|Home | About | Journals | Submit | Contact Us | Français|
Summary: Molecular diagnostic tools have been used recently in assessing the taxonomy, zoonotic potential, and transmission of Giardia species and giardiasis in humans and animals. The results of these studies have firmly established giardiasis as a zoonotic disease, although host adaptation at the genotype and subtype levels has reduced the likelihood of zoonotic transmission. These studies have also identified variations in the distribution of Giardia duodenalis genotypes among geographic areas and between domestic and wild ruminants and differences in clinical manifestations and outbreak potentials of assemblages A and B. Nevertheless, our efforts in characterizing the molecular epidemiology of giardiasis and the roles of various animals in the transmission of human giardiasis are compromised by the lack of case-control and longitudinal cohort studies and the sampling and testing of humans and animals living in the same community, the frequent occurrence of infections with mixed genotypes and subtypes, and the apparent heterozygosity at some genetic loci for some G. duodenalis genotypes. With the increased usage of multilocus genotyping tools, the development of next-generation subtyping tools, the integration of molecular analysis in epidemiological studies, and an improved understanding of the population genetics of G. duodenalis in humans and animals, we should soon have a better appreciation of the molecular epidemiology of giardiasis, the disease burden of zoonotic transmission, the taxonomy status and virulences of various G. duodenalis genotypes, and the ecology of environmental contamination.
Giardia sp. is one of the most common intestinal parasites of humans; about 200 million people in Asia, Africa, and Latin America have symptomatic infections (301). Once infected, Giardia causes a generally self-limited clinical illness (i.e., giardiasis) characterized by diarrhea, abdominal cramps, bloating, weight loss, and malabsorption. However, asymptomatic giardiasis occurs frequently, especially in developing countries (113, 256). Giardia is also a very common enteric parasite of domestic animals, including livestock, dogs, and cats (257, 260), and wildlife (19). One species within this genus, Giardia duodenalis (syn. Giardia lamblia and Giardia intestinalis), causes giardiasis in humans and most mammals. Thus, giardiasis is considered a zoonotic disease.
The life cycle of Giardia is direct, and the infective stage of the parasite, the cyst, is encysted when released into the feces and is immediately infectious (123). Cysts remain infectious for months in cool, damp areas and rapidly accumulate in the environment. In soil, cyst infectivity was reduced by only 11% after 49 days at 4°C and was noninfective after 7 days at 25°C (75). In tap water, Giardia cysts were infectious for 56 days at 0°C to 4°C and 14 days at 20°C to 28°C. Similar results were obtained in lake water, with 56 days of survival at 0°C to 4°C or 6°C to 7°C and 28 days at 17°C to 20°C. Longer survival was noticed in river water, with 84 days of survival at 0°C to 4°C and 28 days at 20°C to 28°C. In seawater, Giardia cysts could survive for over 65 days at 4°C (75). When ingested by the host, cysts excyst in the duodenum, releasing the trophozoites. The latter undergo repeated mitotic division and form environmentally resistant cysts in response to the stimulation of bile salts and other conditions. Cysts pass through the intestine in feces and are spread by contaminated water, food, and fomites and by direct physical contact.
An important aspect of the epidemiology of giardiasis is to understand the host range of different Giardia species and strains/genotypes, the potential for cross-species transmission, and risk and environmental factors involved in the exposure of the pathogen. This is particularly important in determining the zoonotic potential of Giardia infections in domestic animals and in determining the human disease burden attributable to parasites of animal origin. It has been only recently, with the advent of molecular typing tools, that the epidemiology of Giardia is systematically addressed.
Giardiasis exerts a significant public health impact because of the high prevalence and disease burden of the infection, its propensity in causing major outbreaks and emergency responses, and its effects on growth and cognitive functions of infected children. Giardiasis is also a common disease in livestock and companion animals; thus, it is of veterinary health importance.
Giardia duodenalis has a global distribution (Tables (Tables11 and and2).2). Infection rates for giardiasis in humans are generally lower in developed countries, and data from some recent studies are listed in Table Table1.1. Recent studies reported infection rates of 4.0% in Belgium (93), 1.5% in Germany (227), 0.4% to 6.2% in Italy (58, 97), 3.7% in Portugal (10), 5.4% in Spain (167), 1.3% in the United Kingdom (61), 1.4% in the United States (52), 1.1% to 6.6% in Saudi Arabia (8, 136), 2.5% in South Korea (126), 1.6% to 7.6% in Australia (113, 213), and 7.6% in New Zealand (152). Most of the surveys were conducted with asymptomatic children. The occurrence of giardiasis is probably higher in children with diarrhea. Thus, in the Nordic countries of Denmark, Finland, Norway, and Sweden, the infection rate of giardiasis was estimated to be 2.9% and 5.8% for asymptomatic and symptomatic persons, respectively (121). A similar trend was seen in the Netherlands, with infection rates of 14.0% in patients with persistent diarrhea and 2.0% in asymptomatic subjects (117). Although Giardia is not considered an opportunistic pathogen in immunocompromised patients, the infection rates of giardiasis in HIV-infected people ranged from 3.5% to 6.2% in Italy before the introduction of highly active antiretroviral treatment (97). The infection rates for giardiasis are 3.1% in HIV/AIDS patients in Iran (60), 6.6% in immunocompromised patients in Saudi Arabia (8), and 13.9% in 252 patients with common immunodeficiency in France (193). Occasionally, high infection rates have also been reported for some poor regions in developed countries. For example, an infection rate of 42.9% was reported for a socially deprived small Rom community in Italy (168).
