Amoebiasis and Giardiasis the Global Impact of Two Common Intestinal Protozoan Infections

Drug Invest. 8 (Suppl. 1): 1-18,1994 0114-2402/94/OOO1-OOO1/S09.00/O

Amoebiasis and Giardiasis The Global Impact of Two Common Intestinal Protozoan Infections

C.c. Cook Hospital for Tropical Diseases, London, England

Summary Intestinal protozoan infections involving humans are by no means confined to tropical/subtropical countries; however, it is here that maximal prevalence, and consequently, morbidity, assumes a major practical importance. Coccidial infec• tions (Cryptosporidium spp., Isospora belli, Sarcocystis hominis and Cyclospora spp.), and BLastocystis hominis and Microsporidium spp., which were previously underrecognised, have come to the fore in the present era, largely in association with HlV infection. Nevertheless, these organisms can induce self-limiting infec• tion in travellers and other individuals in countries in which standards of sanita• tion/public health are less than satisfactory. Overall, however, there can be no doubt that Giardia Lamblia (also known as G. intestinalis and G. duodena lis) and Entamoeba histolytica are numerically the most important protozoan para• sites to involve the gastrointestinal tract. Whereas trophozoites of the former organism are virtually confined to the small intestinal lumen, the extra-intes• tinal manifestations of E. histoLytica infection are of greater importance than its colorectal pathogenic properties. This review concentrates on the prevalence of the 2 infections in different populations; this reflects both incidence and outcome, including duration of ill• ness. Epidemiology is also addressed. Routes of transmission of the 2 infections are covered, together with morbidity and, in the case of E. histoLytica infection, mortality also.

When viewed within a broad scenario, data 'killers' in developing countries, such as diarrhoeal collected over the last 15 to 20 years suggest that disease, Plasmodium Jalciparum infection, tuber• prevalence and transmission patterns of En• culosis (and other respiratory tract problems), tamoeba histolytica infection have not changed schistosomiasis and childhood infections. significantly during that periodV] However, in• creasing migration of the population into urban 1. Entamoeba histolytica Infection areas in developing countries, coupled with an in• 1,1 Prevalence of E. histolytica Worldwide crease in the size of urban slums and wars (frequent in these countries), are probably accelerating the In a world context, E. histolytica is very wide• spread of infection; consequently, resultant dis• spread and produces an enormous burden of mor• ability might well be greater in future years. bidity and mortality resulting from colitis and in• Whereas E. histolytica infection causes a great vasive hepatic disease (liver 'abscess').[l] Some deal of morbidity and mortality (see sections 1.4 10% of the world's population harbours this proto• and 1.5), this is less than that caused by the major zoan parasite (in most as the encysted form), al- 2 Cook

though only a small proportion experiences dis• parts of rural Tanzania;[I] faecal parasitological ease. In certain countries, e.g. Mexico, this infec• surveys have given a range of 0 in rural parts of the tion constitutes one of the leading 10 causes of Dominican Republic to 49.4% in urban Peru. No death. In 1984, it is probable that 500 million indi• method assessing the potential pathogenicity of viduals were infected with E. histolytica and, of E. histolytica (i.e. its zymodeme status) is currently these, 40 million experienced resultant disabling in routine use. colitis or extra-intestinal disease. Furthermore, at least 40 000 deaths during that year were estimated 1. I. 1 Geographical Evidence for Local to have resulted from this infection)ll The accu• E. histolytica Prevalence Probably the most reliable data available for racy of reported prevalence rates depends very prevalence rates comes from Mexico and. to a largely on the quality of diagnostic procedures, lesser extent, other parts of South America. In which are principally parasitological and serolog• northeastern Brazil. 24 of 334 serum samples ex• ical, although seroepidemiological surveys are amined by an enzyme-linked immunoabsorbent as• very important. Recorded prevalence is also de• say (ELISA) gave positive results for E. histolytica pendent on the enthusiasm of medical/scientific antibody.l4J In a suburban community with low so• workers in a particular locality in reporting dis• cioeconomic status in Venezuela, E. histolytica in• ease, and in most developing countries this is inad• fection prevalence (as assessed by faecal micros• equate. Many surveys have used a nonrepresenta• copy) was 8.7% (29 of 342), with most cases being tive population sample;lll even hospitalised cyst carriers.[SI The Republic of South Africa is patients supposedly without intestinal symptoms another country from which extensive data are can, for example, be construed as a representative available.[61 In a series of 5087 adult admissions sample in a survey of E. histolytica infection. Fur• involving invasive amoebiasis to a medical unit at thermore, most surveys conducted outside hospital Durban, intestinal and liver disease accounted for do not mention the sampling technique used to en• 60% and 40%, respectively, and the mortality rate sure a random sample. and patient compliance is for both forms of disease was 1.9%. In the same not always stated. city, the overall mortality rate from intestinal dis• Those experiencing diarrhoea (or another ill• ease was 27% (see also section 1.4). ness) may be more willing to give a faecal and/or In a study carried out on the Thai-Cambodian blood sample. in the hope of benefiting from the border, the highest incidence of amoebic dysentery study. Interestingly, an age-related antibody re• was 63 per 1000 (6.3%) in children 12 to 23 months sponse to E. histolytica has been documented in 0Id.[71 On the West Bank of Jordan. during 1981- India)21 It is also essential to appreciate that with 1986, of 22 900 faecal samples obtained from pa• both of the infections covered in this review excre• tients attending the Central Medical Laboratory at tion of cysts tends to be intermittent. In addition, Nablus,l81 E. histolytica was present in 22.9%. As many surveys do not state the number of faecal with other intestinal parasites, the peak incidence specimens examined, or indeed the interval be• occurred during the summer and early Autumn, tween their collection. Arroyave et al.[3) have with a low incidence in winter and early spring. In stressed that circannual variability of infection a survey at Mahe, Seychelles, 21 of 313 cultures with E. histolytica in Mexico should be carefully grew E. histolytica,l91 and when zymodeme analy• taken into account in intervention studies. sis was employed, 8 organisms were pathogenic There are huge differences in the prevalence and 40 nonpathogenic (E. dispar as defined by rates of E. histolytica infection at different geo• Brumptf101 see section 1.3.1). graphical locations. Using available serological Despite these findings, in some tropical coun• techniques, reported differences in prevalence tries the prevalence rate is low. In Jamaica, a recent vary from 0 in Surinam blood donors to 57% in report of 3 cases of such infection represented the

first documentation of amoebiasis there for more (also known as G. intestinalis and G. duodenalis); than 2 decades. Also, invasive hepatic disease can E. histolytica was present in 18.4% of the residents occur in individuals who have never travelled to at the rehabilitation centre. In a 190-patient institution an endemic areaJIl] These authors have recorded for mentally retarded individuals in Japan, 20% a case contracted in Australia. In Europe and North had either cysts or trophozoites of E. histolytica in America, the disease usually occurs in well de• a faecal sample;f22] 38% of them were serologi• fined groups, e.g. recent travellersfl2 ] and immi• cally positive. grants, active homosexual men,l13.14] and those confined to institutionsJl5] 1.2 Epidemiology of E. histolytica Infection

