (WHO, 1993), five of which (Schistosoma haematobium, S. mansoni, S. japonicum, S. mekongi and S. intercalatum) are of major pathological importance, while the others are essentially parasites of non-human mammals, although some zoonotic transmission to man does occur. An estimated 600 million people are at risk for schistosomiasis; 200 million are currently infected in 74 countries (WHO, 1993). Probably more than 95% of human infections are due to S. mansoni and S. haematobium. Several of the ‘non-human’ species, including S. mattheei and S. bovis, are of veterinary importance, and both domestic and feral animals are major reservoirs of infection with S. japonicum (but not with any of the other species) (Taylor, 1987).

This monograph is restricted to S. haematobium, S. mansoni and S. japonicum.

1.1.2. Structure

Unlike all other pathologically important trematodes, schistosomes are dioecious (rather than hermaphroditic). The adult worms are about 1 cm long, and the male has a deep ventral groove or schist (hence the term ‘schistosome’) in which the female worm resides permanently in copulo. Worms of each sex have a mouth at the anterior end, which also serves as the anus since there is only one gut opening. Around the mouth is the oral sucker, while nearby, further back, is the ventral sucker. These suckers are much better developed in male worms; they are used mainly for hanging on to the venous epithelium of the host and for locomotion of the worm pair. In order to obtain amino acids for protein synthesis, the adult worms ingest red blood cells and break down the haemoglobin with a haemoglobinase. Small molecules, including glucose, amino acids, purines and pyrimidines, are taken up via transtegumentary absorption; there is evidence that the female derives much of her nutrition via transtegumentary absorption from the male worm. The metabolism of adult schistosomes is largely anaerobic, by glycolysis (Rumjanek, 1987).

1.1.3. Life cycle and biology of the adult worm

Schistosoma do not multiply in the human body. The life cycle of schistosomes is illustrated in Figure 1.

Figure 1.

Life cycle of blood flukes

Adult worms are found either in the vesical plexus of the urinary bladder (S. haematobium) or in the mesenteric veins (other species). Adult worms live for up to 30 years (von Lichtenberg, 1987), with a mean lifespan of 3–6 years (Anderson, 1987). They produce large numbers of eggs: 300 per day per female S. mansoni and S. haematobium and 10 times as many per female S. japonicum. About one-half of the eggs transit to the lumen of the urinary bladder (S. haematobium) or the intestine (other species), from where they leave the body in the urine or faeces, respectively. A substantial number of eggs are retained in the tissues, where they survive for a further three weeks; these are responsible for inducing most of the pathological manifestations of disease (Warren, 1978).

The eggs are large (e.g. 144 × 58 μm for S. haematobium) and consist of an egg shell of tanned protein containing, when laid, about 40 yolk cells and the oocyte. After about one week in the tissues, the mature egg contains the large (150 × 70 μm) ciliated miracidium larva (von Lichtenberg, 1987). It is this life-cycle stage that infects the snail host. Thus, embryonated eggs excreted from the body in urine or faeces and deposited in water hatch to liberate the free-swimming miracidium larvae. If the miracidia can locate an appropriate snail host within a few hours, they penetrate it; if not, they die, as they do not feed.

Within the tissues of the snail, the miracidium is transformed into the mother sporocyst, within which are formed several hundred daughter sporocysts. These migrate from the site of penetration to the digestive gland and reproductive tract of the snail, in which they proliferate internally to produce cercariae, the stage that infects man. This process takes about one month, and from one miracidium several million genetically identical cercariae may be produced by this asexual process during the lifetime of the infected snail.

The cercariae are shed from the snail in response to temperature and light and aggregate at the surface of the water, ready to infect the definitive human host. They swim tail first, locate the host by a combination of chance and Chemotaxis and adhere to the skin by their suckers. A cercaria is approximately 0.5 mm long and consists of a head-end, bearing the oral and ventral suckers, and a tail with a pronounced fork. Cercariae respire aerobically, using glycogen as a substrate, but do not feed; therefore, if they do not penetrate the final host within a few hours they die. Cercariae penetrate intact skin rapidly, using proteolytic enzymes produced by the paired penetration glands at their anterior ends; the tail is discarded in the water. Once they are within the skin, a profound metamorphosis takes place, and the cercaria is transformed into the ‘skin-stage schistosomulum’. Metamorphosis includes shedding of the cercarial glycocalyx, transformation from the single-lipid bilayer tegument of the cercaria into the double-lipid bilayer of the schistosomulum and various physiological changes, such as a change from aerobic to anaerobic respiration and the acquisition of host molecules, particularly lipids, some of which are incorporated into the tegument (Wilson, 1987).

The schistosomulum then penetrates the basement membrane of the epidermis, using proteinases secreted by the residual penetration glands of the cercaria stage. In mice, this process takes about three days, after which time the schistosomulum enters a lymphatic vessel or capillary in the dermis and is carried passively to the lungs via the right side of the heart (Wilson, 1987). The young schistosomula embolize in the capillaries—being too large to pass through the pulmonary veins—whereupon they again metamorphose, this time to the ‘lung-stage schistosomulum’, which, unlike the skin stage, is capable of stretching out its body to become long and thin and can cross the capillary bed of the lungs, taking three to six days to reach the left side of the heart (Wilson, 1987). Schistosomula are then distributed all over the body via the left ventricle, in proportion to cardiac output. Those that embolize in various capillary beds migrate through these to regain the heart and recirculate until they reach the hepatic portal system, a process usually completed within three recirculations. When the hepatic portal system is reached, a third metamorphosis takes place: the elongated migratory forms return to the squat shape of the skin-stage schistosomulum. Blood feeding begins and this is followed by growth, organogenesis and sexual maturation. The mature worms pair up in the intrahepatic portal venules from about four weeks onwards and then migrate to the final sites of oviposition in the vesical plexus (S. haematobium) or in the mesenteric veins (all other species).

1.2.1. History taking

Information derived from simple questionnaires eliciting a history of haematuria is sufficiently accurate to identify nearly all heavily infected people (Mott et al., 1985), and such questionnaires can be used for rapid identification of communities with a high prevalence of S. haematobium infection (Lengeler et al., 1991a,b). Validation of the use of questionnaires on history of S. mansoni infection for identifying infected people showed a specificity of about 60% (Barreto, 1993). In community-based epidemiological studies of S. japonicum, although symptoms of weakness, colicky abdominal pain and diarrhoea were observed in a greater proportion of infected than uninfected individuals, these were not specific to infection (Olveda et al., 1983).

1.2.2. Clinical diagnosis

Macroscopic and microscopic haematuria are highly sensitive, specific signs of S. haematobium infection in most endemic areas of Africa and the eastern Mediterranean (Savioli & Mott, 1989; Savioli et al., 1990; Eltoum et al., 1992; Lengeler et al., 1993). Testing with chemical reagent strips to detect microscopic haematuria consistently results in the identification of 80% or more of people excreting S. haematobium eggs, and gross haematuria is associated with urinary egg counts greater than 50 per 10 ml of urine (Savioli et al., 1990).

Schistosomiasis is a protean, multisystem disease, and the clinical signs and symptoms are often nonspecific (Abdel-Wahab & Mahmoud, 1987; von Lichtenberg, 1987; Olveda & Domingo, 1987; Prata, 1987; Wilkins & Gilles, 1987; Chen & Mott, 1989). Thus, multiple abdominal symptoms may be found in patients infected with S. mansoni and S. japonicum, of which only a history of bloody diarrhoea is significantly associated with heavy infection (Sleigh & Mott, 1986; Olveda & Domingo, 1987). Schistosome eggs and associated granulomas and fibrosis are frequently detected by liver biopsy. The degree of periportal fibrosis can now be assessed accurately by ultrasonography of the liver (S. mansoni, S. japonicum) or urinary tract (S. haematobium), the latter having replaced intravenous pyelography which was formerly the standard method of assessment (Hatz et al., 1992a,b,c,d; Jenkins & Hatz, 1992; Wei-min et al., 1992).

1.2.3. Parasitological tests

The best method for diagnosing infection with mature, egg-producing adult worms is to demonstrate the presence of eggs in the urine (S. haematobium) or faeces (other species). In routine medical practice, diagnosis is usually qualitative rather than quantitative. In both techniques, some form of concentration is used to increase sensitivity. Thus, urine samples may be centrifuged or filtered to concentrate the eggs, while eggs in faecal samples are frequently concentrated by the formol–ether technique.

For most epidemiological purposes, however, eggs are counted, although the sensitivity is limited owing to small sample size (de Vlas & Gryseels, 1992; de Vlas et al., 1993). The quantitative relationship between the presence of viable adult worms and faecal or urinary egg counts has been established experimentally (Cheever, 1969) and in autopsy studies (Edington et al., 1970; Smith & Christie, 1986).

For S. haematobium infections, filtration through standard filter paper, cellulose membranes, polycarbonate or nylon in filter holders attached to a syringe is a standard quantitative technique. The Kato technique for examination of faeces for the eggs of other Schistosoma involves use of a glycerine-impregnated cellophane coverslip over a measured volume of stool, ranging from 10 to 50 mg.

Light and heavy infections can be distinguished reliably by the available faecal and urinary examination techniques in all endemic areas. The limitations of their sensitivity have been well described (Savioli et al., 1990; de Vlas & Gryseels, 1992; de Vlas et al., 1993). A single filtration of a random 10-ml urine sample allows detection of all infected individuals with more than 50 eggs/10 ml of urine (Savioli et al., 1990). Although several quantitative techniques are available for faecal egg counting, their sensitivity is dependent on the amount of stool examined, and the Kato technique has become the standard, allowing comparison of the results of epidemiological studies. A single Kato thick smear allows detection of all people with more than 100 eggs/g of faeces (Barreto et al., 1990; Feldmeier & Poggensee, 1993).

In people with chronic or light infections, eggs may be difficult to demonstrate with these techniques. In such cases, rectal biopsy is sometimes used, followed by microscopic examination of compressed mucosal specimens for eggs. The sensitivity of rectal biopsy is unknown; however, it appears to be highly sensitive clinically, even if the viability of the infection cannot be determined. Sometimes eggs (or adult worms) are found by histopathological examination of lesions taken by biopsy from other anatomical sites or in cytological smears. S. haematobium eggs are frequently reported in diverse parts of the urogenital system, and ‘ectopic’ lesions of the central nervous system caused by S. japonicum or S. mansoni are seen (Chen & Mott, 1989).