High infection rates for giardiasis have been reported for developing countries, and some recent data are listed in Table Table2.2. In various Asian countries (Bangladesh, Cambodia, China, India, Indonesia, Laos, Malaysia, Nepal, Philippines, Thailand, Turkey, Saudi Arabia, and Vietnam), North America (Cuba, Mexico, and Nicaragua), South America (Argentina, Brazil, Colombia, and Peru), and Africa (northern Africa, west Africa, and South Africa), most of the studies focused on children, and infection rates for G. duodenalis fell into the range of 8% to 30% in the majority of those studies (12, 29, 56, 62, 63, 65, 69, 144, 160, 166, 174, 184, 186, 202, 203, 229, 234, 237, 239, 252, 253, 281). In a few studies, the infection rate was lower than 4% or higher than 30% (65, 137, 145, 230, 243, 274, 280). High infection rates for giardiasis were also reported for adults in developing countries, with rates of 25.1% in pregnant women in Minatitlan, Mexico (224); 11.7% in adults in Settat, Morocco (72); and 5.0 to 14.0% in African refugees and new immigrants in the United States, the Netherlands, and Spain (167).
Giardiasis is highly underreported for various reasons (302). In the United States, the total number of reported cases remained about 20,000 annually from 2006 to 2008, and the incidence ranged from 7.4 to 7.6 cases per 100,000 population (302). Although the true burden of giardiasis in the United States is unknown, an estimated 2 million cases occur annually (302). In Britain, there are approximately 3,500 cases per year according to a report of Health Protection Scotland (http://www.hps.scot.nhs.uk/giz/giardia.aspx), with an incidence of giardiasis in the United Kingdom of 5.5 cases per 100,000 population in 2005. In Scotland, the annual number of reported Giardia infections remained constant at approximately 300 to 400 between 1988 and 1998 but decreased consistently thereafter; in 2003 only 192 infections were reported (205). In Germany, on average, 3,806 notified giardiasis cases (range, 3,101 to 4,626) were reported between 2001 and 2007, which corresponded to an average incidence of 4.6 cases/100,000 population (227). Much higher incidence rates were reported for some other countries. In the Netherlands, there were 11,600 cases in 2004, corresponding to 69.9 cases/100,000 population (286). In Canada, there were 5,295 cases in 2000, with an average annual incidence of 16.3 cases per 100,000 population (http://www.cureresearch.com/g/giardia/stats-country.htm). In New Zealand, 16,471 cases were reported in the 10 years of 1997 to 2006, with an average annual rate of 44.1 cases per 100,000 population (241). In China, there are approximately 28.5 million cases per year (290). The majority of giardiasis cases remain unreported. It was estimated that in the Nordic countries of Denmark, Finland, Norway, and Sweden, for every reported giardiasis case there were 254 to 867 cases of undetected and/or unreported cases and that the true incidence of symptomatic giardiasis was 4,670 cases annually per 100,000 general population (121). The disease burden in other countries is even more difficult to assess, as giardiasis reporting is not required in most countries, and data are available for only selected populations.
Deleterious effects of giardiasis on growth and development have been observed in many studies. Malnutrition is commonly seen in Giardia-positive children (237). In a case-control study including 30 children between 1 and 10 years old with giardiasis and 30 healthy children, the infected group had significantly lower body weight, serum iron levels, and zinc levels than controls (5). It is generally assumed that malnutrition results from malabsorption caused by giardiasis-associated chronic diarrhea. Nutrient malabsorption has been reported for at least 50% of patients with symptomatic giardiasis (27, 51, 238). In addition, a negative effect of giardiasis on growth and weight gain has been commonly reported (11, 186, 209). Results of cross-sectional studies suggested that symptomatic giardiasis delayed childhood growth, and diarrhea could cause growth delays up to 4 months after a diarrheal episode (51, 238). The duration of giardiasis episodes and their association with diarrhea appeared to be the most important factors associated with growth retardation (123). Even in asymptomatic giardiasis, reduction in growth (208), probably via malnutrition, was noticed (46, 186). This is in agreement with the observed association between an improvement in small intestinal mucosal function and better weight-for-age and weight-for-height Z scores (101) and the observation that giving antigiardial treatment three or four times a year improved growth in children in Brazil (208) and serum zinc levels in children in Mexico (209). Even more disturbing were the potential effects of stunting on cognition, intelligence, and psychosocial development, especially language-cognitive and fine-motor development (28, 238).
Although most cases occur sporadically, outbreaks of giardiasis are well documented. A recent review (139) indicated that there have been at least 132 reported waterborne outbreaks of giardiasis worldwide since 1954. Among them, 104 were related to drinking water, 18 were related to recreational water, and 10 were related to foreign travel. The number of cases in each outbreak varied from several to 50,000. The majority of the outbreaks were reported in North America and Europe because of better surveillance and reporting systems. In addition, food-borne outbreaks of giardiasis linked to infected food handlers and food handlers who changed diapers of infected children prior to handling food have been reported (115). Food-borne outbreaks of giardiasis have also been associated with ice, vegetables, and chicken salad (302). Outbreaks resulting from person-to-person transmission in child care centers are common (15). Community-wide outbreaks might be waterborne initially but might spread subsequently by person-to-person transmission (142). Few direct animal-to-human outbreaks have been documented. Two food-borne outbreaks of giardiasis were linked to animals: the consumption of a Christmas pudding contaminated with rodent feces and tripe soup made from the offal of an infected sheep (240).