1.1.2 Prevalence ofE. histolytica in High prevalence rates of E. histolytica infection Special Groups in developing countries are associated with pov• Analysis of faecal samples from 2700 individ• erty, poor sanitation, overcrowding, and a warm uals who travelled from a tropical to a temperate climate)61 Available epidemiological data are lim• country f16] showed that 4.0% had evidence of ited for several reasons, including a) a tendency to E. histolytica infection, but in only 5 was a patho• include the most accessible people, i.e. those living genic zymodeme identified. The authors con• nearest to the road; b) intermittency of cyst excre• cluded that travellers to the tropics had a 0.3% tion; c) the fact that faecal analysis reflects preva• (1 :340) risk of acquiring invasive amoebiasis, and lence at one point in time and not the actual prev• a 92.3% risk of an E. histolytica infection remain• alence of infection (see above), which might ing asymptomatic. In a group of travellers to approach 100% in an endemic area; d) variability Phuket, Thailand, E. histolytica infection showed in the laboratory methods used;f23] and e) lack of a significant relationship to consumption of drinks suitable/consistent criteria for diagnosing the dis• containing ice, ice-cream, and raw fruit in ice) 17] ease and defining its severity. Most data relating to 525 Czechoslovaks who had worked in 50 tropical prevalence havc been obtained from faecal analy• and subtropical countries in Asia, Africa, and Latin sis and serological surveys. E. histolytica-specific America were tested for evidence of E. histolytica serum IgG, IgA, IgM, and IgE antibodies might infection.fI8 ] A total of 74 (14.1 %) were infected prove to be valuable indices in the detection of with one or more intestinal pathogen, and 3.8% of infection in an endemic area.l24] the organisms were identified as E. histolytica. Several host factors influence the preva• E. histolytica does not act as an 'opportunistic' lence/epidemiology of severe disease. Overall, the infection in the presence of HIV infection. A recent infection is more common in children than in study in Mexico has confirmed that patients with adults,r6,25] and it is more often encountered in AIDS did not have an increased prevalence of in• men, in pregnant women, during the summer fectionJl9] In another study, 16.3% of homosexual months, and in association with undernutrition and men were infected. fI6] Sorvillo et a1J20] have con• altered iron status. Ahigh prevalence ofHLA-DR3 cluded that 'amoebiasis trends (as a result of the has been recorded in patients with invasive hepatic association with male homosexuality) may be a disease among Mexican children;f26] however, useful predictor of human (HIV) transmission and further studies are required. A study carried out future rates of AIDS among gay men'. in rural Indiaf27 ] suggested that' ... invasive 218 residents in a nursery, a foster home, and a E. histolytica infection (mostly asymptomatic) rehabilitation centre for handicapped children at Abha, evokes good gut immunity in the host with clearing Saudi Arabia, were examined for the presence of of the parasite from the colon and/or resistance to intestinal parasites,l21] Approximately 30% were reinfection', and that a 'high prevalence of amoe• shown to possess an asymptomatic infection bic antibodies indicates good 'herd' immunity'. with either Entamoeba spp. or Giardia lamblia Moreover, pathogenic zymodemes are essential

Fig. 1. Entamoeba histolytica trophozoite from amoebic dysentery faeces. The circular bodies in the cytoplasm are red cells. The parasite is actively motile and generally moves in one direction at a time. The background shows bacteria. red cells and degenerate white cells. Interference contrast x 1000. Enlarged by 9.6.(From Zaman[106]). for invasive disease to occur and, indeed, demon• months. Desiccation following exposure to bright stration of these is of paramount importance in any sunlight or a high ambient temperature signifi• epidemiological studyPS.30] These reports were cantly reduces cyst viability. Trophozoites, which made in India and Mexico, respectively. are extremely sensitive to environmental influ• In endemic areas, some 12% of the population ences, are unimportant with regard to transmission. is infected with E. histolytica; however, only 10% They rapidly die on exposure to dry air and are also present with acute disease,f61 and a similar percent• destroyed by gastric acid and other digestive en• age probably harbours noninvasive organisms zymes. However, when trophozoites are directly (i.e. E. histolytica acting as a 'commensal'). Be• introduced into the colon of another animal, dys• tween these 'polar' forms of infection, many gra• entery results (fig. I). Cyst excretion, either by dations of pathological change and clinical expres• convalescent or asymptomatic carriers, constitutes sion occur, although only sparse reliable data are an important vehicle for transmission (fig. 2). available on this group, which warrants much greater study. 1.3.1 The Protozoan-Host Equilibrium Viable cysts ingested in water and food or from 1.3 Transmission of E. histo/ytica faecally contaminated hands hatch in the intestinal Various factors involved in the transmission of lumen, producing an infection which may give rise E. histolytica have been reviewed by Walsh)3l] to symptoms. The reservoir for infection lies in hu• With all pathogenic organisms, both parasite and mans; however, evidence of transmission from a host factors are involved. The mature quadrinucle• subhuman primate to man has been recorded.[33] ate cyst is the stage in the life-cycle that is trans• Canine and feline amoebae have also been experi• mitted from one individual to another (less mature mentally transmitted to many mammalian species. cysts have a lower survival) and cysts can remain Host susceptibility is complex, with immuno• viable in a damp/moist environment for several logical, age related, sexual, ethnic, nutritional, and

cultural factors all being involved. After tissue in• humidity and extremes of temperature. Moreover, vasion, humoral antibodies and cell-mediated im• the infecting dose may be lower and the persist• munity result, but the degree of protection against ence longer, and these strains may possess the abil• further infection remains unclear. ity to multiply in the intestinal lumen under more On theoretical grounds, Brumpt suggested in variable conditions, as well as being able to repro• 1925 that there are 2 strains of E. histolytica: duce/mature more rapidly so that a larger propor• E. dysenteriae, which produces invasive disease, tion of cysts is produced. and E. dispar, which is noninvasive.[IO] Recent re• An oral inoculum of 2000 to 4000 cysts consis• search using DNA technology has confirmed the tently produces E. histolytica infection, but it is hypothesis that pathogenic and nonpathogenic likely that a single cyst is sufficient under certain zymodemes are genetically distinct. It has also conditions. After ingestion, mature cysts appear in been suggested that transmission characteristics a faecal sample in approximately 5 days; however, differ for pathogenic and nonpathogenic the median incubation period from ingestion to the strainsJ3l] N onvirulent strains might possess production of symptomatic disease is usually greater environmental resistance and be able to cir• much longer (of the order of 2 to 6 weeks), and culate under more hygienic conditions, and in less depends on the inoculum size. Invasive hepatic

, .... -_ ... ------.... , I , t Extraintestinal

Luminal drugs (cysts) (1) Diloxanide (2) Diiodohydroxyquinoline (3) Paromomycin

Invasive intestinal amoebias is (trophozoites) (1) TInidazole (2) Metronidazole (3) Emetine or dehydroemetine

Extraintestinal amoebiasis (trophozoites) (1) Tinidazole (2) Metronidazole (3) Emetine or dehydroemetine (4) Chloroquine , '------~

Fig. 2. Life-cycle of Entamoeba histolytica showing transmission, site of infection, and sites of action of chemotherapeutic agents[32)

disease may not occur until many years later. Ex• are used over a long period. Although information cretion of cysts is erratic, and can amount to 15 on the most appropriate methods for sewage treat• million daily. ment is limited, the viability of cysts is substan• tially reduced by exposure to bright sunlight and a 1.3.2 Risk Factors for E. histolytica Carriage warm temperature. In a hot dry climate, cyst sur• A variety of factors are involved in E. histolytica vival may be less than 24 hours, and dry weather infection, and many of these are interconnected. for 3 days kills cysts on the surface of crops that Individuals of low socioeconomic status are usu• have been irrigated with polluted water or fertilised ally exposed to poor sanitation, inadequate water with night-soil. supplies, poor hygiene, crowded living conditions Water also plays a significant role in the trans• and, in addition, are poorly educated. Most of these mission of histolytica infection, and 2 factors are factors operate in a developing country setting. E. important: 1) the purity of the supply, which must Seasonability and environmental conditions, and be free from faecal contamination; and 2) adequate sexual practices are also involved in transmission. water per capita and easy accessibility. Personal Personal hygiene and attention to hand washing hygiene is not directly affected by the quality of are of paramount importance in the prevention of the water supply. Although E. histolytica cysts amoebiasis. Culturally appropriate health educa• have only rarely been isolated from the domestic tion must be instituted and children in particular water supply, infections are common when the wa• must be taught to defecate in a latrine. Personal ter is faecally contaminated. In an endemic area, hygiene is crucially important in mental hospitals person to person transmission is probably more im• and other institutions. The value of installing a pu• portant than water-borne infection. rified water supply with regard to transmission rates has received a great deal of attention. It is E. histolytica cysts can be mechanically trans• probable that only a minor proportion of cyst trans• mitted by insects, including cockroaches and flies, mission takes place via contaminated water; infec• which contaminate food with their faeces and/or tion more often follows ingestion of contaminated vomit. Although the importance of insect transmis• food. sion is not well established, transmission by house flies is certainly possible under field conditions. 1.3.3 Methods ofE. hisfolytica Transmission E. histolytica cysts have been detected on the ex• The importance of the food handler in convey• ternal surface of flies 4 minutes after exposure, in ing infection is well established.131 ] In one study, the intestinal tract for up to 240 minutes, in vomit the author concluded that ' ... the infected mother for 64 minutes, and in faeces for 254 minutes. In bears most of the responsibility for infection of the cockroach, cysts can survive both in the gut and young children with this parasite in Nigeria'. A faeces for 48 hours after feeding on infected human closely related means of transmission is direct con• faeces. Overall, however, it seems likely that the tamination of the oral cavity by faecally soiled role of insects in transmission is relatively small, hands and fingernails. This is especially important and insect control should therefore not receive ma• in children, and in mental hospitals. jor attention in control strategiesJll In developing countries, sewage contamination of foodstuffs is important, and human night-soil 1.3.4 Identification of Transmitters of E. hisfolytica is frequently used as a crop fertiliser. Also, fresh• Although any effective control strategy should ening vegetables and skinned fruit with polluted concentrate on major excreters of cysts, to date no water, should be actively discouraged. These satisfactory method for their identification has be• foodstuffs should be carefully washed using fil• come available. Such individuals a) excrete a large tered/purified running or boiled water. Treatment number of mature virulent cysts; b) have poor of water with iodine, chlorine, or silver solutions hygiene, especially in food preparation and food kills E. histolytica cysts only ifhigh concentrations handling for others; c) have poor excreta disposal