1.2.4. Immunological tests

In the past, immediate hypersensitivy-based intradermal tests for S. mansoni and S. japonicum were widely used in epidemiological studies, but they have been rarely used since 1970 because of the lack of correlation with active infection and the availability of improved parasitological techniques. Using S. mansoni adult worm antigens, the sensitivity ranged from 82 to 100% and the specificity from 96 to 99% (Mott & Dixon, 1982); with S. japonicum adult antigens, the sensitivity ranged from 77 to 99% and the specificity from 95 to 99% (Mott et al., 1987). The age distribution of intradermal reactivity is not known. The specificity is not influenced by other intercurrent infections, except for certain trematode infections; the sensitivity is lower in children than in adults, and the sensitivity of the test and the intensity of the hypersensitivity reaction are greater in infections of long duration. Reactivity persists for years after a successful treatment (Kagan & Pellegrino, 1961).

Researchers have concentrated on S. mansoni and S. japonicum infections because of the ease with which the parasites can be maintained in the laboratory. Many immunodiagnostic techniques have been described and used experimentally, but so far none has been used consistently or validated in epidemiological studies (Mott & Dixon, 1982; Mott et al., 1987). Difficulty in maintaining S. haematobium in the laboratory has limited research in immunodiagnosis of urinary schistosomes (Xue et al., 1993).

Antibody detection assays are very sensitive; however, in epidemiological studies, a positive serological result may be due to either a light infection or the presence of residual antibody from a resolved infection. This is a particular disadvantage now that large-scale chemotherapy campaigns are so frequently carried out (Bergquist, 1992). Antigen detection assays may solve these problems. Several systems are being developed, the most advanced of which involve an enzyme-linked immunosorbent assay with monoclonal antibodies to detect circulating antigens of S. mansoni (de Jonge, 1992).

1.2.5. Establishment of absence of infection

The absence of infection cannot be established unequivocally. The variation in sensitivity of the diagnostic techniques is such that a combination of diagnostic tests is appropriate to establish absence of infection (Feldmeier & Poggensee, 1993). In the field, at least three successive urine filtration examinations are required to establish the absence of infection with S. haematobium (Savioli et al., 1990). For S. mansoni infection, five consecutive Kato examinations are required (Barreto et al., 1978). If available, antigen detection assays can be used (see section 1.2.4).

1.3.1. Geographical distribution (see Table 1 and Figures 2 and 3)

Table 1.

Geographical distribution of schistosomiasis by species.

Figure 2.

Global distribution of schistosomiasis due to Schistosoma haematobium, S. japonicum and S. mekongi

Figure 3.

Global distribution of schistosomiasis due to Schistosoma mansoni and S. intercalatum

It has been estimated that over 600 million people in 74 countries are exposed to the risk of schistosomal infection, and 200 million are currently infected (WHO, 1993). Schistosomiasis may be the second most important parasitic disease in man after malaria. About 95% of cases are due to S. mansoni and S. haematobium infections and the remainder to S. japonicum, S. intercalatum and S. mekongi. The geographical distribution of the schistosomes roughly corresponds to the distribution of susceptible snail hosts, which are present in many tropical and subtropical regions. S. mansoni is the most widespread species, being prevalent in 52 countries in Africa, the eastern Mediterranean, South America and the Caribbean. S. haematobium and S. mansoni have a similar distribution in Africa and the eastern Mediterranean; S. haematobium does not occur in the Americas. There is a small focus of S. haematobium in India, but neither S. mansoni nor S. haematobium occurs in central or east Asia; S. japonicum is endemic in three countries (China, the Philippines and Indonesia), while the related S. mekongi is restricted to the Mekong River basin in the Lao People's Democratic Republic and Cambodia. In Africa, S. mansoni and S. haematobium often coexist, and mixed infections are common. S. intercalatum, a much rarer species than S. mansoni or S. haematobium, is restricted to foci in 10 central and West African countries.

1.3.2. Risk factors for infection

Contact with contaminated freshwater is the major risk factor of infection (Jordan & Webbe, 1993). Many other environmental factors influence the distribution, prevalence, intensity of infection, morbidity and mortality of schistosomiasis (WHO, 1993). Among these are the type and size of intermediate snail host populations, human population density and behaviour in relation to freshwater bodies, and climatic and hydrological features. Infection may be constant in endemic areas owing to repeated contact with water, particularly among children.

Susceptibility to infection is influenced by genetic factors (Abel et al., 1991), but genetic differences between parasites are not known to influence their infectivity. Acquisition of infection depends on the duration of exposure, proportion of the body surface exposed, degree of body movement during exposure, presence of intermediate snail hosts, cercarial concentration in the water and water temperature. These conditions are fulfilled in endemic areas, usually in open water bodies where frequent recreational contact occurs.

Since schistosomes, like most other helminths, do not multiply in man, it is a striking feature of schistosome epidemiology that, although the prevalence of infection may be very high, significant symptoms are present in only the small segment of people who are most heavily infected. The decline in prevalence and intensity of infection after the second decade of life is believed to be due mainly to the gradual acquisition of immunity, although other age-related factors, such as decreasing contact with infected water and physiological changes associated with the onset of puberty, may also be important (Hagan et al., 1991; Rihet et al., 1991; Dessein et al., 1992; Dunne et al., 1992) (see Figures 4 and 5).

Figure 4.

Relationship between overall prevalence and intensity of infection with Schistosoma mansoni as determined by the Kato technique in different endemic areas in various studies

Figure 5.

Age–prevalence patterns based on faecal and urinary egg counts

1.3.3. Aggregation of infection

Within any endemic area, transmission is highly focal and the prevalence and intensity of infection vary between households, communities and progressively more population agglomerations. This heterogeneity or aggregation is determined by the risk factors cited in section 1.3.2. In common with other worms, Schistosoma are not randomly distributed among infected persons but are aggregated in heavily infected people in a manner best described by a negative binomial distribution. The amount of tissue damage caused by the Schistosoma infection is roughly proportional to the numbers of worms present; it is the heavily infected segment of the population that is at greatest risk of developing disease and which contributes the most to transmission of the parasite.

1.3.4. Prevalence and intensity of infection

For epidemiological studies, the intensity of infection is measured by the number of eggs/10 ml of a urine sample (S. haematobium) or per gram of faeces (all other species). Definitions of ‘heavy’ infections are routinely included in most epidemiological studies (Sleigh & Mott, 1986). Throughout areas endemic for urinary schistosomiasis, most infection in people who excrete more than 50 eggs/10 ml of urine is associated with haematuria (Mott et al., 1983). The definition of heavy infection due to S. mansoni varies from a mean of 16 eggs/g of faeces in areas of low prevalence such as Puerto Rico (Hiatt et al., 1980) to 1000 eggs/g of faeces in Burundi (de Vlas & Gryseels, 1992). About 10% of infected people in areas endemic for S. mansoni have heavy infections. S. japonicum infections have been considered to be heavy when more than 400 eggs/g of faeces are found; they occur in up to 4% of some populations (Olveda & Domingo, 1987).

Analysis of 11 methodologically similar studies (Jordan & Webbe, 1993) showed that there is a general trend to a proportional relationship, i.e. the higher the prevalence, the higher the intensity (Figure 4). A similar relationship was seen for S. haematobium infection, but few similar population-based studies have been reported using comparable methods.

The peak prevalence of all Schistosoma infections occurs in the second decade of life. In general, the decrease in intensity of S. haematobium infection after that time is accompanied by a comparatively greater decrease in prevalence than in S. mansoni infection. That is, while the intensity of S. mansoni infection tends to decrease in the same period, the prevalence remains high, i.e. a few eggs are excreted over a long period.

Few studies have been carried out on the interaction of S. mansoni and S. haematobium infections. The data reported by Robert et al. (1989) suggested that the intensity of S. haematobium in mixed infections was greater than that in infections with S. haematobium alone.

1.3.5. Sex-related patterns of infection

Differences in the sex distribution of infection were seen in three selected epidemiological studies (Figures 4 and 5). In general, although not universally, the prevalence and intensity of infection are higher in men than in women, owing to greater employment in agricultural work. The interpretation of any statement about sex differences must, however, take into account the focality of infection and its variable distribution (see section 1.3.3). In predominantly Islamic countries such as Egypt, the prevalence and intensity of urinary schistosomiasis tend to be lower in girls and women than in boys and men (El-Malatawy et al., 1992) owing to lower rates of contact with water.

1.3.6. Relationship of morbidity to intensity of infection

Morbidity due to Schistosoma infection becomes apparent during the period of peak prevalence and intensity of infection as well as many years later. In urinary schistosomiasis due to S. haematobium, the intensity of infection is correlated with morbidity, especially in children. The degree of haematuria and proteinuria detectable by chemical reagent strips is correlated with the intensity of infection (Mott et al., 1983; Savioli et al., 1990). Changes in the urinary bladder and ureters detected radiologically (Forsyth, 1969; Pugh et al., 1979; Warren et al., 1979) or by ultrasound (Hatz et al., 1992a), cystoscopic abnormalities of the urinary bladder (Abdel Salam & Ehsan, 1978) and pathological signs at autopsy (see section 4.1) are also correlated with intensity of infection.

Although S. japonicum adults lay more eggs per day (see section 1.1.3), the rates of hepatic and splenic enlargement are similar to those observed in S. mansoni infections when the egg counts are similar (Olveda & Domingo, 1987).

Kloetzel (1964) showed in population-based studies in Northeast Brazil that the rates of splenomegaly associated with S. mansoni infection are proportional to the intensity of infection, as measured by faecal egg counts, particularly in the first two decades of life.

1.3.7. Relationship of morbidity to mortality from infection

Annual mortality due to S. haematobium infection in East Africa has been estimated at 2/1000 infected adults (Forsyth, 1969). The proportional contribution of urinary bladder cancer or hydronephrosis leading to end-stage renal disease is not known.

In 1984, annual mortality due to portal hypertension caused by schistosomiasis from S. mansoni in Brazil was estimated at 0.5/100 000 total population; at the same time in Suriname, the figure was estimated to be 2.4/100 000 inhabitants. The control of schistosomiasis through large-scale chemotherapy in Brazil was associated with a decline in related annual mortality between 1977 and 1988 from 0.67 to 0.44 deaths per 100 000 inhabitants (WHO, 1993).