For cattle, the infection rate varied markedly in different studies (Table (Table3),3), being 17.4% to 31.3% in Belgium (90, 91), 43.6% in Denmark (149, 163), up to 38.0% in Germany (135), 30.0% in Italy (30), 49.0% in Norway (107), 2.2 to 14.0% in Poland (22), 9.0% in Portugal (173), 26.6 to 30.1% in Spain (48, 49), 8.7 to 57.0% in Canada (17, 55, 103, 171, 192, 275), 19.1 to 52.0% in the United States (114, 231, 270-273), 3.7% in Taiwan (122), 10.2% in Vietnam (94), 58.0% in Australia (192), 4.5 to 40.6% in New Zealand (127, 152, 182, 295), and 8.0 to 10.0% in Uganda (104, 137). The age of the cattle and housing, feeding, and management practices probably contributed to the different infection rates observed. The infection rates were also different when different detection methods were used (285). Several longitudinal studies revealed cumulative infection rates of 73 to 100% in cattle (135, 210, 231, 275).
Giardia infection rates in sheep in some studies are listed in Table Table3.3. The infection rates were 25.5% in Belgium (95), 1.5% in Italy (98), 26.8% in Norway (220), 1.3% in Poland (22), 19.2 to 42.0% in Spain (47, 48, 99), 55.6% in Mexico (66), 25.4% in the United States (232), and 11.1 to 44.0% in Australia (189, 226, 298). In goats, the infection rates were 35.8% in Belgium (95), 13.0 to 42.2% in Spain (48, 50, 225), and 12.3% in Uganda (137).
Giardia infections have been reported for pigs from Australia, Asia, Europe, and North America, with infection rates ranging from 0.1% to 20% in most studies (108). An infection rate of 31.1% was observed for preweaned piglets, postweaned piglets, and sows in Australia (20) (Table (Table33).
Giardiasis in dogs and cats was reviewed recently (23), and infection rates in some studies are listed in Table Table4.4. In dogs, the infection rates were 24.8% in a large study in Europe (74), 22.7% in Belgium (53), 1.1% in the Czech Republic (68), 5.3% in Finland (215), 2.3% in Germany (73), 4.3% in Greece (198), 7.7 to 26.6% in Italy (30, 44, 197, 216, 235), 15.2% in the Netherlands (282), 8.2% in Norway (110), 2.0 to 36.0% in Poland (22, 242), 14.6% in Serbia (188), 1.0 to 7.0% in Spain (169, 175), 8.4 to 21.0% in the United Kingdom (25, 278), <0.1 to 12.9% in Canada (194, 236), 8.0% in Nicaragua (153), 3.3 to 15.6% in the United States (45, 85, 158), 1.3 to 8.9% in Argentina (82, 244), 0.8 to 36.8% in Brazil (43, 124, 141, 143, 172, 185, 287), 22.0% in Chile (161), 0.7% in Iran (235a), 0.9 to 37.4% in Japan (131, 133, 297), 11.2% in South Korea (158a), 7.9 to 56.8% in Thailand (129, 158a), and 9.4% in Australia (196). Among the various methods used, higher infection rates were detected by PCR or enzyme-linked immunosorbent assay (ELISA) than by conventional microcopy. For example, in two studies, the infection rates were 3.3% and 7.5% when microscopy was used, whereas the rates were 12.9% and 58.8% when ELISA was used (194, 251). In another study, the infection rates were 3.0% by microscopy and 20.0% by PCR (265). Similarly, the infection rate determined by PCR was almost twice as high as that determined by microscopy in another study (235).
The prevalence of Giardia has been studied less in cats than in dogs. The infection rates were 20.3% in a multicountry study in Europe (74), 15.8 to 37.0% in Italy (199, 200), 13.6% in the Netherlands (195), 5.3% in the United Kingdom (102), 10.8 to 44.4% in the United States (45, 79, 284), 4.1% in Canada (194), 5.9% in Brazil (54), 19.0% in Chile (161), 6.5% in Colombia (233), 40.0% in Japan (132), and 2.0% in Australia (196). However, low infection rates were reported in some studies, such as 0.2% in 8,160 feline specimens examined in Canada (236), 0.6% in 211,105 cats in 40 U.S. states during 2003 to 2004 (64), 2.3% in 1,566 cats in another U.S. study (85), 1.1% of 441 cats in Germany (73), none of 1,079 cats in Japan (297), and 0.9% of 113 cats in Iran (177). The health status of the cats (normal or diarrheic), age variations, and diagnostic techniques used probably contributed to the variations in infection rates in different studies. In one case, the infection rate obtained by ELISA (4.1%) was higher than that obtained by microscopy (1.0%) (194).
Giardiasis in animals is often asymptomatic but has been associated with the occurrence of diarrhea and ill thrift in calves, puppies, and kittens (257). Although it is commonly believed that infection with Giardia is associated with economic losses through the occurrence of diarrhea, poor growth, and even death in farm animals (48, 89), only a few studies have been conducted to assess the effect of giardiasis on the production or growth rates in livestock. In bottle-fed specific-pathogen-free lambs experimentally infected with Giardia cysts, the infection was associated with delays for lambs to reach slaughter weight and decreased carcass weight (191). In calves experimentally infected with G. duodenalis and treated with fenbendazole, a significant difference in weight gain was noticed between fenbendazole-treated and untreated calves. Animals in the treatment group gained on average 2.86 kg (equal to 102 g per day) more than the animals in the control group (P < 0.031) (96). In an outbreak of giardiasis on a sheep farm, Giardia-infected lambs (30 to 90 days of age) had malabsorption, decreased weight gain, and reduced feed efficiency. After treatment with fenbendazole, giardiasis-infected animals recovered rapidly from the symptoms and poor weight gain (13). However, in some other treatment studies with fenbendazole or paromomycin sulfate, differences in mean body weight, average daily weight gain, or feed intake between the control and treated groups were not significant, although there was a slightly higher weight gain and lower occurrence of diarrhea in the treated groups (88, 190). Rapid reinfection after antigiardial treatment was one possible reason for the failure to detect some benefits of the treatment.