habits (children and institutionalised individuals spontaneous eradication within 5 months is per• are especially relevant); and d) although carriers, haps the usual sequel. Data obtained in Mexico in feel well and circulate freely among a susceptible 1984 (population 77 million) indicated that 3.9 population)3!] If major transmitters could be eas• million people seroconvert for E. histolytica annu• ily identified, they could be given a luminal ally, and 5.5 to 7.0 million moderate to severe cases amoebicide to prevent or minimise cyst excretion. of clinical disease result. [1] A similar number will Alternatively, a vaccine (when available) would probably have had mild, transient diarrhoea (5 to induce local immunity and reduce cyst excretion. 7 days) resulting in a visit to a health worker. In The proportion of carriers in an endemic equilib• Mexican children, up to 15% of cases of acute di• rium is a measure of transmissability within the arrhoea necessitating hospitalisation are associ• population. ated with E. histolytica infection. At Caracas, Ven• At present, mass chemotherapy of open popu• ezuela, 11 % of those suffering from colorectal lations is not feasible; available chemotherapeutic disease were infected. [I] agents require several days to several weeks of ad• Available evidence indicates that the preva• ministration to eliminate carriage in the majority lence of E. histolytica infection in most of Asia is of those treated. If a satisfactory short course lu• lower than that in southern America, although sat• minal amoebicide were produced, mass chemo• isfactory data are often sparse)l] In sub-Saharan therapy might be possible. Africa, however, the proportion of the population E. histolytica infection differs from a viral dis• suffering from invasive disease is probably com• ease, in which an accumulation of immune individ• parable with that in Mexico. Therefore, morbidity uals limits the overall number of cases: a) infected is likely to be high, a conclusion that is supported individuals recover spontaneously, but only sev• by personal observation. The overall effect of E. eral months to several years after exposure; b) histolytica infection on a population may be more because infection does not impart full immunity, significant than the mortality rate, as those more and after recovery (spontaneous, or after chemo• severely infected are often in their economically therapy) individuals are immediately susceptible productive years of life. to a further infection; and c) spread of infection is There can be no doubt that because of the sub• by the person to person route, either directly or less stantial morbidity (and mortality) associated with frequently via water, food, or flies. E. histolytica infection, this disease warrants far greater study with regard to preventive methods, 1.4 Morbidity Associated with diagnosis, and satisfactory chemotherapy. E. histolytica Infection 1.5 Mortality Associated with A major problem in delineating morbidity (and E. histolytica Infection mortality) involves difficulty in diagnosing dis• ease and in defining mild, moderate, and severe One estimate of the mortality associated with cases)!] Disease produced by E. histolytica infec• E. histolytica infection is that amoebiasis causes tion is often indistinguishable from that caused by 10 000 to 30 000 deaths annually in Mexico other intestinal pathogens. Unless clear definitions alone,fll the majority in men during the third or of 'dysentery', 'colitis', 'subacute colonic amoe• fourth decades of life. Pregnant and postpartum biasis', and 'carrier state' are established, it is ex• women also have an excessive risk of severe dis• ceedingly difficult to compare different studies of ease and death. Furthermore, the death of a mother morbidity (and mortality), incidence, treatment in a developing country often results in malnutri• outcome, and serological response. There is, for tion and death in her young children)'] example, good evidence that the cyst carrier rarely For every case of liver 'abscess', some 1 to 5 experiences significant clinical disease;f34,351 adults are admitted to hospital with colitis. Inva-

sive amoebiasis, involving liver and/or colon, fre• faecal sample in 73% of patients, and in 85% when quently presents as an acute abdominal emer• 3 consecutive samples were examined. In contrast, gency.l36] The case fatality rate of liver 'ab• G. Lamblia was detected in 44% of duodenal fluid scess'/colitis sufficiently severe to necessitate aspirates obtained from patients with a positive hospitalisation may reach 13%.[1] faecal result. L39J Simultaneous estimation of infec• Some data for the mortality rate resulting from tion rate, cure rate, and detectability of G. LambLia invasive disease in South Africa are provided in infection was carried out in Kenyan children by use section 1.1.1. In Mexico, the mortality rate for pa• of a new statistical model.[40] In a study in central tients with hepatic amoebiasis before 1970 varied Arkansas, the authors examined records from 3 between 9% and 12.8%.[6] After the introduction clinical laboratories over a 7-year period and de• of the 5-nitroimidazoles, this dropped to 3.8% in tected a circannual rhythm.[41] This was consid• the major centres. Elsewhere, the mortality rate re• ered to be ' ... important for the prevention, diagno• mains high in fulminating colitis (72%), amoebic sis and treatment of this infectious order'. Such a appendicitis (20%), and amoeboma (6.4%). cycle has also been described by workers in Mex• ico.l42] A further factor to be taken into account is 2. Giardia lamblia Infection the change in geographical patterns of infection. A In common with E. histoLytica, G. Lamblia is study in the USA by Kappus et aI.l43] demonstrated very widely distributed in both tropical and tem• 'changes in rates of identification and in geo• perate countries. It is, however, not an invasive graphic patterns compared with state laboratory protozoan; although it accounts for significant data collected a decade earlier'. morbidity it is only very rarely (if ever) associated Certain human populations are at especial risk with mortality. All aspects of the infection have for G. lambLia infection. Prevalence rates of 2: 20% recently been reviewed)37] have been recorded in Bangladesh (21 to 33%), Giardia spp. have widespread distribution in the Guatemala (20%), Thailand (21 %), Seychelles animal kingdom. The organisms have remarkable (43%), India (20%), Egypt (35%), and Zimbabwe morphological similarities (this applies especially (22% ).[44] Other risk factors are age (5- to lO-year• to the cysts) and controversy exists concerning the old children are especially at risk), male sex, number and identity of Giardia spp. in different impaired nutritional status, gastric hypoacidity,[45] species.l38] It is therefore difficult to know which an immunodeficiency disorder (e.g. hypo gamma• species contribute to human disease via water con• globulinaemia), lack of breast-feeding in infancy, tamination. Giardia spp. isolated from beavers and dietary factors, travel to an endemic area,[46] preg• calves are, for example, indistinguishable by light nancy, high carbohydrate intake, living in an urban microscopy from the organism causing human dis• environment, exposure to cool and wet weather, ease. Early differentiation was based on the mor• high population density, and a poor socioeconomic phological appearance of the 'median body', an environment.[44] Although G. Lamblia infection is organelle shown by ultrastructural studies to be more common in male homosexuals, there is no composed of microtubules. However, from more satisfactory evidence that this protozoan parasite recent observations based on molecular biology, it constitutes an 'opportunistic' agent in AIDS. is clear that the entire classification of the genus Giardiasis is by no means uncommon in Aus• Giardia requires urgent revision. tralia and New Zealand. In the former country, prevalence of G. Lamblia was documented in an ab• 2.1 Prevalence of G. lamblia Infection original community at Kimberley, western Aus• As with E. histoLytica, the reported prevalence tralia,L47] where 32.1 % of children and 12.5% of rate is largely dependent upon methodology. In one adults were infected. Incidentally, a high preva• study in India, G. lamblia was detected in the first lence rate in cats and dogs was documented in this