Before the introduction of praziquantel in China, severe acute schistosomiasis due to S. japonicum resulted in a mortality rate of 2.5–20.7%. Mortality from schistosomiasis during 1975–79 in 10 counties in the Jiaxing area, Zhejiang Province in China was reported to vary from 0.69 to 39.8/100 000 in men [median, 15.1] and from 0.45 to 44.6/100 000 in women [median, 7.7] (Liu et al., 1983). Cumulative (0–64) mortality rates during 1973–75 were reported from 49 counties in various Chinese provinces. No mortality was seen among men in 29 counties or among women in 36 counties; the rates in counties with some deaths from schistosomiasis varied from 0.03 to 37.2/1000 for men [median, 1.3/1000] and 0.07–42.1/1000 for women [median, 1.4/1000] (Chen et al., 1990). In Leyte, the Philippines, annual mortality among 135 untreated patients was 1.8% (Blas et al., 1986). More widespread use of antischistosomal drugs in highly endemic areas should reduce both morbidity and mortality.

1.5.1. Treatment

Safe, effective chemotherapy has been available for the past 20 years against all the schistosomes that affect man (WHO, 1993). The most versatile drug is praziquantel which is effective in a single oral dose against all species of schistosomes (and some other trematodes and cestodes). Large-scale treatment is costly (US$ 0.35 per treatment), and, in areas of infection with S. haematobium only, the much cheaper metrifonate may be preferred, which, however, must be given in two or three doses at two-week intervals. Metrifonate is effective only against S. haematobium, while the third available drug, oxamniquine, is effective only against S. mansoni, for which it provides safe and effective treatment. None of these drugs is significantly effective against infections by immature worms; thus, prophylactic treatment is not available. Katayama syndrome is usually treated symptomatically for hypersensitivity reactions, but praziquantel is also given to kill adult worms as they mature. In advanced or ectopic disease, surgery for anatomical consequences and complications of infection may be necessary, but, even in advanced cases, antischistosomal drug therapy usually produces great improvement.

Treatment of all forms of schistosomiasis with safe, effective antischistosomal drugs (i) results in a high rate of resolution of infection, even in endemic areas where reinfection is a risk; (ii) prevents development of disease in people with heavy infection; (iii) arrests progression of existing severe disease; and (iv) reverses some manifestations of disease, such as haematuria and proteinuria, particularly in children. Liver fibrosis caused by S. mansoni and S. japonicum infection is usually arrested by the treatment and may even be reversed (Mohamed-Ali et al., 1991; Wei-min et al., 1992). Similarly, in cases of S. haematobium infection, hydroureter and hydronephrosis are reversible by treatment (Doehring et al., 1986; Hatz et al., 1990; King et al., 1990).

1.5.2. Control

Control of schistosomiasis in the community may in practice be achievable by removing the adult worms by chemotherapy, by eliminating the snail intermediate hosts by modification of their habitat or by chemical attack, by changing human behaviour through health education, by providing safe water supplies and sanitation, so that excreta containing live eggs do not reach water containing snails, and by ensuring that people avoid water contaminated with cercariae.

Effective drugs are available. Trivalent antimonials were introduced in 1918, although these toxic compounds were far from ideal for control programmes since they required repeated intravenous injections. Chemical control of snails by molluscicides became possible in the 1920s, when copper sulfate was introduced for the control of the aquatic vectors of S. mansoni and S. haematobium, and when lime was first used to attack the amphibious vectors of S. japonicum.

Using integrated control measures since the 1920s, the Japanese eventually eradicated schistosomiasis by the end of the 1970s (Kitani & Iuchi, 1990). Similarly, in the much more extensive endemic areas of S. japonicum in China, unremitting integrated control measures over a 40-year period have reduced the prevalence of schistosomiasis by 90% (Chen, 1989; Anon., 1992). Eradication has also been achieved in two other countries: S. haematobium has been eliminated in Tunisia, and S. mansoni in Monserrat (WHO, 1993). In several countries, particularly those where schisosomiasis was identified early on as a major public health problem, such as Brazil, Egypt, the Islamic Republic of Iran, the Philippines and Venezuela, significant reductions in disease prevalence have been achieved, usually by national control programmes that incorporate integrated measures. Even in cases where prevalence of infection has remained high, the prevalence of serious disease manifestations (such as Symmers' fibrosis and fibro-obstructive lesions of the urogenital tract) has often been reduced, largely by the use of population-based chemotherapeutic campaigns (WHO, 1993).

Set against this, however, is the demographic increase in younger people, who are most affected by the disease, thus increasing the size of the susceptible population. This, combined with the expansion of water resource developments and irrigation, has led to spread of the disease to new areas and to intensification of transmission in existing endemic areas. The WHO (1993) report on schistosomiasis control thus concluded that the global number of infected cases was similar to that in 1984. Furthermore, in only a very few areas has the snail vector been eradicated, so that, if control measures break down or are relaxed, the disease will rapidly sweep back and may in fact become worse than before because of loss of immunity by the population. Currently, no antischistosomal vaccine for humans is available, although intensive efforts are being made to develop one.

2.1.1. Schistosoma haematobium

The proportion of all cancers represented by urinary bladder cancer varies greatly within Africa and the Middle East, and the ratio of male to female frequency of occurrence is nearly as variable (Parkin, 1986). In Egypt, the proportion of bladder cancers among all cancers in men is twice that in Zambia, four times that in Zimbabwe and 10 times that in Algeria. Very few formal assessments of the correlation between bladder cancer incidence and the prevalence of S. haematobium have been done, but there are many informal descriptions of geographical correspondence between the areas affected by the two diseases.

Most of the early clinical descriptions of urinary bladder cancer in connection with evidence of schistosomiasis come from the Nile Delta, where there are few unexposed populations and no population-based incidence data (see section 2.2.1); however, in countries with less universal exposure, observations have been made on the geographical relationship between exposure to S. haematobium and bladder cancer occurrence. The common geographical pattern of occurrence of S. haematobium and bladder cancer has been noted by investigators in almost all endemic African countries (Table 2).

Table 2.

Descriptive studies of infection with Schistosoma haematobium and urinary bladder cancer.

In addition to the link between the risk of a subpopulation for a haematobium schistosomiasis and the risk of the same population for urinary bladder cancer, a slightly more direct link has been noted; the proportion of bladder cancers that are squamous histologically in the population of a country is related to the proportion of cancerous bladder specimens from that population which contain evidence of past schistosomal infection in the form of eggs or egg remnants (Lucas, 1982a). This has been noted even within countries; in Iraq, for example, 36.1% of bladder cancer cases from the north are squamous-cell tumours and 4.9% have evidence of S. haematobium, whereas in the south, where S. haematobium is more prevalent, 54.8% are of the squamous variety and 32.2% have evidence of S. haematobium; those from the central part of the country show intermediary rates of 48.5% and 20.7%, respectively (Al-Fouadi & Parkin, 1984).

The two diseases have other characteristics in common. In a description of the pattern of urinary bladder cancer by occupation in the Nile Delta, 99% of the bladder cancers occurring in high-risk male agricultural workers (fellahin) were found to be associated with histological evidence of S. haematobium infection, whereas only 52% of the cases occurring in men with lower-risk occupations showed such evidence (Makhyoun et al., 1971).

Whereas in the Nile Delta, where men do most of the agricultural work, the ratio of male to female cases of urinary bladder cancer with histological evidence of past infection may be as high as 12:1 (Makhyoun et al., 1971), the sex ratio among those without such evidence approximates the 4:1 ratio seen in the United Kingdom (Prates & Gillman, 1959). In contrast, in Mozambique (Prates, 1963) and adjacent regions of the Transvaal in South Africa (Keen & Fripp, 1980), where women do most of the agricultural labour and are therefore more commonly infected, the sex ratios are reversed to 1:1.1 or even 1:2, even though ratios of 2:1 prevail among cases referred from nearby areas. The sex ratio of bladder cancer cases has also been linked to the histologically measured intensity of infection in tumour specimens, and ranged from 8.7:1 in heavily infected people, to 4:1 in those who are lightly infected, to 2:1 in those without eggs in Egypt (Tawfik, 1988).

In a community in Angola, where both males and females work in agriculture, the minimal age of infection with S. haematobium was 11 years. The mean age of patients with urinary bladder carcinomas associated with schistosomiasis was 44 years. The sex ratio was 1.6:1 for bladder carcinoma associated with schistosomiasis and 3.2:1 for bladder carcinoma not associated with schistosomal disease (p ~ 0.05) (da Silva Lopes, 1984).

It should be noted, however, that in Uganda, squamous-cell carcinomas of the urinary bladder are commoner than in Europe or North America in the absence of any relationship to known S. haematobium prevalence (Anthony, 1974).

Because of the lack of population-based cancer registration, the secular trends in incidence of squamous- or transitional-cell carcinomas of the urinary bladder have not been formally evaluated. In an area of the Nile Delta where the prevalence of S. haematobium infection was brought from a level of 60% in 1968 to 10% in 1988, no impact upon the rate of bladder cancer was clinically evident at the end of that period, although the mean age at diagnosis had increased (Tawfik, 1988).

2.1.2. Schistosoma mansoni

No description has appeared of the geographical occurrence of cancer in relation to the prevalence of S. mansoni infection. In relation to liver cancer, one observer pointed out that the pattern of occurrence in Africa and South America does not correspond to that which would be expected on the basis of a strong association with S. haematobium (Edington, 1979). The absence of any geographical relationship between colorectal cancer and colorectal schistosomiasis in Africa is even clearer. Despite wide variations in the geographical distribution of S. mansoni, colorectal cancer occurs in Africa with remarkable uniformity, insofar as the proportion of cases among all cancers provides pertinent information (Parkin, 1986). Moreover, reports from multiple centres in north, east, south and west Africa all indicate that evidence of schistosomal infection in colorectal tumour specimens is no commoner than would have been expected on the basis of the known prevalence of infection (Murray, 1967).