Species of the genus Giardia infect numerous hosts, ranging from mammals to amphibians and birds. The taxonomy of Giardia was reviewed previously (42, 180, 260). Currently, six Giardia species are accepted by most researchers. Among them, G. agilis, G. ardeae, G. muris, G. microti, and G. psittaci infect various animals, whereas G. duodenalis infects humans and many mammals (Table (Table5).5). A unique species is probably also present in reptiles, as a G. duodenalis-like parasite was found in lizards. This parasite, however, mostly lacked median bodies and had binucleated cysts, and it was considered G. varani (279). Although it is expected that fish have unique Giardia spp., a recent study of cultured and wild freshwater and marine water fish in Australia showed the occurrence of zoonotic (assemblages A and B) and artiodactyl-specific (assemblage E) genotypes of G. duodenalis and G. microti (300). It was not clear whether the fish were infected with these Giardia spp. or merely served as mechanical vectors for the dissemination of waterborne Giardia cysts. Thus, Giardia species differ significantly in host range, with G. duodenalis having the broadest host range and greatest public health significance.
The species names Giardia duodenalis, Giardia intestinalis, and Giardia lamblia are used interchangeably in current literature referring to the same organism (296). Both G. duodenalis and G. intestinalis are used in equal frequency in referring to the Giardia species infecting most mammals, including humans, their companion animals, and livestock, and opinions differ regarding the legality of the name G. intestinalis, largely because of different interpretations of the complex rules of the International Code of Zoological Nomenclature (180, 259). For purposes of consistency, G. duodenalis is used in this review. In the medical field, G. lamblia is still commonly used to discuss Giardia species infecting humans (180, 259). Because numerous recent biological and genetic analyses have shown that the same Giardia species present in humans are also found in a range of other mammalian species (296), there is no taxonomic basis for the use of the name G. lamblia, which was preempted by both G. duodenalis and G. intestinalis.
Although G. duodenalis is the only species found in humans and many other mammals, including pets and livestock (261) (Table (Table5),5), it is now considered a multispecies complex. Historically, allozyme analyses placed all isolates from humans into two genetic assemblages (assemblages A and B) encompassing at least four genetic clusters (groups I to IV) (reviewed in reference 181). Phylogenetic analyses of a large set of nucleotide sequence data from the small-subunit (SSU) rRNA gene and several housekeeping genes coding for glutamate dehydrogenase (gdh), β-giardin (bg), elongation factor 1 alpha (ef1α), and triosephosphate isomerase (tpi) confirmed the genetic uniqueness of assemblages A and B. Additional lineages of G. duodenalis from animals were identified: assemblages C and D from dogs, assemblage E from artiodactyls, assemblage F from cats, and assemblage G from rodents (42, 261) (Fig. (Fig.11 and see Fig. S1 and S2 in the supplemental material).
In view of the observed host specificity and genetic characteristics of the G. duodenalis assemblages, attempts have been made recently to resurrect some of the previously used species names based on host occurrence (180). Thus, it has been proposed to adopt the names of G. duodenalis for assemblage A, G. enterica for assemblage B, G. canis for assemblages C and D, G. bovis for assemblage E, G. cati for assemblage F, and G. simondi for assemblage G (180, 260, 261). This suggestion seemingly makes logical sense and could lead to reduced confusion in Giardia taxonomy and a better understanding of giardiasis epidemiology. Because of the uncertainty related to the identity of the parasites at the initial description of these species, in compliance with the International Code of Zoological Nomenclature, redescriptions of these Giardia species that incorporate modern biological and genetic data may be needed before the proposed new Giardia species nomenclature can be accepted by the research community. This probably should include morphological descriptions and measurements of the cysts and trophozoites, data on natural host spectrum and, if any, infectivity in animal models, other biological characteristics (prepatent period, patent period, and infection site in the gastrointestinal tract, etc.), and a summary of the nucleotide sequence uniqueness of the assemblage. Such data are already available for most of the proposed Giardia species, although they are scattered in the literature. The only difficulty is probably in the naming of the species that infects canines, as lumping assemblages C and D into one Giardia species may be problematic; the sequence divergence between the two assemblages is of an order of magnitude similar to that separating the other assemblages, and at some genetic loci, such as tpi, they do not form a monophyletic group (see Fig. S2 in the supplemental material).
Among assemblages of G. duodenalis (Fig. (Fig.11 and see Fig. S1 and S2 in the supplemental material), assemblages A and B have the broadest host specificity, having been found to infect humans and various other mammals (42, 261, 296). Assemblage A is frequently found in livestock (cattle, water buffalo, sheep, goats, alpacas, and pigs) and companion animals (dogs, cats, and horses) (Tables (Tables33 and and4).4). In comparison, assemblage B is less frequently reported for livestock and companion animals, with only a few reports of infection of cattle, sheep, horses, dogs, cats, and rabbits (Tables (Tables33 and and4).4). Assemblage A and, to a lesser extent, assemblage B are commonly found in wild animals, with the exception of beavers and muskrats, which seemingly have a high occurrence of assemblage B (Table (Table6).6). Both assemblages A and B are commonly reported to infect humans (Tables (Tables11 and and2).2). Because they are found in humans and numerous species of mammals, both assemblages A and B are considered to have broad host specificity and can be transmitted zoonotically (41, 261, 296). The host range of assemblages A and B is shown in Table Table5,5, and their distributions in humans and various animals are shown in Tables Tables11 to to44 and and66.