community. In Canterbury, New Zealand, an over• The infectious stage of Giardia spp. is the cyst, all attack rate of 4.0 per 10000 population per year which measures 10Ilm x 61lm. After excretion in has been documented.l48 ] Major mechanisms of faeces, the cyst is immediately infectious to a new infection were probably the same as those pre• host, no period of maturation or latent period being viously identified overseas. At Denver, Colorado, necessary. Furthermore, the organism cannot mul• a 16% prevalence in the toddler age group was tiply outside the host, and an intermediate one is recorded.l49] Risk factors for all children in the not required for transmission. An infected individ• sample included travel to the Colorado mountains, ual excretes up to 4 x 108 cysts daily, and this can large family size, and attending a daycare centre. continue for several months[67] although excretion In a rural village in Kenya, a G. lamblia• is intermittent. Cysts can survive for several weeks infected group was delineated in children aged 19 in cool water, but not after chlorination, drying, or to 24 months.l50] At Niger, a 28.5% prevalence has heating. Filtration is recommended for the removal been documented.l51 ] In institutionalised children of cysts from water collected from surface sources (l to 61 months old) at a Thai orphanage, 20% were such as lakes and rivers. Volunteers given as few infected with G. lamblia.[52] Prevalence has also as 10 cysts have been shown to be infected.[671 been documented for a residential housing estate The biology of Giardia spp. (fig. 3), the variety in Malaysia.l53] At Cairo, Egypt, infants and chil• of possible modes of transmission, and the vari• dren showed a progressively elevated antibody ti• ability of human host responses are all involved in tre, which reached the adult level at 16 years.l54] the complex epidemiology of this infection. Some Numerous studies have documented high rates of the factors responsible for its ability to remain of infection in children, especially preschool endemic in most countries of the world are as fol• groups[55-59] and those at daycare centres in many lows: a) the infection period of G. lamblia is rela• parts of the world.l60-64] High rates have also been tively long, typically lasting for months; b) a very documented at a nursing home,[65] and in lactating small number of cysts (<10 cysts) can establish an women in Egypt.l66] infection; c) cysts can survive in the environment for several weeks, assuming that conditions are ap• 2.2 Epidemiological Aspects of propriate; and d) infections do not render protec• G. lamblia Infection tive immunity, at least not in all individuals.[67] G. lamblia infection is probably the most wide• The latent period (i.e. that interval between the spread intestinal parasitic infection to involve hu• moment of infection and the beginning of infec• mans worldwide,l67] and it is by no means con• tiousness) is equivalent to the prepatent period (i.e. fined to developing countries. The fact that the time of infection and appearance of the infec• giardiasis persists in a technologically advanced tive stage in a faecal sample). The incubation pe• country in the face of measures to prevent trans• riod for this infection is difficult to determine be• mission 'suggests that there are important features cause many cases are asymptomatic; the precise (involving) its biology and epidemiology which reasons for this are largely unclear. However, most contribute to its transmission in human popula• studies indicate that the incubation period is usu• tions living in very different circumstances' .l67J ally 1 to 2 weeks after exposure to infection. It is Since 1965, over 100 outbreaks of waterborne important to draw a distinction between infection G. lamblia infection have occurred in the USA.l38] with G. lamblia, and being diseased as a result of Up to 21 million individuals in that country may this infection.[67] Adequacy of case definition is be at risk because their drinking water is derived essential in epidemiological studies.l68] from unfiltered supplies. One report indicates that Patterns of G. lamblia infection have been over 80% of surface water samples from 66 sites summarised by Hall.[67J In developed countries, in northern America contain G. Lamblia cysts. whilst endemic in institutions, epidemics also oc-

Fig.3. (top) Giardia trophozoite from culture. The ventral surface is characterised by a large disc at the anterior end. Scanning electron micrograph x 6000. (below) Giardia trophozoite undergoing division x 6000. (From Zaman[1061).

cur, and these are usually associated with a con• creational, environmental, and social aspects of taminated water, or possibly food supply, In devel• waterborne infection should also be considered. oping countries, infection is usually acquired in Details of waterborne outbreaks have been well early childhood, Reinfection is possible immedi• documented in the USA, Canada, Australia, New ately after eradication of infection. Season ability Zealand, Scotland, and Sweden.l7S-78] Whereas of infection is another factor that must always be most outbreaks have been traced to sewage con• taken into account.[69] tamination of the water supply, the importance of Owing to intermittent excretion of cysts, beaver and muskrat faeces remains a possibility in G. lamblia infection is underdiagnosed, especially some. In addition to health aspects, other factors in developing countries, where faecal samples are including economic costs, lost tourist revenues, usually used for diagnosis. A capture enzyme im• lowered workforce productivity, and maintenance munoassay (EIA) has been shown to provide an costs should also be considered.[7S] efficient means of processing large numbers of Giardia spp. cysts have been shown to remain samples for prompt and accurate assessment of a viable in natural, tap, and distilled water for peri• 70 Giardia spp. epidemic.l ] It also facilitates rapid ods of up to several weeks.f7S] Cold temperatures tracing of epidemic sources. Levels of parasite• enhance survival, but other water-associated factors specific IgA seem to accurately reflect G. lamblia seem relatively unimportant. In unchlorinated tap infection in Colorado and Thailand)7I] water, the maximum survival time for G. lamblia has been shown to be 77 days at SoC; the maximum 2.3 Transmission of G. lamblia survival time in natural water for G. muris is 56 to Faecal transmission of G. Lamblia from person 84 days at 3.3°C. to person is a major problem, although the scale The methodology for detecting Giardia spp. remains unclear.l38,67] It seems probable that in cysts has been reviewed.f75 ] At Dunedin, New Zea• a developing country setting, where hygiene land, the incidence rate ratio for a population re• is poor and sanitary conditions inadequate, this ceiving unfiltered (micros trained) water relative to constitutes the major mode of transmission. While one using sand-filtered water was 3.3 (90%, CI = much evidence indicates that a contaminated water 1.1, 10.1), and in a parallel cross-control study of supply is the most common vehicle for a hu• incident cases, the ratio for Giardia spp. infection man G. Lamblia infection,l72,73] other sources and unfiltered (microstrained) water supply was should also be considered. Food contamination is 1.8 (90%, CI = 0.5, 6.9).l79] Not all studies, how• undoubtedly important in some outbreaks. Vene• ever, have shown an association between endemic real spread is a further possibility. The importance Giardia spp. infection and a drinking water of an animal reservoir remains unclear. With re• source.l80] A questionnaire used to assess risk fac• gard to domestic animals, the dog has received tors in the greater Vancouver water district showed most attention, and although 'man's best friend' that the significant ones were a <6-year-old child can certainly be infected,l74] the role of this reser• in the household, and travel (both within British voir in the context of human disease remains to be Colombia and internationally). Interestingly, shal• elucidated. low wells have been implicated in G. lamblia in• fection in New Hampshire, USAl8I 1 2.3. 1 Contaminated Water Supply - A Vehicle for G. lamblia Infection The CDC [CUS) Centers for Disease Control The importance of a contaminated water supply (and Prevention)] and US Environmental Protec• in the transmission of G. Lamblia infection seems tion Agency recently undertook a collaborative clear, although it is difficult to balance the fre• surveillance programme on the occurrence of wa• quency ofthis against person to person (faecal-oral terborne diseases.f82] Over the previous 5 years, and venereal) spread,PS] Economic, regulatory, re- the number of outbreaks had not changed substan-