2.1.3. Schistosoma japonicum

The geographical co-occurrence of S. japonicum and cancer has been assessed formally (Table 3). Unfortunately, interpretation of the geographical patterns of occurrence of liver and colorectal cancers in Asia is difficult, because of known variations in the distribution of other causes of the same neoplasms, including hepatitis viral infection, dietary nutrients and carcinogenic dietary contaminants such as aflatoxins. In particular, a large correlation study from China assessed the association between mortality from schistosomiasis and from colorectal, liver, oesophageal and gastric cancers (Liu et al., 1983). Correlations were calculated at two geographical levels: in 24 provinces of varying endemicity and in 10–98 counties within six provinces of high endemicity. [The Working Group noted that, in addition to the problems common to the interpretation of all correlation studies (see Preamble, p. 22), interpretation of studies correlating mortality from cancer and from schistosomiasis are complicated by the low diagnostic specificity of the latter cause of death; however, such misclassification of cause of death would probably lead to an underestimated correlation coefficient.]

Table 3.

Descriptive studies of infection with Schistosoma japonicum and cancer.

2.2.1. Schistosoma haematobium

Subsequent to the early reports, large series of cases of urinary bladder cancer have been reported in association with evidence of S. haematobium infection (see Box).

The case descriptions have repeatedly emphasized the preponderance of squamous-cell urinary bladder tumours among cases with evidence of schistosomal infection, the somewhat different distribution over the surface of the bladder (notably the rarity of occurrence in the trigone) in comparison with bladder tumours in developed countries, and the prevalence of metaplastic changes in conjunction with evidence of infection (da Silva Lopes, 1984). Clinically, the most notable and consistent feature described in these series is the relative youth of the cases with evidence of a link to S. haematobium infection. While this observation is made in almost all of the reports, and is usually interpreted as constituting evidence of etiological heterogeneity, the finding does not constitute strong evidence because evidence of the infection is known to decrease in frequency with age.

Other than urinary bladder cancer, the malignancies most frequently reported in association with S. haematobium infection are those of the female genitalia. A few dozen cases of squamous cervical carcinoma have been reported from endemic areas (Badawy, 1962; Youssef et al., 1962; Berry, 1966; Sharma et al., 1970; Youssef et al., 1970; Bognel et al., 1980; Schwartz, 1984; El Tabbakh & Hamza, 1989), and the same authors and others (Shafeek, 1957; Iskander & Kamel, 1968; Sunder-Raj, 1976; Al-Adnani & Saleh, 1982; El-Maraghy et al., 1982) have reported certain other genital squamous malignancies, ovarian cystadenocarcinomas, Brenner tumours and teratomas. It has been alleged that breast cancers in men infected with S. haematobium constitute a relatively high proportion of all male breast cancers in Egypt (El-Gazayerli & Abdel-Aziz, 1963; Sherif et al., 1980), but the reported numbers are small and cannot be evaluated. Relatively small numbers of other malignancies that have been reported in association with evidence of S. haematobium infection include hepatocellular carcinoma (Nkrumah, 1964; Hashem, 1971), bladder sarcoma (Alwan et al., 1988) and lymphomas (Edington et al., 1970; Cheever et al., 1978).

2.2.2. Schistosoma mansoni

Cases of liver cancer have been reported in connection with evidence of S. mansoni infection from Egypt (Hashem, 1971), Mozambique (Prates & Torres, 1965), Brazil (Cheever & Andrade, 1967; Lyra et al., 1976), Puerto Rico (Martinez-Maldonado et al., 1965), Saudi Arabia (Nouh et al., 1990) and Nigeria (Edington et al., 1970). Similarly, cases of colorectal cancer have frequently been described from Egypt (Afifi, 1948; Dimmette et al., 1956; Cheever et al., 1978) and Lebanon (Uthman et al., 1991). Andrade and Abreu (1971) reported the occurrence of eight giant follicular lymphomas in 863 spleens removed from patients with portal hypertension due to infection with S. mansoni; subsequently, six additional cases of this neoplasm were described (Paes & Marigo, 1981) in a similar series of 714 spleens. Of these 14 lymphomas, four were further confirmed in biopsy samples or at autopsy; the rest were lost to follow-up. Although other individual cases of diverse lymphomas have been reported in patients with schistosomiasis (Andrade & Abreu, 1971; Cheever et al., 1978; de Andrade et al., 1982; Chirimwami et al., 1991), no reports of giant follicular-cell lymphoma have subsequently appeared.

Other malignancies that have been reported in association with evidence of S. mansoni infection include prostatic cancer (Alexis & Domingo, 1986; Godec et al., 1992), ovarian teratoma (Kahn et al., 1978), uterine leiomyosarcoma (Joyce et al., 1972), renal-cell carcinoma (Oro Ortiz et al., 1991), rectal carcinoid tumour (Satti et al., 1988) and cancer of the cervix (Coelho et al., 1979; Wright et al., 1982).

2.2.3. Schistosoma japonicum

Most of the cases or series of cases of liver cancers reported in association with S. japonicum infection have come from Japan (Iuchi et al., 1971; Nakashima et al., 1975; Kojiro et al., 1986; Fujimoto et al., 1989; Kitani & Iuchi, 1990; Uetsuji et al., 1990). Within such series, cases of liver cancer have been reported to occur commonly in patients who responded positively to a skin test or were shown histologically to have S. japonicum infection (Iuchi et al., 1971; Nakashima et al., 1975; Kojiro et al., 1986); in patients who had evidence of hepatitis viral infection (Nakashima et al., 1975; Kitani & Iuchi, 1990; Kojiro et al., 1986); and in those with schistosoma-associated cirrhosis (Iuchi et al., 1971; Kitani & Iuchi, 1990). In one small series from an endemic area, S. japonicum was not found to be especially common in cases of liver cancer (Kamo & Ebato, 1982).

Series of cases of gastric cancer associated with histological evidence of S. japonicum infection have been reported from both Japan (Amano, 1980) and China (Wang, 1979; Qian & Yi, 1980; Feng & Shi, 1981; Wang & Kuang, 1983; Zhou, 1986).

Series of cases of colorectal cancer found in association with infection with S. japonicum have been reported from Japan (Shindo, 1976; Inoguchi et al., 1978; Naito et al., 1979; Amano, 1980; Hashimoto et al., 1986; Sekiguchi et al., 1989), the Phillipines (Abanilla, 1986) and China (Chen & Chen, 1957; Tsou & Ying, 1958; Wu et al., 1960; Chuang et al., 1979; Chen et al., 1980; Chen et al., 1981; Zhao & Wong, 1981; Liu et al., 1983; Zhuang et al., 1985; Chen, 1986). As in studies of bladder cancer, schistosomally infected patients are of younger average age in most series (Abanilla, 1986; Chen, 1986). This observation is difficult to interpret in the light of differences in the prevalence of infection with age.

Other malignancies that have been reported as individual cases in relation to S. japonicum infection include squamous-cell carcinoma of the skin (Ohtake et al., 1991), malignant schwannoma (Schwartz, 1982), carcinoma of the parotid gland (Tangchai & Poshayalakshana, 1968), bronchogenic carcinoma (Ishihara et al., 1984) and breast cancer (Zhou, 1983).

2.4.1. Schistosoma haematobium

Mustacchi and Shimkin (1958) identified 48 male and 7 female hospitalized patients with urinary bladder cancer in the Egyptian Nile Delta city of Tanta among 1472 consecutive admissions to the hospital. All patients were evaluated in relation to the presence of S. haematobium eggs in a urine sample taken at admission and to any subsequent evidence of S. haematobium infection [the latter but not the former could have been obtained on the basis of knowledge of the presence of bladder cancer]. After multivariate adjustment for age, sex and urban or rural origin, odds ratios of 2.1 (p = 0.04) were seen for the finding of eggs at the time of admission and 2.2 (p < 0.01) for any subsequent evidence of schistosomal infection.

Prates and Gillman (1959) compared 100 urinary bladder cancer cases in Maputo, Mozambique, with 185 cases found at autopsy in people over 40 years of age with respect to the frequency of identification of S. haematobium eggs in relation to the histological type of bladder cancer. Eggs were found in 33 of the cases found at autopsy and in 61% of controls [odds ratio, 0.3; 95% confidence interval (CI), 0.2–0.5]. Eggs were found in 56% of the 59 squamous-cell cancer patients but in none of the transitional-cell cancer patients. [The methods used to examine the biopsy and autopsy specimens were dissimilar, and there was no reconciliation of the high rate in cadavers, despite the absence of eggs in the bladders of people with transitional-cell cancer. The causes of death of the controls were not described, and no adjustment was made for differences in specific age or place of origin.]

Hinder and Schmaman (1969) compared the prevalence of histologically identified eggs in punch biopsy specimens from 79 patients with urinary bladder carcinoma in Johannesburg, South Africa, with the prevalence in two or more full-thickness biopsy specimens from 101 people over the age 15 who came to autopsy. Eggs were identified in 34.2% of the cases but in only 9.0% of the autopsied patients [odds ratio, 5.3; 95% CI, 2.3–12]. The causes of death of the controls were not provided, and no adjustment was made for differences in specific age or place of origin. When cases were analysed by histological type, 19% of transitional-cell carcinomas and 68% of squamous-cell carcinomas contained eggs.

Gelfand et al. (1967), in Harare, Zimbabwe, compared 33 patients with urinary bladder cancer with other hospital patients who had been ‘submitted to similar investigation’ and were matched on age, sex and race. Comparisons were made on the basis of the results of pelvic X-rays (33 pairs) and rectal biopsies (31 pairs). Among the 16 pairs discordant for calcified eggs identified by X-ray, the case was positive in 15, giving an odds ratio of [15; 95% CI, 2.0–114]; among the 15 pairs discordant for the results of rectal biopsy, the case was positive in 13, giving an odds ratio of [6.5; 95% CI, 1.5–29]. The diagnoses of disease in the controls were not described, and no adjustment was made for differences in smoking habits or place of origin.

In a project for cytological screening of urinary bladder cancer conducted from 1976 to 1979 in a location in the Nile Delta highly endemic for S. haematobium, participants over 20 years of age were characterized by occupation, on the presumption that the 4769 agricultural labourers in this age group had a higher prevalence of infection than 1112 people with other occupations (El-Bokainy et al., 1982). All 10 cases of bladder cancer detected and confirmed histologically appeared among the agricultural workers [prevalence ratio, ∞]. Although the ages of the subjects were not analysed in detail, it was concluded that adults of this working class were at increased risk of bladder cancer.

In Zambia, Elem and Purohit (1983) compared the bladders of 50 patients who had died of urinary bladder cancer with bladders from age- and sex-matched cadavers (mostly trauma victims matched on age and sex to the decedent) by means of X-ray examination and digestion of tissues away from the eggs they contained. The bladders of the cases were [3.8] [95% CI, 1.4–10] times as likely to show schistosomiasis by X-ray and [14] [95% CI, 4.6–43] times as likely to contain S. haematobium eggs.