There is substructuring within assemblage A, which consists of mostly two subgroups, subgroups or subassemblages AI and AII. The separation of subgroups AI and AII was initially made by allozyme analysis and is supported by phylogenetic analyses of assemblage A sequences at the gdh locus (Fig. (Fig.2).2). The existence of numerous subtypes related to subgroups AI and AII, however, made the separation of subgroups AI and AII less obvious at some other loci, such as tpi and bg (see Fig. S3 and S4 in the supplemental material). Recently, a third subgroup within assemblage A, subgroup AIII, was identified and appears to be associated mostly with wild hoofed animals (37, 219, 282). It has significant sequence differences from subgroups AI and AII at all loci examined thus far (Fig. (Fig.22 and see Fig. S3 and S4 in the supplemental material). More recently, based on an analysis of the tpi, gdh, and bg loci, subgroup AIII was found in three cats, four cattle, and 45 wildlife but has not been found in dogs, goats, sheep, pigs, and humans thus far (247). In comparisons with subgroup AIII, subgroups AI and AII always form one cluster with high bootstrap values in phylogenetic analyses of nucleotide sequences of all loci (Fig. (Fig.22 and see Fig. S3 and S4 in the supplemental material). In contrast, there is no clear subgrouping within assemblage B (293) (Fig. S5 to S7).
Assemblages C, D, E, F, and G have strong host specificities and narrow host ranges. Assemblages C and D have been found mostly in dogs and other canines (foxes and coyotes) and canine-related animals (seals) (Table (Table4).4). Assemblage E has been found largely in cloven-hoofed domestic mammals (cattle, water buffaloes, sheep, goats, and pigs) (Table (Table3).3). Assemblages F and G have been found mostly in cats and rodents, respectively (Tables (Tables44 and and6).6). However, there are occasional exceptions to the host specificity. Thus, assemblages C and D were reported in a few cats (196, 214) and humans (264), assemblage D was also reported in two pigs (149, 247), assemblage E was reported in cats (154, 214, 247) and humans (83), and assemblage F was reported in one pig (20) and seven humans (87). It is commonly believed that humans are infected only with assemblages A and B (Tables (Tables11 and and2),2), and the recent identification of unusual G. duodenalis genotypes, such as assemblages C, D, E, and F in humans (83, 87, 264), requires thorough genetic characterizations of the parasites.
In addition to the above-described accepted assemblages, several novel genotypes have been reported. Giardia cysts isolated from U.S. seals were reported to belong to a new genotype based on a sequence analysis of the gdh gene (86), and the same genotype was found in gray seals, harbor seals, and a gull in a later study and was recently named assemblage H (151) (Fig. (Fig.1).1). However, the existence of this new assemblage was supported by the sequence analysis of only gdh but not tpi (151), and the gdh sequence of the new genotype is placed outside the G. duodenalis cluster in phylogenetic analyses (Fig. (Fig.1).1). One Giardia isolate from a southern brown bandicoot (Isoodon obesulus, commonly known as quenda) in Western Australia was reported to be a novel genotype based on a phylogenetic analysis of the SSU rRNA (GenBank accession number AY309064) and ef1α (accession number AY309065) gene sequences. Those authors believed that this isolate constituted a distinct species, because phylogenetic analysis shows it to be distinct from other recognized species such as G. microti, G. psittaci, G. ardeae, and G. muris (7). This new genotype was identified in another six quenda specimens in a recent study (262). Two novel genotypes of G. duodenalis were found in house mice on Boullanger Island, Australia, based on a sequence analysis of a 130-bp region of the SSU rRNA gene (183), although the sequence data were not presented. All these new genotypes need to be characterized extensively, at least at the common loci used for genotyping G. duodenalis.
Molecular biology has provided powerful new tools for characterizing Giardia, and the analysis of previously unrecognized genetic differences within this genus has revolutionized our understanding of the taxonomy, population genetics, and epidemiology of giardiasis in humans and domesticated animals. Although some simple PCR assays have been used for the detection of Giardia in clinical and environmental samples, most recent molecular tools are used for the differentiation of Giardia at the species/assemblage and genotype levels. These tools are widely used for the identification of G. duodenalis genotypes in clinical specimens (293).
The utility of molecular diagnostic tools is determined by the gene targeted (such as the SSU rRNA, gdh, tpi, ef1α, bg, and variant surface protein [vsp] genes), the number of loci used in the analysis, the specificity of the assay (Giardia specific, G. duodenalis specific, or assemblage specific), and downstream procedures (restriction fragment length polymorphism [RFLP] analysis or DNA sequencing of PCR products). The usage of these loci for the genotyping and subtyping of G. duodenalis and their sequence characteristics were reviewed recently (293). Some commonly used primers for the species/genotype and subtype differentiation of Giardia isolates in animal and human specimens and water samples in recent studies are listed in Table Table7.7. Among them, the SSU rRNA, gdh, tpi, and bg genes are frequently targeted.
Since the genetic loci of Giardia differ in substitution rates, the resolution for parasite typing is different among loci. For example, substitution rates for the partial SSU rRNA, bg, gdh, and tpi genes were reported to be 0.01, 0.03, 0.06, and 0.12 substitutions per nucleotide, respectively (293). Thus, the conserved SSU rRNA gene is traditionally used for species and assemblage differentiation (mostly genotyping), whereas the most variable locus, tpi, is frequently used for subtyping. The bg and gdh loci, with substitution rates between those of the SSU rRNA and tpi genes, have a broad application spectrum (293).