tially, although Escherichia coli 0157:H7 and of the parasite is stimulated by bile. It should be Cryptosporidium spp. were being reported more assumed, on the basis of available evidence, that frequently and from new settings. any Giardia spp. detected in water (from whatever Few data are available on the prevalence of G. species they are derived) are potentially patho• lamblia cysts in raw and treated sewage. The re• genic to humans. moval efficiency of various treatment systems ap• Cross-species transmission of Giardia spp. iso• pears to be very high, and the viability of cysts may lates from man has been demonstrated in experi• increase as they pass through the treatment system. mental hosts, e.g. the Mongolian gerbil, the neona• Recent advances in the understanding of the tal mouse, and the neonatal rat.[38] This does not, transmission dynamics of this and other parasitic however, constitute evidence for zoonotic trans• infections of the intestinal tract have been re• mission of the infection. viewed.f831 Potential natural hosts and domestic animals[75] occupy ecological environments in wetlands en• 2.3.2 Food Contamination: A Source of G. lamblia Infection abling them, via their aquatic habits, to directly Strict hygiene is clearly an important factor in contaminate (via faecal pollution) watersheds used preventing G. lamblia infection. In a recent out• for human drinking water. In beavers, a prevalence break involving a family party of 25 individu• of infection of 7 to 16% has been recorded in parts als,[841 9 who had eaten fruit salad became ill com• of the USA and, in muskrats, a 95% rate has been pared with one who had not. The preparer of the documented.f38] Both of these species can be in• salad had a diapered child and kept a pet rabbit at fected with human G. lamblia isolates, although home; both were infected with G. lamblia. A fur• the usual animal strains appear phenotypically dis• ther outbreak highlighted a similar problem in a tinct. A study carried out in the northern states of commercial setting.f85 ] A total of 18 laboratory Canada and in Minnesota showed a regional vari• confirmed and 9 suspected cases were recorded in ation in the prevalence of Giardia spp. in beavers, insurance company employees, and the infection but not in muskratsJ871 Other animals known to was traced (using case controls) to a food handler harbour G. lamblia are wading birds, such as blue infected with G. lamblia. In another outbreak, 27 herons, green herons, black-crowned herons, and (75%) of 36 training centre staff who had eaten at egrets. Erlandsen[38] and her coworkers have re• a restaurant became ill compared with 1 (3%) of 31 cently demonstrated cross-species transmission who had not done so. Although no specific food between avian and mammalian Giardia spp. The item could be incriminated, circumstantial evi• possibility that at least some of these species may dence strongly suggested that ice contamination by act as sentinel animals, being infected by human a food handler might have been the source of in• cysts discharged into their environment, should fection. Another group of workers has addressed also be considered. the appropriateness of providing mass chemother• Despite this evidence, the possibility of the bea• apy to food handlers)86] ver,[75] or indeed of any other animal, being signif• icantly involved in water-borne human infection 2.3.3 Is a Zoonotic ReseNoir Important in Human Giardiasis? remains circumstantial. Possible zoonotic sources of infection can be Dogs and cats have also been claimed to be im• divided into natural (e.g. beaver, muskrat, vole, portant in human disease transmission.[38] In one and other wild animals) and domestic ones (e.g. study involving a Western Australian aborigine dog, cat, gerbil, neonatal mouse, and rat). Giardia community, G. lamblia was detected in 16.5% of spp. colonise and proliferate within the gastroin• 182 dogs investigated, but in none of 33 catsJ39] testinal mucosa of a wide variety of different ani• However, the importance of G. lamblia infection mal species)38,44] Binary fission (and encystment) in dogs in the context of human disease remains in

doubt)44,88-90] Other domestic animals, including 2.3.4 Intewenfion Programmes andG. lamblia sheep, cattle, goats, pigs, and birds have also been The paucity of precise information on the biol• implicated.144] In one study, the prevalence of in• ogy and epidemiology of Giardia spp. has hin• fection was 17.7% in sheep and 10.4% in cattle, dered the development of effective control mea• 44J and it was significantly higher in lambs (35.6%) sures.l However, the following factors should be and calves (27.7%))91] The evidence suggested considered in any viable programme: a) identifica• that these domestic animals might form a reservoir tion of reservoirs; b) diagnosis and treatment of for human infection and vice versa. infected cases; c) prevention of environmental fae• Erlandsen[38) has outlined future directions that cal contamination; d) provision of safe excreta dis• could be taken in an attempt to unravel the 'zoo• posal and a safe water supply; e) health education notic dilemma'. Precise data are required on the aimed at improved personal and health hygiene; and f) improved host resistance, with attention to species of Giardia recovered from a given host, the nutrition and possible immunisation. actual number of G. lamblia species in humans, As discussed above, G. lamblia infection is a and the organisms involved in water-borne out• major practical problem in daycare centres, al• breaks. Delineation of genetic diversity, possible though the efficacy of control strategies has not taxonomy, and determination of the numbers of been systematically evaluated. When a concurrent species involved will lie in several different mo• Cryptosporidium spp. infection is present, symp• lecular approaches, e.g. isoenzymic analysis, toms may be worsened. Bartlett et al.[93) have car• southern blots, and random amplified polymorphic ried out a prospective, randomised, controlled trial DNA markers. One group has documented signif• in an attempt to throw light on this problem. A total icant differences in molecular structure between 3 of 31 daycare centres were involved, incorporating species: G. lamblia, G. muris, and G. ardeae.[38) 4180 child months of observation. In group A, ex• These are based on nucleotide sequencing of the clusion and treatment of symptomatic and asymp• rRNA genes, their size in kilobases, and their tomatically infected children were implemented; guanosine-cytosine content. Other work, using the in group B, exclusion and treatment of symptom• polymerase chain reaction (PCR), suggests that G. atic infections only were incorporated, and in psittaci and G. microtus are distinct from the other group C, exclusion and treatment of symptomatic 3. Epidemiologically, such techniques are of value infections and treatment of asymptomatic infec• in comparing Giardia cysts in faeces with those in tion were implemented. G. Lamblia prevalence was environmental samples. Using fluorescent in situ monitored before intervention, and at 1,2,4, and hybridisation, it has been possible to detect 6 months later. Prevalence was 8%, 12%, and 7% a) human cysts in a model system containing for groups A, Band C, respectively, at 1 month, G. lamblia, G. ardeae, and G. muris; b) multiple and 7%, 8%, and 8%, respectively, at 6 months. species of Giardia in the same sample by the use Therefore, a more strict and costly intervention of 3-colour fluorescence detection of rDNA and programme did not result in a significantly better immunological probes; and c) cysts in human level of control. faecal and environmental samples from sewage in In a study involving preschool children carried lagoons. out in Lesotho, a developing African country, the One group of investigators has concluded, in authors concluded that the 'amount of water used the light of available evidence, that ' ... although for personal and domestic hygiene may be more mammals and man do not seem to possess their important than the quality of drinking water' )94) own unique species of Giardia, in reality the major In Mexico, the possible protective effect of methods of transmission ... probably remain basi• breast-feeding against Giardia spp. infection has cally host-specific' .[92] been assessedJ95] A total of 197 infants in a poor

area of Mexico City were followed up from birth to 10 (-5), respectively, of Giardia cysts' )98] In to age 18 months, and symptoms and feeding status another study in Canada, Giardia spp. cysts were were recorded weekly. Lack of breast-feeding was detected in 17% of 83 filtered water effluents. 1991 a significant risk factor for the first G. lamblia in• Evaluation of the data indicated that '24% of the fection at all ages. However, while an absence of utilities examined would not meet a 1/10,000 an• breast-feeding was associated with a symptomatic nual risk of Giardia infection', and for cold water G.lamblia infection, breast-feeding did not protect concentrations (O.S°C), 46% of the plants would against chronic G. lamblia carriage. These authors not achieve the 1110000 risk level. FlanaganlJOOI showed that the presence of animals in the household has summarised the present position in the UK: and the use of water/nonmilk liquid for infant feed• 'the importance of potable water supplies as a ing were significant (p = 0.005 and p = 0.035, re• source of infection ... is not clear, nor is the role of spectively) risk factors for G. lamblia infection. zoonotic spread. The apparent susceptibility to in• The effectiveness of water treatment with chem• fection in certain population groups requires fur• icals, and filters for control of Giardia spp. cysts ther exploration as does the role of the asymptom• in areas where treated water is not available has atically infected in transmission'. been investigated)96] Four filters and 7 forms of High G. lamblia infection rates have been doc• chemical treatment were evaluated for both clear umented in malnourished children in many devel• and turbid water at 10'C, and contact disinfection oping countries.l44 ,101] Clearly, therefore, an im• devices were also investigated for cyst inactiva• provement in nutritional status is of paramount tion. The authors concluded that none of the de• importance and this should be coupled with health vices provided significant cyst inactivation, al• education, which is especially important in devel• though heating water to:::: 70°C for 10 minutes was oping countries where high levels of illiteracy, ig• an acceptable alternative form of treatment. A norance, and cultural taboos exist. Some evidence has been presented indicating that G. lamblia in• Giardia spp. outbreak from a chlorinated, un• fection associated with malnutrition does not re• filtered surface water supply at British Columbia spond satisfactorily to chemotherapy.[lOII (population 25000) over a 3-month period resulted in 363 confirmed cases ofinfection.l97 ] Areservoir With regard to a potential vaccine, little prog• ress has been made so far. l441 The effect of sys• containing Giardia spp.-infected beavers was in• temic oral immunisation with a 56 kDa protein of criminated as the source. Giardia spp. has been tested in experimental mice, Microscopical detection methods have been and although initially encouraging results have adapted, using various filtration systems, for detec• been obtained, further studies are required. tion of G. lamblia cysts in water supplies; when 1 to 25% of a population is infected, levels of cysts 2.4 Morbidity Associated with in raw sewage have been estimated at 9.6 x 103 to G, lamblia Infection 2.4 x 105 per litre. Orlon filters have been shown to be superior to cellulose acetate and polypropyl• A range of morbidity results from a G. lamblia ene yarn-wound filters, and epoxy-fibreglass fil• infection - from travellers' diarrhoea (a self-limit• ter-tubes. Specific monoclonal antibody has been ing clinical syndrome) to severe malabsorption ac• utilised to detect G. lamblia-antigen on the filter. companied by weight loss and, in extreme cases, If the increase in waterborne G. lamblia out• malnutrition. l1021 However, the vast majority of breaks in the US is to be brought under control, cases remain asymptomatic) 100,1 03] public health officials will 'need to work with the A surveillance programme carried out by means water industry to ensure a risk ofless than 1/10,000 of a questionnaire at Bristol, England, revealed 22 for source waters with 0.7 to 70 cysts per 100 liters patients with G. lamblia infection who had trav• through treatment achieving reduction of 10 (-3) elled abroad in the month preceding the onset of