In the Bulawayo region of Zimbabwe, cancer registration procedures from 1963 to 1977 included collection of information about exposures, including past history of clinical schistosomiasis (‘bilharzia’ or ‘blood in the urine’) (Skinner et al., 1993). Some difference in the availability of information about past schistosomiasis is evident between cases of urinary bladder cancer (61%) and cases of cancer of other types (50%). The exposures of 305 patients with bladder cancer were compared with those of 3145 other cancer patients, with and without exclusion of people with cancers known to be linked to smoking. The occurrence of bladder cancer was associated with place of origin, a lower level of education and a more menial occupation. No effect of smoking was found for squamous-cell cancers and only a modest effect (1.3) for other cancers. For a history of schistosomiasis in men, the odds ratio (using all other cancer cases as controls, relative to no such history and adjusted for age, tobacco use, province of origin, education and occupation) was 3.9 (95% CI, 2.9–5.1) for all types of cancer and 3.4 for both squamous-cell cancer and other specified carcinomas. When the cases of smoking-related cancers were excluded from the control group, the odds ratio for squamous-cell cancers increased to 3.9 and that for other cancers dropped to 3.1.

These studies are summarized in Table 5.

Table 5.

Case–control studies of infection with Schistosoma haematobium and urinary bladder cancer.

Of interest is an additional case–control study of urinary bladder cancer, which was not performed to test the hypothesis of schistosomal etiology (Table 6). Makhyoun (1974) compared males admitted to hospital for urinary bladder cancer in Alexandria and Tanta, Egypt, with other admitted males, matched on age and smoking history, after stratification of cases and controls on the basis of history of clinical schistosomiasis. In 80% of people without such a history who had smoked heavily or moderately, the malignancies were strongly associated with cumulative smoking history. Only 23% of the schistosomiasis patients had smoked moderately or heavily, and the link between cancer and smoking in these subjects, while present, was weaker. [While the role of past clinical schistosomiasis in bladder cancer was not assessed within groups comparable for past smoking history, the low level of smoking among the patients with S. haematobium-associated cancer makes it extremely unlikely that the pattern of smoking could explain the strong links between infection and bladder cancer.]

Table 6.

Case–control study of urinary bladder cancer in relation to smoking and history of infection with Schistosoma haematobium among males admitted to hospital.

2.4.2. Schistosoma japonicum

In a comparison based on skin testing for antigens to S. japonicum, Iuchi et al. (1971) found that 85.2% of 52 cases of hepatocellular carcinoma and 68.2% of 217 other hospital in-patients over 40 years of age had antigens. No adjustment was made for evidence of past hepatitis viral infection.

Inaba et al. (1984) used skin testing and medical histories to compare 62 cases of liver cancer diagnosed in seven hospitals in an endemic area, Yamanashi Prefecture, Japan, in 1977–79 with age- and sex-matched hospital controls admitted for various diseases other than liver disease. While the univariate relative risk was 9.5 [95% CI, 2.2–41], restriction of the analysis to the 88 subjects seronegative for hepatitis B surface antigen and their controls gave a relative risk of [6.7; 1.5–30] for 39 alcohol users and of [4.7; 1.2–19] for 49 non-users.

Guo and Lu (1987) compared 166 patients who had died of liver cancer with 166 people who had died from other cancers and with 166 healthy people, both groups matched on age, sex and place of residence with respect to a history of S. japonicum infection. The matched odds ratio for schistosomal infection based on both series of controls was 2.2 (p < 0.01). Relative risks of [2.5; 95% CI, 1.4–4.4] and [2.3; 1.3–4.1] were found in relation to cancer decedents and healthy controls, respectively, after adjustment for smoking and family history of liver cancer but not for evidence of hepatitis viral infection. The relative risk estimates increased significantly with the interval since diagnosis of schistosomiasis, whether cancer or healthy controls were used. Dietary exposure to aflatoxins was considered not to be prevalent in this area.

Amano (1980) compared 362 patients with stomach cancer who were treated surgically in Yamanashi Prefecture, Japan, with 897 surgical cases with non-malignant disease of the stomach and duodenum, and found S. japonicum eggs [1.8] [95% CI, 1.3–2.6] times more frequently in the tissues of cases than in those of controls. No adjustment was made for potential confounders. In the same study, eggs were found [1.2] [0.62–2.5] times more often in the tissues of 103 colon cancer cases than in the 96 controls with benign disease of the colon. No adjustment was made for diet or other potential confounders.

In endemic Kunshan County in Jiangsu Province, China, Xu and Su (1984) gathered medical histories on schistosomiasis for all colorectal cancer patients and for patients with other cancers and from healthy neighbours, each matched on age, sex, occupation and work unit. While no significant association was found between colon cancer and past history, odds ratios of 8.3 and 4.5 were found for rectal cancer in comparisons with cancer controls and healthy controls, respectively. No adjustment was made for diet or other potential determinants of colorectal cancer.

In the same county, Guo et al. (1987) compared people who had died of colon cancer with those who had died of lung cancer and with healthy people, with respect to any medical history of early- or late-stage schistosomiasis. In relation to healthy controls, odds ratios of 2.4 [CI not given] were found for a history of early-stage infection and 5.5 [CI not given] for a history of late-stage schistosomal disease. After adjustment for smoking and a family history of colon cancer, but not for diet or exercise, significant associations were still found: 2.1 (95% CI, 1.1–3.8) and 4.2 (1.2–15) in relation to lung cancer controls for early- and late-stage disease, respectively, and 2.4 (1.1–5.0) and 5.7 (1.3–25) for the same exposures in relation to healthy controls. The risk was found to increase stepwise from 1.2 to 4.3 after < 10 years to > 30 cumulative years of infection [CI not given].

These studies are summarized in Table 7.

Table 7.

Case–control studies of infection with Schistosoma japonicum and cancer.

A number of studies have addressed the association between infections with S. mansoni and S. japonicum and cancer of the liver. The possible confounding of schistosomal infection with hepatitis viral infection (see IARC, 1994) in these studies has rarely been addressed empirically. A recent review of the coincidence of infection with hepatitis B virus and with S. mansoni and S. japonicum in population-based studies (Chen et al., 1993) showed no significant increase in the prevalence of hepatitis B surface antigenaemia in people with these schistosomal infections. The prevalence of joint infection is, however, higher in hospital patients than in members of the corresponding general population; in particular, patients hospitalized with hepatosplenic schistosomiasis are more likely to be seropositive for hepatitis B surface antigen than those with latent or intestinal schistosomiasis. These observations suggest that hepatitis B viral infection may confound the association between schistosomal infection and liver cancer in hospital-based studies of individuals. There are no similar data that would allow evaluation of the possibility of confounding between hepatitis C viral infection and schistosomal infection.

3.1.1. Mouse

Groups of 6–10 male C3H mice, two months old, received single subcutaneous injections of 1.5–10 mg of lyophilized S. haematobium eggs or worms in saline or tricaprylin. Another group received a single intraperitoneal injection of 4 mg of egg material, while a further group received two subcutaneous injections of 3 or 4 mg of egg material at an interval of three months. Animals were examined for the presence of tumours 13, 17 and 20 months after injection. The experiment was terminated at 20 months. No tumours were seen at the injection site. Of the 20 mice killed at termination, five had pulmonary tumours and three had hepatomas. The authors noted that the frequency of pulmonary and liver tumours was similar to that in historical controls (Shimkin et al., 1955). [The Working Group noted the limited reporting and that an infectious agent was not employed.]

As part of an experiment on S. haematobium in combination with 2-acetylaminofluorene (see below), a control group of 20 Swiss mice [sex and age unspecified] was repeatedly treated by subcutaneous injection with cercariae [schedule unspecified] and kept for 44 weeks. Urinary bladder epithelial hyperplasia beginning as early as three weeks was observed in the majority of the mice. Hyperplasia was not observed in 100 untreated control animals (Hashem & Boutros, 1961). [The Working Group noted the absence of ova or worms in the bladder and the short duration of the study.]

A group of 30 male BALB/c mice, three to four weeks of age, received subcutaneous implants of ligated urinary bladder cysts from donor mice. Bladder cysts were prepared by distending bladders with 0.15 ml mineral oil containing 1000 lyophilized S. haematobium ova before ligation. The mice were observed for 44 weeks. A control group of 29 mice received donor bladder cysts containing 0.15 ml mineral oil alone. Hyperplasia of the bladder cyst epithelium was seen in 4/30 mice and 5/29 controls. One transitional-cell tumour and one squamous-cell tumour in the implanted bladder cyst [tumour pathology not described] were reported in the experimental group but not in controls (Al-Hussaini & McDonald, 1967). [The Working Group noted the unusual design of the experiment and the lack of significance of the tumour response.]

3.1.2. Rat

In a combination experiment, a group of 100 white rats [sex, strain and age unspecified], weighing 140–160 g, were exposed to water containing 2000 S. haematobium cercariae per litre. The urinary bladders from half of the rats were examined after 12 months, and the remaining rats were examined at 24 months. No malignant change was reported; a few rats had bladder lesions reported as ‘sessile polyps’ (Gawish, 1975). [The Working Group noted the inadequate reporting and found no compelling evidence for sustained bladder infection.]

3.1.3. Hamster

In a combination experiment reported in a proceedings volume (James et al., 1974), a group of 50 hamsters [sex unspecified] were exposed to 80 S. haematobium cercariae. No adverse pathological finding in the urinary bladder was reported in the 56-week experiment. [The Working Group noted the inadequate reporting on e.g. the presence of infection.]

A group of 18 male hamsters, eight weeks of age, was exposed to S. haematobium by immersion in water containing 250 cercariae for 1 h. Exposure was repeated three months later. The animals were killed 7–11 months after the first exposure, and the viscera were examined histopathologically. Eleven animals developed manifestations of schistosomal cystitis. In four animals, epithelial hyperplasia of the urinary bladder was related to sites of submucosal reaction to ova. In four other animals, the bladder epithelial changes consisted of both hyperplasia and squamous metaplasia (El-Morsi et al., 1975). [The Working Group noted that hyperplasia of the bladder is unusual in untreated hamsters and that the duration of observation was short in comparison with the lifespan of the animals, so that tumours might have developed in the animals if they had been allowed to live longer.]