The SSU rRNA gene is a commonly used marker for the species and assemblage differentiation of Giardia, and different regions have been targeted for different applications. The variable 5′ and 3′ ends of the SSU rRNA gene locus can be used for identifying the relatively closely related assemblages, whereas the more conserved regions would provide sufficient information only for the differentiation of Giardia species. Thus, when the SSU rRNA gene locus is used for assemblage differentiation, primer selection should be careful because the products by some primer sets are too small to differentiate all G. duodenalis assemblages. For example, some primers (118, 283, 291) amplify only the first 1/5 or the last 1/10 of the gene (~1,400 bp in total). Using one such set of primers (118), cat isolates were genotyped as assemblage A in one study (30) and assemblage F in another study (79) due to the fact that assemblage F is mostly identical to assemblage A in the fragment under analysis (293). The identification of unusual assemblages (assemblages C, D, E, and F) in humans involved mostly the use of SSU rRNA gene-based tools. One reason for targeting small fragments of the SSU rRNA gene is the difficulty in the PCR amplification of the locus; the use of dimethyl sulfoxide or special PCR buffers designed for GC-rich targets is frequently needed for efficient PCR amplification of the target.
The tpi, gdh, and bg loci are also common genotyping markers. Because most primers amplify 40% to 60% of the gdh gene and 60% of the tpi and bg genes, these loci have been used for both genotyping and subtyping. However, inconsistent genotyping results are sometimes generated among different loci. In a study of human and dog specimens in a tea-growing community in India, SSU rRNA gene sequencing suggested that assemblages C and D were the dominant genotypes in humans but were absent in dogs. The finding of assemblages C and D in humans was not supported by a sequence analysis of the specimens at the ef1α and tpi loci (265).
The occurrence of infections with mixed assemblages in humans and animals is common. For 908 isolates from human and various animals characterized at two or more loci, mixed infection of Giardia assemblages was found for 121 (13.3%) isolates. Among them, mixed infection was found for 46 of 392 (12%) human specimens and 45 of 134 (34%) dog specimens analyzed. Mixed infection was also found in cat, cattle, goat, sheep, pig, and wildlife specimens. Mixed infection involving assemblages A to E is especially common. In some dogs, the concurrent occurrence of three assemblages (assemblages A, B, and C or B, C, and D) was seen (247). Therefore, assemblage-specific PCR assays are now used to assess more accurately the occurrence of mixed infections. The usage of such a system targeting the tpi gene in analyses of specimens from dairy and beef calves led to the identification of mixed assemblage A and E infections in 31% of positive calves and a more frequent occurrence of assemblage A in dairy calves (59% of all positive samples) (91). Primers specific for assemblage A or B based on the tpi gene were reported and should be useful for an accurate assessment of the occurrence of mixed infections in clinical specimens (14, 31, 223). Such data will improve our understanding of the clinical spectrum of giardiasis, intensity of giardiasis endemicity in study areas, tracking of infection sources, and identification and differentiations of outbreaks of giardiasis.
Ideally, genotyping should be performed at the single-cyst level, as this will allow the differentiation between mixed infections and the occurrence of recombinants, because there is a possibility of genetic exchanges between isolates of assemblage A (57) or even between isolates of assemblages A and B (254). The use of real-time PCR appears to be promising in reaching this technically demanding level of sensitivity and specificity. A real-time PCR assay targeting the bg gene was developed. It could differentiate assemblages A and B of G. duodenalis with a sensitivity of detecting an equivalent of one cyst of G. duodenalis (105). More recently, three real-time PCR assays targeting the tpi, gdh, and open reading frame C4 sequences were developed to differentiate assemblages A and B. The assays had high specificity and detected DNA from a single trophozoite or cyst (9). When these assays and a TaqMan assay targeting the bg gene were used to analyze 30 human stool samples, a simultaneous detection of both assemblages was seen for 37 to 83% of specimens, depending on the genes targeted. PCR analyses of DNA extracted from single cysts purified by immunomagnetic separation from the same specimens showed that the simultaneous amplification of DNA of both assemblages was attributable to mixed infections. However, potential genetic exchanges between assemblages A and B were suggested by the detection of both assemblages in individual cysts (9). There is also a single-tube multiplex real-time PCR assay based on the SSU rRNA gene for the differentiation of assemblages A and B in fecal specimens but with a low sensitivity of an equivalent of 20 trophozoites per PCR (187).
Assemblages A and B are zoonotic; therefore, genetic differences within them may provide information on the relationship of subtypes and hosts, the zoonotic potential of each subtype, contamination source tracking, outbreak investigation, as well as characterizations of transmission dynamics. Based on sequence analysis of the tpi, gdh, and bg genes, multiple subtypes of assemblage A were noticed. At the gdh locus, these subtypes can be grouped easily into three subassemblages, subgroups AI, AII, and AIII, and there are multiple subgroup- or subassemblage-specific polymorphisms (Fig. (Fig.2).2). At the tpi and bg loci, subassemblage AIII also has many distinct nucleotide substitutions, but there is no formation of robust subgroup AI and AII clusters in phylogenetic analyses of the sequences (see Fig. S3 and S4 in the supplemental material). The initially identified two subassemblage-specific substitutions in the tpi gene apply only to the common A1 and A2 subtypes, which belong to subassemblages I and II, respectively. Thus, the A5 subtype (belonging to the AI subassemblage) (37) has a tpi sequence more similar (two nucleotide substitutions versus four) to A2 (an AII subtype) than to A1 (an AI subtype).