symptoms;[104] most others were preschool chil• G. lamblia. Overall, Graceyr39] concluded that' ... dren or individuals engaged in recreational water there is still disagreement about whether diar• activities. rhoeal disease is a major, community-wide cause In individuals living in underprivileged condi• of malnutrition in under-fives in developing coun• tions, G. lamblia constitutes a significant patho• tries'. In a further study, G. lamblia-infected chil• gen. Graceyl39J has reviewed the clinical signifi• dren tended to achieve greater weight and height cance of infection in Australian aboriginal for age than those who were not infected.[I03] children. In this group, G. lamblia infection was These authors concluded that 'Healthy daycare shown to be 2 to 3 times higher than in nonaborigi• children with asymptomatic (G. Lamblia) infection nal children. Almost 30% of aboriginal children show no disadvantage and perhaps even an advan• studied in the late 1960s had evidence of infection tage in nutritional status and freedom from other in faecal samples, and in a later study at south-east illnesses'. However, Cheek et al.[102] have studied Queensland, 70% were infected. In a survey of aboriginal children at Yaluta, South Australia, and 1600 faecal samples carried out in Western Aus• concluded that G. LambLia infection was prevalent, tralia, prevalence rates of 49% at 4 years of age and and that because it 'had been shown elsewhere that 36.7% at 9 to 10 years of age were documented. G. lamblia is capable of inducing malabsorption However, in a hospital-based study of aboriginal with resulting nutritional deficiencies', it probably children in Western Australia, the isolation rate in accounted for at least part of the impairment of those with diarrhoea was similar to that in those several nutritional indices. without. 139J This study also demonstrated that many other small-intestinal pathogens are of far 2.5 Mortality Resulting from greater significance than G. lamblia. G. lamblia Infection Hall[67J has summarised the possible nutritional consequences of a G. lamblia infection. Data ob• Mortality associated with G. lambLia appears to tained from case studies (e.g. 'failure to thrive'), be exceedingly rare. Such an occurrence would cross-sectional studies, and prospective studies presumably be related to malnutrition consequent were assembled. The significant nutritional conse• upon a heavy small-intestinal infection with resul• quences of such an infection are anorexia and mal• tant malabsorption. One death attributed to acute jejunal ulceration has been documented. L105 ] absorption, the pathogenesis of which remains un• clear, despite much research. However, as Halll671 has pointed out, most studies concerning the nutri• References tional impact of G. LambLia infection contain seri• I. Walsh JA. Prevalence of Entamoeba histolytica infcction. In: Ravdin n, editor. Amebiasis: human infection by Entamoeba ous flaws or limitations. Much of the evidence de• histolytica. New York: Churchill Livingstone, 1988; 93-105 pends on clinical improvement after chemotherapy 2. Shetty N, Narasimha M, Elliott E, et al. Age-specific sero-prev• alence of amoebiasis and giardiasis in southern Indian infants with a nitroimidazole compound. However, this and children. J Trop Pediatr 1992; 38: 57-63 agent also eliminates many concurrent organisms, 3. Arroyave RJ, Ayala DE, Hermida RC. Differences in circannual characteristics of the incidences of amebiasis and giardiasis. including some bacteria. In addition, it must be Prog Clin Bioi Res 1990; 341B: 717-27 remembered that improvement in a child's appe• 4. Goncalves JF, Tanabe M. Medeiros F de P, et al. Parasitological tite, digestion, or absorption can only lead to im• and serological studies on amoebiasis and other intestinal par• asitic infections in the rural sector around Recife, northeast proved growth if sufficient protein and energy are Brazil. Rev Inst Med Trop Sao Paulo 1990; 32: 428-35 available. 5. Chacin-Bonilla L, Bonilla E. Parra AM, et al. Prevalence of Entamoeba histolytica and other intestinal parasites in a com• Although failure to thrive, associated with in• munity from Maracaibo. Venezuela. Ann Trop Med Parasitol fectious diarrhoea, is a common problem in aborig• 1992; 86: 373-80 inal children at Kimberley, Western Australia, 6. Jalan KN, Maitra TK. Amebiasis in the developing world. In: Ravdin JI, editor. Amebiasis: human infection by Entamoeba few reports satisfactorily record the prevalence of histolytica. New York: Churchill Livingstone, 1988: 535-55