3.1.4. Opossum

Eight opossums (Didelphis marsupialis) were exposed to 1000–2000 cercariae of S. haematobium on the shaved skin for 30 min. Between 18 and 53 weeks after infection, two animals were reported to have mucosal fibrous plaques in the urinary bladder. A third animal had multiple epithelial lesions of the bladder that were variably described as hyperplastic, papillomatous, polypoid or tumourous. The presence of eggs was associated with the lesions in two animals, and all three animals had evidence of infection (Kuntz et al., 1971).

3.1.5. Nonhuman primate

Young adult primates, including one talapoin (Cercopithecus talapoin), seven capuchins (Cebus apella), seven squirrel monkeys (Saimiri sciureus) and 11 African baboons (Papio cynocephalus), were exposed percutaneously to 1000–2000 cercariae of S. haematobium and observed for up to 24 months. Epithelial lesions of the urinary bladder, reported to be papillary transitional-cell carcinomas, were found in a talapoin monkey which died 21 weeks after infection and in a capuchin that was killed 56 weeks after infection. An epithelial lesion reported as a papilloma of the ureter was associated with the presence of schistosomal eggs in an African baboon killed one year after infection (Kuntz et al., 1972).

In a further experiment, nine capuchin monkeys (Cebus apella) were exposed via the skin to 1000–2000 cercariae of S. haematobium and were examined for pathological changes in the urinary bladder by laparotomy and cystotomy 94–164 weeks after infection. Six of nine animals showed papillary hyperplasia with or without nodular hyperplasia. In two animals, only focal nodular hyperplasia was seen (Kuntz et al., 1978).

In a study to detect C-type viral particles in tumours, it was reported that four of six capuchin monkeys that had been infected experimentally with S. haematobium developed lesions described as papillary carcinomas of the urinary bladder during periods of observation of 109–111 weeks. Three of the animals also had squamous metaplasia of the bladder epithelium (Kalter et al., 1974). [The Working Group noted the lack of experimental details and of documentation of the pathological findings.]

Kuntz et al. (1975) described the pathological findings and parasitological and radiological observations in two gibbons (Hylobates lar) infected by skin application with 1000 cercariae of S. haematobium. Both animals developed evidence of infection, the most striking change being extensive calcification of the eggs in the urinary bladder. One animal had evidence of papillary and nodular transitional-cell hyperplasia of the bladder, and the other had similar lesions in the ureter. The lesions were described as morphologically similar to the grade-I and grade-II papillary transitional-cell carcinomas that are seen in the bladders of humans. [The Working Group noted the small number of animals and the equivocal diagnoses of the lesions.]

In combination experiments with baboons (Papio sp.) (Hicks et al., 1980; Hicks, 1982), five animals were infected by an abdominal pouch method with 1000 cercariae of S. haematobium and kept for 2.5 years. Four animals had polypoid hyperplasia of the urinary bladder and one had endophytic papillary hyperplasia of the ureter. None of these lesions was considered to be a tumour.

3.2.1. 2-Acetylaminofluorene

Mouse: Two groups of 20 Swiss mice [sex and age unspecified] were administered 0.2 ml of a 1.5% suspension of 2-acetylaminofluorene (2-AAF) in olive oil by stomach tube three times a week [duration unspecified]. One of the groups was repeatedly infected with S. haematobium cercariae by subcutaneous injection [dosing schedules and duration unspecified]. Animals were observed up to 44 weeks, at which time survivors were killed. A third group of 20 mice was infected with S. haematobium alone. Epithelial hyperplasia of the urinary bladder was observed in S. haematobium-infected mice. One of the 2-AAF-treated mice developed a benign villous papilloma of the bladder after 43 weeks. Four of the carcinogen-treated animals infected previously with S. haematobium developed bladder neoplasms at 36–44 weeks; one had an anaplastic infiltrating carcinoma and three had papillomas, two of which had malignant areas (Hashem & Boutros, 1961). [The Working Group noted the short duration, the lack of verification of infection and inadequate documentation of experimental details.]

Rat: A group of 100 white rats [sex, strain and age unspecified], weighing 160 g, were exposed to water containing 2000 S. haematobium cercariae per litre; 45 days after exposure, the rats received intraperitoneal injections of 50 mg/kg bw 2-AAF three times per week for four weeks, followed by a diet containing 0.06% 2-AAF and 1.6% indole for one year. A control group of 100 rats received the carcinogen alone. Ten rats were killed every two months and the bladders examined microscopically. In 80 rats in the combined group killed after six months, all but five had transitional-cell carcinomas of the urinary bladder, as did 7 of the 10 control animals treated with 2-AAF and killed after 10 months (Gawish, 1975). [The Working Group noted the inadequate experimental design, the lack of verification of infection and the fact that the results for the two groups did not differ statistically.]

3.2.2. ortho-Aminoazotoluene

Hamster: In a study reported in a proceedings volume (James et al., 1974), groups of 50 hamsters were exposed to 80 S. haematobium cercariae. Ten weeks later, 0.02 or 0.1% ortho-aminoazotoluene was incorporated into the diet. Administration of the carcinogen alone caused hyperplasia of the urinary bladder epithelium. In the combined group at the 0.1% dose level, malignant changes were seen in the bladder within 24 weeks. [The Working Group noted the inadequate reporting.]

3.2.3. N-Nitrosamines

Hamster: In a combination study reported as an abstract (Hicks et al., 1977), groups of hamsters received a single intravesicular instillation of N-methylnitrosourea and were infected with S. haematobium. Urinary bladder tumours developed in 5/16 hamsters receiving the combined treatment, 0/26 uninfected controls [p < 0.001; Fisher exact test], 0/28 infected animals and 0/19 hamsters treated with N-methylnitrosourea alone. In groups of hamsters treated with N-nitrosobutyl-4-hydroxybutylamine (NBHBA), bladder tumours developed in 9/24 infected hamsters and 5/30 uninfected controls [p = 0.057; Fisher exact test]. [The Working Group noted the inadequate reporting.]

Nonhuman primate: In an experiment designed to simulate the possible proliferative stimulus of S. haematobium infection on cancer growth due to exposure to low doses of N-nitrosamines in humans, small groups of baboons (Papio sp.) were either infected through an abdominal pouch with 1000 cercariae of S. haematobium alone (five animals); received intramuscular injections of 5 mg/kg bw (two animals) or 50 mg/kg bw (three animals) NBHBA per week up to the end of the experiment; or were infected with S. haematobium and administered 5 mg/kg bw NBHBA per week throughout the experiment (10 animals). All surviving animals were killed after 2.5 years. No urinary bladder tumour was found in animals receiving either S. haematobium or NBHBA alone, but three of the baboons receiving the combined treatment had adenomatous lesions of the urinary bladder described by the authors as ‘early or latent adenocarcinomas’ and a fourth had a papillary carcinoma. Three baboons had papillary growths in the ureter (Hicks et al., 1980; Hicks, 1982). [The Working Group had difficulty in interpreting some of the diagnostic terms used in these reports.]

3.3.1. Mouse

Groups of eight male C3H mice, three months old, were injected subcutaneously with one, six or 10–16 lyophilized, immature worms of S. mansoni and were examined for palpable tumours at the injection site every two weeks until termination of the experiment at 21 months. No tumours were found at the injection site. The numbers of survivors were 19/24 at 12 months, 11/24 at 18 months and 9/24 at termination. Of the nine mice killed at termination, three had hepatomas and one had a single pulmonary tumour. The authors reported that the frequency of pulmonary and liver tumours in this strain of mice was similar to that in historical controls (Shimkin et al., 1955). [The Working Group noted that an infectious agent was not employed.]

In several combination experiments in mice (Domingo et al., 1967; Haese et al., 1973; Haese & Bueding, 1976; Bulay et al., 1977; El-Aaser et al., 1978; Kakizoe, 1985) in which control groups of untreated mice or mice infected with S. mansoni only were used, no increase in the frequency of liver tumours was reported. Some of the experiments lasted less than 50 weeks (see section 3.4).

As part of an experiment to study the carcinogenic potential of hycanthone, groups of female Swiss-Webster mice, four weeks of age, were infected by intraperitoneal injection with 40 or 80 cercariae of S. mansoni. Eighteen months later, the incidences of livers with nodules were 15/60 [p < 0.001; Fisher exact test] in the group given 40 cercariae and 1/49 [not significant] in that given 80 cercariae. No nodule was found in uninfected paired groups of 61 and 54 animals (Yarinsky et al., 1974). [The Working Group noted that histological examination was not performed.]

3.3.2. Mastomys natalensis

A group of 200 Mastomys natalensis, about three weeks of age, were injected intraperitoneally with 100 S. mansoni cercariae and maintained until death (up to 2.5 years). Infection was confirmed by examination of faeces for ova during life and examination of liver, gut and mesentery for ova and adult worms after death. At the end of the experiment, 106 animals with evidence of infection were available for evaluation. The incidence of adenocarcinomas of the glandular stomach (23/106) did not differ significantly from that expected in controls (~20%). [The common stomach tumours in M. natalensis were then described as adenocarcinoma but are now recognized as carcinoids.] In contrast, hepatomas were observed in 22 infected animals; such tumours had not been observed in several hundred historical controls. Two animals also developed reticulum-cell sarcomas of the ileum and colon, respectively, associated with schistosomal granulomas (Oettlé et al., 1959).

3.3.3. Hamster

Groups of 35 male and 35 female Syrian golden hamsters were infected by intraperitoneal injection of 15 cercariae of S. mansoni. No increase in tumour incidence was observed over that in uninfected hamsters within 73 weeks (Bulay et al., 1977). [The Working Group noted the lack of verification of infection.]

3.3.4. Nonhuman primate

One case report of a hepatocellular carcinoma in a 12-year-old female chimpanzee (Pan troglodytes) has been published. The animal had been captured in the wild in Sierra Leone when two years of age and had no hepatitis B surface antigen, no antibodies to hepatitis B surface or core antigens and no viral RNA of hepatitis C on arrival at the laboratory, although granulomatous inflammation was seen. After 10 years in captivity, during an intervention before the start of a study of hepatitis, a firm white nodule was discovered in the liver which, upon histological examination, was found to be a well-differentiated hepatocellular carcinoma. No cirrhosis was present, but a severe granulomatous inflammatory reaction was apparent, with remnants of schistosomal egg capsules. On the basis of morphological examination, the eggs were considered to be S. mansoni (Abe et al., 1993).