To systematically characterize intra-assemblage genetic diversity and provide more-robust typing of G. duodenalis, Caccio et al. proposed a subtype nomenclature system (Table (Table8)8) for assemblage A based on multilocus genotype (MLG) analysis of the bg, gdh, and tpi genes (37). This would reduce confusions in subtype terminology and improve understandings of the host segregation of subtypes. Nevertheless, the subgroup AI, AII, and AIII classification system proposed was determined largely by the extent of nucleotide substitutions at the gdh locus, although concatenated sequences of the gdh, tpi, and bg genes were used in the phylogenetic assignment of the subassemblages. The designation of rare subtypes within the AI and AII subassemblages may need support from genetic characterizations of other loci.
The BIII and BIV subassemblages originally described by allozyme electrophoretic studies are not supported by DNA sequence analyses. The sequences from known subgroup BIII or BIV isolates are insufficient to assess the existence of subgroup BIII/BIV-specific sequence polymorphisms, and subtyping analyses of field isolates produced inconsistent subassemblages among different loci, most of which were not supported by bootstrap analyses (293) (see Fig. S5 to S7 in the supplemental material). Assemblage B has many more subtypes than does assemblage A.
Many subtypes are also seen in assemblage E. Although assemblage E also infects a range of hoofed livestock, there are no host-specific subgroups in nucleotide sequences of various genetic loci. There are insufficient data to assess the substructure of other assemblages (assemblages C, D, F, and G). Host-specific sequence polymorphism was not seen at the gdh locus when assemblage G isolates from rats and mice were examined (293).
The low resolution of current genotyping tools has limited their potential for the characterization of the transmission of human giardiasis (42). Thus far, the vast majority of studies have relied on the sequence characterization of human- and animal-derived cysts at one or two genetic loci (41). In early studies, there was a bias toward the use of the SSU rRNA gene target because of its multicopy nature and high degree of sequence conservation. This has led to some problematic interpretations of the data acquired. In addition, recent data have shown that inconsistent genotyping results may be obtained when different genetic loci are targeted. Thus, MLG analysis is increasingly used for the characterizations of G. duodenalis from humans and animals (37).
Several recent studies using the MLG approach showed that some isolates of both human and animal origins could not be unequivocally assigned at the assemblage level (37, 265). In one study where a sequence analysis of four genes (the SSU rRNA, bg, gdh, and tpi genes) of human and animal isolates was performed, the data generated showed that although congruent genotyping results were obtained at the four genetic loci for isolates of assemblage A, it was difficult to assign some assemblage B isolates from humans and nonhuman primates to an assemblage, because the result obtained at one locus was not consistent with data obtained at the other three loci. In addition, there were variable levels of intraisolate sequence heterogeneity in assemblage B isolates, which prevented the unambiguous identification of MLGs (37). Similar results were reported by other studies of human and animal isolates using some of the same targets (93, 265). This inconsistency in genotyping results was more frequently reported for isolates from dogs, where, depending on genetic loci, isolates were typed as either host-adapted assemblages C and D or zoonotic assemblage B (265). This has important repercussions, as different conclusions may be reached. These findings also raise concerns about the interpretation of genotyping data based on single markers. It was reported that 15% of isolates genotyped (including 2,400 Giardia sequences from the GenBank database) in the ZOOnotic Protozoa NETwork (ZOOPNET) had inconsistent typing between two markers among the SSU rRNA, bg, gdh, and tpi genes, and this inconsistency was observed predominantly for specimens from humans and dogs (247).
One possible factor contributing to the inconsistent genotyping among markers is the high level of occurrence of mixed infections as a result of the high prevalence of giardiasis in humans and animals. This was supported by the frequent detection of concurrent infections of both assemblages A and B in humans and the high level of occurrence of assemblage A infection in animals when assemblage-specific PCR was used for genotyping. The grouping of the parasites into individual MLGs was further complicated by the fact that many isolates exhibited double peaks at specific positions in electropherograms of otherwise clean sequences. In one recent study, double peaks in electropherograms were seen in assemblage B, C, D, and E isolates but never in assemblage A, F, and G isolates, suggesting that MLG analysis of G. duodenalis might be more useful for the typing of assemblage A (154).
Giardia species have long been assumed to have exclusively asexual reproduction. Results of recent MLG studies suggest the occurrence of genetic recombination in Giardia (57, 151a, 207, 211). Possible meiotic recombination was seen within some loci under analysis when four regions of three chromosomes of the subgroup AII reference strain JH and five subgroup AII isolates from an area in Peru where the disease is highly endemic were sequenced (57). In another study of 978 human and 1,440 animal isolates at four loci (the SSU rRNA, bg, tpi, and gdh genes), the allelic sequence heterogeneity and genetic recombination within assemblage B were inferred. Mixed genotypes were repeatedly seen for individual isolates, particularly those of assemblage B (247). The high level of allelic sequence heterozygosity in assemblage B in contrast to assemblage A was confirmed by sequence analysis of 30 Thai isolates at the bg locus (144a) and by the whole-genome sequencing of isolate GS, which showed that only a few regions of the genome appeared to be free of heterozygosity (84). The possibility of recombination between different assemblages, especially between assemblage AI and B isolates, was inferred in another study. When PCR products of nine isolates of assemblages AI, AII, and B were clone sequenced at 10 loci, there were rare assemblage B haplotypes that grouped with assemblage AI at several loci. In contrast, all assemblage AII isolates formed a single clade with no evidence of recombination (254).