7. Candler W, Phuphaisan S, Echeverria P, et al. Amebiasis at an 26. Arellano J, Grandos J, Frenk P, et al. Increased frequency of evacuation site on the Thai-Cambodian border. Southeast As• HLA-DR3 in Mexican mestizo pediatric patients with amebic ian J Trop Med Public Health 1990; 21: 574-9 liver abscess (ALA). Arch Med Res 1992; 23: 269-70 8. Ali-Shtayeh MS, Hamdan AH, Shaheen SF, et al. Prevalence 27. Choudhuri S, Prakash V, Kumar A, et al. Protective immunity and seasonal fluctuations of intestinal parasitic infections in to Entamoeba histolytica infection in subjects with antiamoe• the Nablus area, West Bank of Jordan. Ann Trop Med bic antibodies residing in a hyperendemic zone. Scand J Infect Parasitol1989; 83: 67-72 Dis 1991; 23: 771-6 9. Sargeaunt PG. A survey of Entamoeba histolytiea and En• 28. Baveja UK, Francis S, Kaur M, et al. The zymodemes of En• tamoeba dispar (Brumpt) infections on Mahe, the Seychelles. tamoeba histolytica in New Delhi, India. J Diarrhoeal Dis Res Arch Med Res 1992; 23: 265-7 1990; 8: 27-30 10. Ravdin n. Amebiasis: human infection by Entamoeba 29. Martincz-Garcia MC, Guticrrcz-Trujillo G, Sanchez-Pares ME, histolytiea. New York: Churchill Livingstone, 1988 et al. Efficacy of zymodemes of E. histoly/ica technique in an 11. Weinmann AI. Spelman DW, Spicer WJ. Indigenous invasive epidemiological study and report of new zymodemes in Mex• amoebiasis in Australia. Aust N Z J Surg 1992; 62: 235-7 ico. Arch Invest Med (Mexico) 1990; 21 Suppl. I: 203-8 12. Pearson RD, Hewlett EL. Amebiasis in travelers. In: Ravdin n, 30. Martinez-Garcia MC, Munoz 0, Garduna-Rodriguez G, et al. editor. Amebiasis: human infection by Entamoeba histolytica. Pathogenic and non-pathogenic zymodemes of Entamoeba New York: Churchill Livingstone, 1988: 556-62 histolytica in a rural area of Mexico. Concordance with serol• 13. Druckman DA, Quinn TC. Entamoeba histolytica infections in ogy. Arch Invest Med (Mexico) 1990; 21 Suppl. 1: 147-52 homosexual men. In: Ravdin n, editor. Amebiasis: human 31. Walsh JA. Transmission of Entamoeba histolYlica infection. In: infection by Entamoeba histolytica. New York: Churchill Liv• Ravdin JI, editor. Amebiasis: human infection by Entamoeba histolytica. New York: Churchill Livingstone, 1988: 106-19 ingstone, 1988: 563-75 14. Takeuchi T, Okuzawa E, Nozaki T, et al. High seropositivity of 32. Wolfe MS. Amebiasis. In: Strickland GT, editor. Hunter's trop• ical medicine, 7th ed. Philadelphia: WB Saunders, 1988: 550- Japanese homosexual men for amebic infection. J Infect Dis 65 1989; 159: 808 33. Jackson TF, Sargeaunt PG, Visser PB, et al. Entamoeba 15. Petri WA, Ravdin n. Amebiasis in institutionalized popUlations. histolytica: naturally occurring infections in baboons. Arch In: Ravdin n, editor. Amebiasis: human infection by En• Invest Med (Mexico) 1990; 21 Suppl. 1: 153-6 tamoeba hisloly/ica. New York: Churchill Livingstone, 1988; 34. Anand BS, Tuteja AK, Kaur M, et al. En/amoeba hislolytica 576-81 cyst passers. Clinical profile and spontaoeous eradication of 16. Weinke T, Friedrich-Janicke B, Hopp P, et al. Prevalence and infection. Dig Dis Sci 1993; 38: 1825-30 clinical importance of Entamoeba his/olyliea in two high-risk 35. Ruiz-Palacios GM, Castanon B, Bojalil R, et al. Low risk of groups: travelers returning from the tropics and male homo• invasive amebiasis in cyst carriers. A longitudinal molecular sexuals. JInfect Dis 1990: 161: 1029-31 seroepidemiological study. Arch Med Res 1992; 23: 289-91 17. de Lalla F, Rinaldi E, Santaro D, et al. Outbreak of Entamoeba 36. Cook Gc. Gastroenterological emergencies in the tropics. In: histo(vtica and Giardia Lamblia infections in travellers return• Baillieres clinical gastroenterology. Vol. 5. London: Bailliere ing from the tropics. Infection 1992: 20: 78-82 Tindall, 1991: 861-86 18. Jedlicka J, Tolarova V, Svandova E. Intestinal parasitoses in 37. Thompson RCA, Reynoldson JA, Lymbery AJ, editors. Giar• Czechoslovak citizens working abroad. J Hyg Epidemiol dia: from molecules to disease. Wallingford: CAB Interna• Microbiol Immunol (Praha) 1990; 34: 63-8 tional, 1994: 394 19. Jessurun J, Barron-Rodriguez LP, Fernaodez-Tinoco G, et al. 38. Erlandsen SL. Biotic transmission - is giardiasis a zoonosis? In: The prevalence of invasive amebiasis is not increased in pa• Thompson RCA, Reynoldson JA, Lymbery AJ, editors. Giar• tients with AIDS. AIDS 1992; 6: 307-9 dia: from molecules to disease. Wallingford: CAB Interna• 20. Sorvillo FJ, Lieb L, Mascola L, et al. Declining rates of amebi• tional, 1994: 83-97 asis in Los Angeles County: a sentinel for decreasing acquired 39. Gracey M. The clinical significance of giardiasis in Australian immunodeficiency syndrome (AIDS) incidence? Am J Public Aboriginal children. In: Thompson RCA, Reynoldson JA, Health 1989; 79: 1563-4 Lymbery AJ, editors. Giardia: from molecules to disease. 21. Omar MB, al-Awad ME, ai-Madani AA. Giardiasis and amoe• Wallingford: CAB International, 1994: 281-91 biasis infections in three Saudi closed communities. J Trap 40. Nagelkerke NJ, Chunge RN, Kinoti SN. Estimation of parasitic Med Hyg 1991; 94: 57-60 infection dynamics when detectability is imperfect. Stat Med 22. Nagakura K, Tachibana H, Tanaka T, et al. An outbreak of am• 1990; 9: 1211-9 ebiasis in an institution for the mentally retarded in Japan. Jap 41. Pasley IN, Daly JJ, McCullough D, et al. Circannual incidence J Med Sci Bioi 1989; 42: 63-76 of Giardia lamblia. Chronobiol Int (Oxford) 1989; 6: 185-9 23. Yadav SK, Jain AK, Srivastava VK, et al. Comparison of stool 42. Hermida RC, Ayala DE, Arroyave RJ. Circannaul incidence of microscopy and serology (enzyme linked immunosorbent as• Giardia lamblia in Mexico. Chronobiol Int (Oxford) 1990; 7: say) in epidemiology of amebiasis. Indian J Gastroenterol 329-40 1990; 9: 25-6 43. Kappus KK, Juranek DD, Roberts JM. Results of testing for 24. Shetty N, Nagpal S, Rao PY, et al. Detection of IgG, IgA, IgM intestinal parasites by state diagnostic laboratories, United and IgE antibodies in invasive amoebiasis in endemic arcas. States, 1987. MMWR CDC Surveill Sunun 1991; 40: 25-45 Scand J Infect Dis 1990; 22: 485-91 44. Rabbani GH, Islam A. Giardiasis in humans: populations most 25. Fuchs G, Ruiz-Palacios G, Pickering LK. Amebiasis in the pe• at risk and prospects for control. In: Thompson RCA. diatric population. In Ravdin JI, editor. Amebiasis: human Reynoldson JA, Lymbery AJ, editors. Giardia: from mole• infection by Entamoeba histoly/ica. New York: Churchill Liv• cules to disease. Wallingford: CAB International, 1994: 217- ingstonc, 1988: 594-613 49