3.4.1. 2-Amino-5-azotoluene

Mouse: A total of 410 female CBA mice, two months of age, were divided into four groups as follows: 80 untreated, uninfected animals, which served as controls; 95 mice that each received subcutaneous injections of 10 mg 2-amino-5-azotoluene in glycerol once a month for nine months; 100 mice that received a single subcutaneous injection of 30 cercariae of S. mansoni; and 135 mice that were infected with S. mansoni and received 2-amino-5-azotoluene eight weeks later. Between 24 and 52 weeks, six animals from each group were examined periodically for pathological changes in the liver; the remaining animals were maintained until death and were examined for gross liver tumours. At 24 weeks after the beginning of the study, the numbers of animals alive in the four groups were 69/80, 80/95, 31/100 and 35/135, respectively. The authors noted that the high mortality in infected animals was due to the infection. No hepatoma was observed in control animals or in those infected with S. mansoni alone. At 52 weeks of age, the incidence of hepatomas was 1/80 in the group treated with 2-amino-5-azotoluene alone and 13/35 in the group given the combined treatment [p < 0.001; Fisher exact test] (Domingo et al., 1967; Liu et al., 1969).

3.4.2. 2-Naphthylamine and 2-acetylaminofluorene

Mouse: Groups of female Swiss albino mice, six to eight weeks of age, were divided at random into the following groups: one group of 45 mice served as untreated controls; one group of 46 mice was infected with S. mansoni by immersion [technique unspecified] for 1 h in water containing 20–30 cercariae per millilitre; one group of 20 mice received 1% 2-naphthylamine in the diet; a group of 20 mice received 0.06% 2-AAF in the diet; one group of 17 mice was infected with S. mansoni and treated with 1% 2-naphthylamine; and a further group of 22 mice was infected with S. mansoni and treated with 0.06% 2-AAF. Administration of the carcinogens was terminated after 30 weeks owing to severe toxicity. The other experimental groups were continued up to 70 weeks. No liver or bladder tumour was observed in any of the groups. All mice infected with S. mansoni showed granulomatous areas in the portal tracts of the liver and had ova in the faeces (El-Aaser et al., 1978).

A total of 109 female ddY mice, four weeks of age, were divided into three groups: 45 mice received an intraperitoneal injection of 20 S. mansoni cercariae and four weeks later were fed a diet containing 0.03% 2-AAF; 32 mice were infected with S. mansoni and fed normal diets; and 32 uninfected mice were fed normal diet for four weeks and subsequently fed a diet containing 0.03% 2-AAF. A number of animals from each group were killed every 10 weeks for interim examination. The experiment was terminated after 40 weeks. No liver tumour was found in the group infected with S. mansoni only. In the group fed 2-AAF only, the incidence of hyperplastic nodules in the liver was 2/32 (6.3%) at 40 weeks. In the group that was both infected and fed 2-AAF, the incidence of hyperplastic nodules was 9/45 (20%) [p = 0.005; Fisher exact test]. Hepatocellular carcinomas were found in 12/45 mice in the combined treatment group at weeks 29–40 and 0/32 in the group infected with S mansoni [p < 0.001; Fisher exact test] (Kakizoe, 1985).

3.7.1. Dimethylaminoazobenzene

Mouse: Three groups of mice [initial numbers, sex and age unspecified] were either infected with S. japonicum and received no further treatment; were uninfected and fed a diet containing 20 ml 3% dimethylaminoazobenzene in corn oil mixed with 1 kg of rice powder; or were infected with S. japonicum and, 60 days later, fed the diet containing dimethylaminoazobenzene. Groups of mice were killed at various intervals up to 150 days. The authors reported that the mice that received the combined treatment developed severe liver cirrhosis and had faster hepatic cancer formation than the uninfected, carcinogen-treated mice (Shigefuku, 1943). [The Working Group noted the limited reporting of this early study].

3.7.2. 2-Acetylaminofluorene

Mouse: Female ddY mice, four weeks of age, were divided at random into two groups. The first group (77 animals) was infected by immersion of the tail in water containing 40 S. japonicum cercariae and four weeks later were fed a diet containing 0.03% 2-AAF for 40 weeks; the second group (86 animals) was fed basal diet followed four weeks later by a diet containing 0.03% 2-AAF for 40 weeks. Interim killings of animals were made between weeks 9 and 40 of 2-AAF administration. The first liver tumours were observed 16 weeks after administration of 2-AAF in the infected group and at 37 weeks in the uninfected group. At 40 weeks, the incidence of liver tumours was 24/77 in carcinogen-treated, infected mice and 6/86 in carcinogen-treated, uninfected mice (p < 0.005, χ² test). The tumour types in the two groups, respectively, were: hyperplastic type 1 nodules, 6 and 4; hyperplastic type 2 nodules, 10 and 2 (p < 0.01, χ² test); and hepatocellular carcinomas, 8 and 0 (p < 0.005, χ² test) (Miyasato, 1984).

4.1.1. Humans

(a) Schistosoma haematobium

Many of the most severe pathological manifestations of schistosomiasis are due to a large extent to a physical and immunological response of the host to the eggs (Parra et al., 1991). A periovular area of granulomas surrounded by an exudative cellular reaction consisting of many polymorphonuclear leukocytes, lymphocytes and eosinophilic cells is found to occur in most granulomatous areas (Nawar et al., 1992; Lukacs et al., 1993).

Clinical and pathological evidence for ‘early stage of infection’ (haematuria and dysuria) is seen in the majority of infected children and young adults (King et al., 1988). In contrast, ‘late stage of infection’ may be less symptomatic but associated with structural urinary tract diseases. Asymptomatic infection may still be associated with urinary tract lesions (Abdel-Salam & Ehsan, 1978).

da Silva Lopes (1984) reported a pathological study of 210 malignant tumours (206 carcinomas) of the bladder in Luanda, Angola. Of the 164 carcinomas associated with schistosomiasis, 122 were of the ‘spinocellular’ type, 15 were ‘urothelial’, 13 were ‘urothelial plus epidermoid metaplasia’, 8 were adenocarcinomas and 16 were undifferentiated carcinomas. Of the 42 carcinomas not associated with infection, 30 were ‘urothelial’, 6 were ‘urothelial plus epidermoid metaplasia’, 3 were ‘spinocellular’ and 3 were undifferentiated carcinomas.

(i) Early stage of infection

The most significant pathophysiological disease sequelae of the early stage of S. haematobium disease occur in the ureters and urinary bladder. Eggs are deposited in particularly large numbers at the lower ends of the ureters. Ureteric lesions result in anatomical or functional stenosis, leading to hydroureters and hydronephrosis. At the site of egg deposition in tissues, circumoval granulomas, fibrosis and muscular hypertrophy may be demonstrated histologically. The same pattern of tissue involvement is seen in the urinary bladder (Smith et al., 1974).

Two major autopsy studies—one in Ibadan, Nigeria (Edington et al., 1970) and the other in Egypt (Smith et al., 1974)—contributed significantly to our appreciation of the pathological changes in schistosomiasis caused by S. haematobium. Edington et al. (1970) studied 673 unselected cadavers in Nigeria and found S. haematobium in 20%; 183 of the autopsies were performed on individuals under 19 years of age. In Egypt, Smith et al. (1974) examined specimens taken at 190 consecutive autopsies and found evidence of S. haematobium infection in 117 (61.6%).

The morphological findings in early active S. haematobium disease comprise polypoid granulomatous lesions surrounding the parasite eggs. In the urinary bladder, the pathological manifestations are polyposis and/or ulceration. S. haematobium-induced bladder polyps consist of large inflammatory masses containing schistosome eggs. The deposition of eggs may be apicocentric, basocentric or diffuse. Apicocentric ova deposition usually occurs at the apex and dome of the urinary bladder, whereas basocentric deposition occurs predominantly in the trigone and lower posterior wall (Christie et al., 1986b). Bladder polyposis is responsible for the haematuria seen in the early stages of infection and in obstructive disease. The other major morphological lesion is bladder ulceration, which may be due to polyp sloughing. Histological examination of bladder tissue in the early stage of infection demonstrates hyperaemia, granulomas around nests of schistosome eggs and early fibrosis and hypertrophy of muscle. Urethelial hyperplasia, metaplasia and dysplasia were significant in all stages of the disease in the series of Smith et al. (1974), hyperplasia occurring in 38% of autopsied cases and 21% uninfected controls, metaplasia in 31.6% cases and 11.5% controls and dysplasia in 27.2% cases and 8.5% controls.

(ii) Late stage of infection

The change from early-stage to late-stage schistosomiasis caused by S. haematobium occurs with age, decrease in parasite load (as determined by urinary egg excretion) and diminished manifestations of acute inflammatory disease, e.g. haematuria. Morphologically, urinary bladder disease in late-stage infection manifests as schistosomal ulcers or sandy patches (Smith et al., 1974). Chronic schistosome-related bladder ulcers usually occur in individuals with previous heavy infection. They are located mainly in the posterior part of the bladder. Sandy patches occur late in infection, most frequently in the trigone area and are covered by irregularly thickened or atrophic mucosa. Histologically, old granulomas may be found in the submucosa and muscularis surrounding disintegrating or calcified ova. In many instances, fibrosis and scanty round-cell infiltration may be seen. Differences between the early and late stages of infection are summarized in Table 8. The eggs of S. haematobium tend to calcify and to remain in tissues longer than those of S. mansoni (Cheever et al., 1978).

Table 8.

Differences between early and late stages of infection with S. haematobium.

The concordance of lesions of chronic infection with those of urethelial cancer has been known for over a century. In a series of 1095 patients with urinary bladder cancer in Egypt, S. haematobium eggs were found in 82.4% of cases (El-Bolkainy et al., 1981). Well-differentiated squamous-cell carcinomas of the bladder were seen predominantly in patients with eggs and at an earlier mean age than transitional-cell carcinomas. The morphological changes in the urinary bladder associated with the late stage of infection included a spectrum of hyperplasia, squamous metaplasia, dysplastic changes and predominance of squamous-cell carcinoma. Of the 798 squamous-cell carcinomas, 691 occurred in S. haematobium-positive samples and 107 in patients with no eggs. Of the 148 cases of transitional-cell carcinoma, 103 were in patients with eggs and 45 in those without.

Similarly, urethelial hyperplasia and squamous metaplasia have been associated with urinary schistosomiasis. Squamous-cell metaplasia of the bladder occurs at increased frequency in schistosomiasis patients and in young people in populations at high risk of squamous-cell carcinoma (Khafagy et al., 1972). Although granulomas occur in both the ureter and bladder, carcinomas occur predominantly in the bladder. In 30 patients with bladder carcinoma in Egypt, the tissue surrounding the tumours usually contained a higher concentration of S. haematobium eggs than other areas in the bladder: The egg burden in tissue surrounding the tumour was almost twice the mean in the remainder of the urinary bladder (Christie et al., 1986a).

Father pathological sequelae of S. haematobium infection can be seen almost anywhere in the body. The infection may also be associated with other clinical conditions, such as bladder calcification, urolithiasis and pyelonephritis. Most of these lesions are thought to be related to the inflammatory and subsequent fibrotic responses that follow egg deposition in tissues (Cheever et al., 1978).

4.1.2. Experimental systems

4.2.1. Humans

Numerous explanations have been offered for the proposed association between schistosomiasis and human cancers. Generally, these can be categorized as involving: exogenous and endogenous agents which induce DNA damage (Abdel-Tawab et al., 1968a,b; Fripp & Kean, 1980; Hicks, 1982; Gentile, 1991) or possible tumour promoting activity (Ishii et al., 1989); altered host metabolism (El-Aaser et al., 1982; Gentile, 1985; Gentile et al., 1985); pathological changes leading to increased cell proliferation (Ishak et al., 1967; Brand & Brand, 1980a,b; Rosin et al., 1994); and immune reactions (Raziuddin et al., 1991, 1992, 1993; Gentile & Gentile, 1994).

Endogenous agents may be introduced into schistosome-infected organs in several ways. For example, quantitatively altered tryptophan metabolism in S. haematobium-infected patients results in higher concentrations of certain metabolites (e.g. indican, anthranilic acid glucuronide, 3-hydroxyanthranilic acid, L-kynurenine, 3-hydroxy-L-kynurenine and acetyl-L-kynurenine) in pooled urine (Abdel-Tawab et al., 1966a, 1968b). Some of these metabolites have been reported to be carcinogenic to the urinary bladder in implantation experiments (Allen et al., 1957; Bryan et al., 1964; Bryan, 1969; Röhl et al., 1969).

Other endogenous agents may be involved in secondary bacterial infection. Bacterial urinary tract infections such as those that occur subsequent to the late sequelae of S. haematobium infection may play an intermediary role in the genesis of squamous-cell carcinoma. Secondary bacterial infection of Schistosoma-infected bladders is a well-documented event (Laughlin et al., 1978; Hill, 1979; El-Aaser et al., 1982; Hicks et al., 1982).

Nitrosamines formed by bacterial catalysis (or via urinary phenol catalysis) of the nitrosation of secondary amines with nitrites have been detected in urinary bladders from S. haematobium-infected patients; they may be carcinogenic to bladder mucosa (Hicks et al., 1977, 1978, 1982; Tricker et al., 1989, 1991). Mostafa et al. (1994) also demonstrated the presence of nitrates and nitrites in the saliva and increased concentrations of N-nitroso compounds in the urine of S. mansoni- or S. haematobium-infected people who were not on controlled diets. The etiological significance of these findings is, however, unclear in the light of the finding that urine from schistosomiasis patients is not mutagenic (Everson et al., 1983).

Nitrosamines have been detected in the urine of paraplegic patients with urinary tract infections due to urinary stasis (Hicks et al., 1977, 1978).

In a US case–control study in which 2982 urinary bladder cancer patients (97% with transitional-cell carcinomas) were compared with 5782 controls (Kantor et al., 1984), odds ratios of 1.5 (95% CI, 1.3–1.8) in males and 1.2 (0.9–1.5) in females reflect an association with one or two past urinary tract infections, and odds ratios of 2.0 (1.6–2.6) in males and 2.1 (1.6–2.7) in females reflect an association with three or more such infections. For the 39 patients with squamous-cell carcinomas, odds ratios of 1.9 (0.7–4.8) for having had one or two infections and 4.8 (1.9–11.5) for three or more infections were found for the two sexes combined. Adjustments were made for race, age, smoking and, for squamous-cell cancer, sex.

On follow-up of 6744 British paraplegic patients (who are subject to frequent urinary tract infections), 25 urinary bladder cancers were identified (El Masri & Fellows, 1981). On the basis of information for an otherwise comparable population, 1.6% of these would have been expected to be of squamous origin, whereas 44% actually were (estimated relative risk, 49; 95% CI, 20–119). In Uganda, squamous-cell bladder cancers are commonly seen in the absence of S. haematobium infection but in the presence of other urinary tract abnormalities (Anthony, 1974).

One of the prevalent theories for the association between schistosomal infection and cancer is that elevated levels of the enzyme β-glucuronidase in the host could increase the release of carcinogenic metabolites from their glucuronides. No data are available at present to confirm this association, although schistosome-infected humans are known to have elevated β-glucuronidase activity in urine (Fripp, 1960; Abdul-Fadl & Metwalli, 1963; Fripp, 1965; Abdel-Tawab et al., 1966b, 1968a; Norden & Gefland, 1972; El-Sewedy et al., 1978; El-Zoghby et al., 1978; El-Aaser et al., 1979). The cause of the increase in β-glucuronidase activity in individuals suffering from schistosomiasis is unknown.

Several studies provide evidence for genetic damage in schistosomiasis patients. Sister chromatid exchange and micronucleus frequencies are increased slightly in peripheral blood lymphocytes harvested from schistosomiasis patients (Shubber, 1987; Anwar, 1994) and micronuclei were more frequent in urothelial cells from chronic schistosomiasis patients than in controls (Rosin & Anwar, 1992). The mean frequency of micronuclei was reduced significantly after treatment with praziquantel, which may indicate that infection is involved in chromosomal breakage in the urothelium (Anwar & Rosin, 1993).

No mutation was detected at codon 12 of the H-ras oncogene in nine squamous-cell carcinomas associated with schistosomiasis (Fujita et al., 1987). Mutations of the p53 tumour suppressor gene were detected in six of seven squamous-cell carcinomas associated with S. haematobium; no specific pattern of mutation emerged, in contrast to the pattern seen in transitional-cell carcinomas related to tobacco smoking (Habuchi et al., 1993) O⁶-Methyl-deoxyguanosine was detected in DNA from 44 of 46 Egyptian samples of bladder tissue, 38 from tumour tissue and eight from uninvolved bladder mucosa, and in 4 of 12 normal samples of bladder of European origin (Badawi et al., 1992a).

4.2.2. Experimental systems

Schistosoma haematobium

A number of studies from Africa have shown that the estimated incidence of urinary bladder cancer is higher in areas with a high prevalence of infection with S. haematobium than in areas with a low prevalence. For example, urinary bladder cancer as a proportion of all cancer appears to be 10 times commoner among men in Egypt than among men in Algeria. Several other observations support an association between the occurrence of urinary bladder cancer and S. haematobium infection: the estimated incidence of urinary bladder cancer was related to the proportion of cancerous urinary bladder specimens containing S. haematobium eggs or egg remnants; the sex ratio of urinary bladder cancer cases varied widely and corresponded to the relative involvement of men and women in agricultural work (a risk factor for S. haematobium infection); and squamous-cell cancers of the urinary bladder were proportionately commoner in populations with a high prevalence of infection with S. hematobium and a high proportion of urinary bladder cancers showing histological evidence of infection than in areas without these characteristics.

Many cases of urinary bladder cancer have been reported in association with schistosomal infection of the urinary bladder. Other cancers have been reported in association with infection with S. haematobium including, particularly, cancer of the cervix.

Seven case–control studies of the association between S. haematobium infection and urinary bladder cancer have been reported. S. haematobium infection was measured variously by presence of eggs in urine, pelvic X-ray, rectal biopsy, biopsy of the urinary bladder and digestion and centrifugation of urinary bladder tissue. All of the studies were hospital-based and in none was the correspondence between the population giving rise to the cases and that sampled for the controls demonstrated or addressed in the analysis. Possible confounding by age and sex was not considered in four studies. In three of these four studies, the method of measurement of past infection with S. haematobium differed between cases and controls. Possible confounding by smoking was considered in only one study. Six of the seven studies showed significant, positive associations between the occurrence of urinary bladder cancer and infection with S. haematobium, with estimated relative risks ranging from 2 to 14. Confounding is not likely to explain the strong associations seen in these studies.

Schistosoma mansoni

A number of cases of liver cancer, colorectal cancer, giant follicular lymphoma and some other cancers have been reported in association with S. mansoni infection.

Schistosoma japonicum

Mortality from liver cancer and prevalence of infection with S. japonicum have been found to be positively correlated in Japan but not consistently so in China. Mortality from and, in one study, incidence of colorectal cancer were strongly, consistently and significantly correlated with various measures of infection with S. japonicum in many studies across provinces, counties and communes in China.

In three case–control studies of liver cancer and infection with S. japonicum in Japan and China, the estimated relative risks for the association varied from 2 to 10. The relative risk remained elevated in patients who did not have antigens to hepatitis virus. The two studies giving the highest estimated relative risks were hospital-based and did not address the issue of correspondence between the population giving rise to the cases and that sampled for the controls. In one of these studies, possible confounding by age and sex was not controlled for.

In one hospital-based case–control study of gastric cancer in Japan, the estimated relative risk for S. japonicum infection, based on the presence of eggs in tissue, was 1.8 and was significant. Possible confounding by age and sex was not controlled for, and the issue of correspondence between the population giving rise to the cases and that sampled for the controls was not addressed.

Three case–control studies of colorectal cancer and infection with S. japonicum have been reported from China and Japan. In one, the estimated relative risks for cancer of the colon in association with the presence of eggs in tissue was about 2.5 and was significant. Possible confounding by age, sex, area of residence, smoking and family history of cancer of the colon was controlled for in this study. In the two other studies, the estimated relative risks were 1.2 for colon cancer in both studies and 8.3 for rectal cancer in one study with control for possible confounding by age, sex, place of residence and occupation.

Overall evaluations

Infection with Schistosoma haematobium is carcinogenic to humans (Group 1).

Infection with Schistosoma mansoni is not classifiable as to its carcinogenicity to humans (Group 3).

Infection with Schistosoma japonicum is possibly carcinogenic to humans (Group 2B).