45. Doglioni C, De-Boni M, Cielo R, et a!. Gastric giardiasis. J Clin 64. Steketee RW, Reid S, Cheng T, et a!. Recurrent outbreaks of Pathol 1992; 45: 964-67 giardiasis in a child day care center, Wisconsin. Am J Public 46. Peppiatt R, Byass P. A survey of the health of British mission• Health 1989; 79: 485-90 aries. Br J Gen Pract 1991; 41: 159-62 65. White KE, Hedberg CW, Edmonson LM, et a!. An outbreak of 47. Meloni BP, Thompson RC, Hopkins RM, et a!. The prevalence giardiasis in a nursing home with evidence for multiple modes of Giardia and other intestinal parasites in children, dogs and of transmission. J Infect Dis 1989; 160: 298-304 cats from aboriginal communities in the Kimberley. Med J 66. Azab ME, Abdel-Fattah SM, Makled KM, et a!. Prevalence of Aust 1993; 158: 157-9 Giardia lamblia antibodies in serum and milk in lactating 48. Mitchell P, Graham P, Brieseman MA. Giardiasis in Canter• women from different social classes in Egypt. J Egypt Soc bury: the first nine months reported cases. N Z Med J 1993: Parasitol1991; 21: 611-9 106: 350-2 67. Hall A. Giardia infections: epidemiology and nutritional con• 49. Novotny TE, Hopkins RS, Shillam P, el a!. Prevalence of Giar• sequences. In: Thompson RCA, ReynoldsonJA, Lymbery AJ, dia lamblia and risk factors for infection among children at• editors. Giardia: from molecules to disease. Wallingford: tending day-care facilities in Denver. Public Health Rep CAB International, 1994: 251-80 1990; 105: 72-5 68. Hopkins RS, Juranek DO. Acute giardiasis: an improved clini• 50. Chunge RN, Karumba PN, Kaleli N, et a!. Prevalence and fre• cal case definition for epidemiologic studies. Am J Epidemiol quency of Giardia lamblia in children aged 0 to 60 months 1991; 133: 402-7 with and without diarrhoea. East Afr Med J 1992; 69: 311-3 69. Addiss DG, Davis JP, Roberts JM, et a!. Epidemiology of 51. Develoux M, Alarou A, Mouchet F. High prevalence of giardia• giardiasis in Wisconsin: increasing incidence of reported sis in an urban population in Niger. J Trop Med Hyg 1990; cases and unexplained seasonal trends. Am J Trop Med Hyg 93: 355-6 1992; 47: 13-9 52. Janoff EN, Mead PS, Mead JR, et a!. Endemic Cryptosporidium 70. Green E, Warhurst 0, Williams J, et a!. Application of a capture and Giardia lamblia infections in a Thai orphanage. Am J enzyme immunoassay in an outbreak of waterborne giardiasis Trop Med Hyg 1990; 43: 248-56 in the United Kingdom. Eur J Clin Microbiol Infect Dis 1990; 53. Rahman WA. Prevalence of Giardia in dogs in Malaysia: survey 9: 424-8 ofa residential housing estate. Trans R Soc ofTrop Med Hyg 71. Janoff EN, Taylor ON, Echeverria P, et a!. Serum antibodies to 1990; 84: 805 Giardia lambha by age in populations in Colorado and Thai• 54. Abdel-Fattah SM, Maklad KA, Gadallah MA. Age-related rate land. West J Med 1990; 152: 253-6 of seropositivity of antibody to Giardia lamblia in different 72. Birkhead G, Janoff EN, Vogt RL, et a!. Elevated levels of im• age groups in Cairo. J Egypt Soc Parasitol1991; 21: 707-13 munoglobulin A to Giardia lambha during a waterborne out• 55. Ahmed MM, Bolbol AH. The intestinal parasitic infections break of gastroenteritis. J Clin Microbiol 1989; 27: 1707-10 among children in Riyadh, Saudi Arabia. J Egypt Soc 73. Birkhead G, Vogt RL. Epidemiological surveillance for en• Parasitol1989; 19: 583-8 demic Giardia lamblia infection in Vermont. The roles of 56. Bolbol AS, Mostafa SO, al-Sekait M, et a!. Pattern of intestinal walerborne and person-lo-person transmission. Am J parasitic infection in preschool children in Riyadh, Saudi Ara• Epidemiol 1989; 129: 762-8 bia. J Hyg Epidemiol Microbiol Imrnunol (Praha) 1989; 33: 74. Cook Gc. Canine-associated zoonoses: an unacceptable hazard 253-9 to human health. Q J Med 1989; 70: 5-26 57. Chunge RN, Karumba PN, Nagelkerke N, et a!. Intestinal par• 75. Wallis PM. Abiotic trammission - is water really significant? asites in a rural community in Kenya: cross-sectional surveys In: Thompson RCA, Reynoldson JA, Lymbery AJ, editors. with emphasis on prevalence, incidence, duration of infec• Giardia: from molecules to disease. Wallingford: CAB Inter• tion, and polyparasitism. East Afr Med J 1991; 68: 112-23 national, 1994: 99-122 58. Chunge RN, Nagelkerke N, Karumba PN, et a!. Longitudinal 76. Herwaldt BL, Craun GF, Stokes SL, et a!. Waterborne-disease study of young children in Kenya: intestinal parasitic infec• outbreaks, 1989-1990. MMWR CDC Surveill Summ 1991; tion with special reference to Giardia lamblia, its prevalence, 40: 1-21 incidence and duration, and its association with diarrhoea and 77. Isaac-Renton JL, Philion JJ. Factors associated with acquiring with other parasites. Acta Trop (Basel) 1991; 50: 39-49 giardiasis in British Columbia residents. Can J Public Health 59. Kasuya S, Khamboonruang C, Amano K, et a!. Intestinal para• 1992; 83: 155-8 sitic infections among schoolchildren in Chiang Mai, north• 78. Roach PO, Olson ME, Whitley G, et a!. Waterborne Giardia ern Thailand: an analysis of the present situation. J Trop Med cysts and Cryptosporidium oocysts in the Yukon, Canada. Hyg 1989; 92: 360-4 Appl Environ Microbiol1993; 59: 67-73 60. Addiss DG, Stewart JM, Finton RJ, et a!. Giardia lamblia and 79. Fraser GG, Cooke KR. Endemic giardiasis and municipal water Cryptosporidium infections in child day-care centers in Ful• supply. Am J Public Health 1991; 81: 760-2 ton County, Georgia. Pediatr Infect Dis J (Baltimore) 1991; 80. Mathias RG, Riben PO, Osei WD. Lack of an association be• 10: 907-11 tween endemic giardiasis and a drinking water source. Can J 61. Goldin AJ, Apt W, Aguilera X, et a!. Efficient diagnosis of Public Health 1992; 83: 382-4 giardiasis among nursery and primary school children in San• 81. Dennis DT, Smith RP, Welch JJ, et a!. Endemic giardiasis in tiago, Chile by capture ELISA for the detection of fecal Giar• New Hampshire: a case- control study of environmental risks. dia antigens. Am J Trop Med Hyg 1990; 42: 538-45 J Infect Dis 1993; 167: 1391-5 62. Rauch AM, Van R, Bartlett AV, et a!. Longitudinal study of 82. Moore AC, Herwaldt BL, Craun GF, et a!. Surveillance for Giardia lamblia in a day care center population. Pediatr Infect waterborne disease outbreaks - United States, 1991-1992. Dis J 1990; 9: 186-9 MMWR CDC Surveill Surnm 1993; 42: 1-22 63. Shandera W. From Leningrad to the day-care center. The ubiq• 83. Eckert J. New aspects of parasitic zoonoses. Vet Parasilol uitous Giardia lamblia. WestJ Med 1990; 153: 154-9 (Amcrstcrdam) 1989; 32: 37-55

84. Porter J 0, Gaffney C, Heymann 0, et al. Food-borne outbreak 96. Ongerth JE, Johnson RL, Macdonald SC, et al. Back-country of Giardia lamblia. Am J Public Health 1990; 80: 1259-60 water treatment to prevent giardiasis. Am J Public Health 85. Mintz ED, Hudson-Wragg M, Mshar P, et al. Foodborne giardia• 1989; 79: 1633-7 sis in a corporate setting. J Infect Dis 1993; 167: 250-3 97. Moorehead WP, Guasparini R, Donovan CA, et a!. Giardiasis 86. Sanchez JL, Rios C, Hernandez-Fragoso I, et al. Parasitological outbreak from a chlorinated community water supply. Can J evaluation of a foodhandler population cohort in Panama: risk Public Health 1990: 81: 358-62 factors for intestinal parasitism. Mil Med 1990; 155: 250-5 98. Rose JB, Haas CN, Regli S. Risk assessment and control of 87. Erlandsen SL, Sherlock LA, Bemrick WJ, et al. Prevalence of waterborne giardiasis. Am J Public Health 1991; 81: 709-13 Giardia spp. in beaver and muskrat populations in north• 99. Le Chevallier MW, Norton WD, Lee RG. Giardia and eastern states and Minnesota: detection of intestinal tropho• Cryptosporidium spp., in filtered drinking water supplies. zoites at necropsy provides greater sensitivity than detection Appl Environ Microbiol 1991; 57: 2617-21 of cysts in fecal samples. Appl Environ Microbial 1990; 56: 100. Flanagan PA. Giardia - diagnosis, clinical course and epidemi• 31-6 ology. A review. Epidemiollnfect 1992; 109: 1-22 88. Castor SB, Lindqvist KB. Canine giardiasis in Sweden: no ev• 10 1. Sullivan PB, Marsh MN, Phillips MB, et a!. Prevalence and idence of infectivity to man. Trans R Soc Trop Med Hyg 1990; treatment of giardiasis in chronic diarrhoea and malnutrition. 84: 249-50 Arch Dis Child 1991: 66: 304-6 89. Collyer R, Lim KH, Tang R, et a!. Suburban dogs - a reservoir 102. Cheek DB, McIntosh GH, O'Brien V, et a!. Malnutrition in of human giardiasis? MedJ Aust 1992; 156: 814-5 aboriginal children at Yalata, South Australia. Eur J Clin Nutr 90. Sykes TJ, Fox MT. Patterns of infection with Giardia in dogs 1989; 43: 161-8 in London. Trans R Soc Trop Med Hyg 1989; 83: 239-40 103. Ish-Horowicz M, Korman SH. Shapiro M, et al. Asymptomatic 91. Buret A, Hollander N den, Wallis PM. et al. Zoonotic potential giardiasis in children. Pediatr Infect Dis J 1989; 8: 773-9 of giardiasis in domestic ruminants. J Infect Dis 1990; 162: 104. Gray SF, Rouse AR. Giardiasis - a cause of travellers' diar• 231-7 rhoea. Communicable Disease Report CDR Review 1992; 2: 92. Kasprzak W, Pawlowski Z. Zoonotic aspects of giardiasis: a R45-47 review. Vet Parasitol1989; 32: 101-8 105. Cook Gc. Tropical gastroenterology. Oxford; Oxford Univer• 93. Bartlett AV. Englender SJ, Jarvis BA, et a!. Controlled trial of sity Press, 1980: 304-9 Giardia lamblia control strategies in day care centers. Am J 106. Zaman V. Atlas of medical parasitology. 2nd ed. Sydney: Adis Public Health 1991; 81: 1001-6 Health Science Press, 1984 94. Esrey SA, Collett J, Miliotis MD, et al. The risk of infection from Giardia lamblia due to drinking water supply, use of water, and latrines among preschool children in rural Lesotho. Int J Epidemiol 1989; 18: 248-53 95. Morrow AL, Reves RR, West MS, et a!. Protection against in• Correspondence and reprints: Dr C.c. Cook, Hospital for fection with Giardia lamblia by breast-feeding in a cohort of Tropical Diseases, 4 SI Pancras Way, London NWI OPE, Mexican infants. J Pediatr 1992; 121: 363-70 England.