Current population genetic data do not allow us to distinguish between meiotic sex and parasexual reproduction (33). Further investigations are required to better understand the population structure and reproduction of Giardia spp. Emphasis should be placed on intra-assemblage comparisons, as available data indicate that the genetic assemblages of Giardia are conserved in terms of geographic location and host occurrence, suggesting that any recombination is probably not reflected at the assemblage and species levels (180). This was supported by results of a recent MLG study of 114 Giardia isolates from various animals (pets, livestock, wildlife, and captive nonhuman primates) in Sweden at three loci (bg, tpi, and gdh), which showed no evidence of recombination between assemblages, although allelic sequence divergence was commonly noticed (154). However, in a recent study, both intra- and interassemblage recombination and meiotic sex were seen in assemblages A to G, which challenges the rationale for naming G. duodenalis assemblages different species (151a).
The evolutionary advantage of recombination is the capacity for the organism to respond to adverse conditions such as selection pressures imposed by immunity and antigiardial treatment (144a, 180). Thus, it was observed many years ago that metronidazole- or furazolidone-resistant G. duodenalis clones underwent chromosomal rearrangement (277). More recently, a whole-genome sequence analysis suggested that the large vsp gene repertoire (270 to 303 copies) in isolate WB of assemblage A is probably the result of gene duplication and recombination (6). The occurrence of allelic heterogeneity and genetic recombination makes the development of high-resolution subtyping tools more difficult and the analyses of sequence data intricate. In contrast, if G. duodenalis has a clonal population structure, the use of only a few highly polymorphic markers in epidemiological investigations would be justified (249). As the clonality in G. duodenalis is currently challenged by data indicative of both heterozygosity and genetic recombination, the single-genetic-locus approach used by many previous studies should be reevaluated. Thus, investigations of the distribution of G. duodenalis genotypes and subtypes in humans and the identification of infection sources and risk factors now require more discriminatory typing techniques that allow the identification of individual lineages. These techniques will undoubtedly improve the understanding of the molecular epidemiology of giardiasis.
Few epidemiological studies have assessed the importance of zoonotic transmission in the occurrence of human giardiasis. In New Zealand, case-control studies of giardiasis did not identify contact with pets as a risk factor for children or adults, although contact with farm animals was associated with an increased risk of infection for adults (119, 120). In agreement with this finding, the infection rate of human giardiasis in New Zealand was 23% higher in rural areas than in urban areas (241). In the United Kingdom, farm visitation was frequent among case patients, but specified exposures to dogs, cats, horses, cattle, and sheep was not a significant risk factor (248). One case-control study in eastern England found an association of giardiasis with exposure to farm animals and pets, particularly pigs, dogs, and cats (296). Other studies in the United States, Canada, and the United Kingdom did not show such an association (296).
Beavers are probably the animals most commonly implicated in the zoonotic transmission of giardiasis. This originated largely from investigations of waterborne outbreaks and reports of frequent giardiasis in hikers and campers, who recalled drinking water from streams or lakes. The first observation was made in 1976 with an outbreak of giardiasis in Camas, WA, when 128 townspeople had laboratory-confirmed giardiasis. A questionnaire survey revealed that 3.8% of residents had clinical giardiasis, whereas none of 318 residents in a control town were ill. Giardia cysts were recovered from water entering the city water treatment system and two storage reservoirs containing chlorinated and filtered stream water. Trapping in the watershed yielded three beavers infected with Giardia that was infective for beagle pups. Similar observations were made by subsequent investigations of waterborne giardiasis outbreaks in New Hampshire and Nevada, based largely on the finding of Giardia cysts in beavers caught in source watersheds (reviewed in reference 296). These reports did not provide concrete evidence of an involvement of beavers in human infections and reported merely circumstantial evidence based on beavers found at contaminated sites. The possible role of humans and other animals as a source of infection for beavers was not addressed in these reports. In a review of data from case-control studies, the association between the drinking of wilderness water and acquisition of giardiasis was deemed minimal (292).
The consumption of raw surface water clearly represents a significant risk for giardiasis (119). However, the contamination of such water supplies may result from humans, farm animals, and wildlife (112). A 2-year study assessed the significance of each source in the environmental contamination of surface water with Giardia cysts (112). Sewage effluent was shown to have the highest prevalence of Giardia, although the concentration of cysts was minimal compared with that detected in cattle feces. Although the overall prevalence of Giardia was lower in wildlife, giardiasis was prevalent in aquatic mammals such as beavers and muskrats. Nevertheless, an interpretation of these results in the context of the source of human Giardia infections can be made only in conjunction with data on the distribution of Giardia species and genotypes in these potential contamination sources (257).
Numerous isolates of Giardia collected from different host species in various geographical locations have been genotyped, and the occurrence of the same species/genotype in humans and other animals has been well demonstrated (181). Such data are indicative of zoonotic potential, and most experts would agree that G. duodenalis strains are potentially zoonotic, especially assemblages A and B of G. duodenalis. The use of genotyping and subtyping tools, especially the more recent MLG tools, in well-designed epidemiological studies, however, is needed before we can fully assess the human disease burden caused by zoonotic giardiasis (37).
The occurrence of Giardia in aquatic wildlife in investigations of waterborne outbreaks of giardiasis, particularly of isolates that are morphologically identical to G. duodenalis, has been the single most important evidence implicating Giardia as a zoonotic agent. Indeed, recent studies have shown the common occurrence of G. duodenalis assemblage B in beavers and muskrats (78, 249). However, there is little evidence to implicate these animals as the original contaminating source in waterborne outbreaks. It has been suggested that these animals are more likely to become infected from water contaminated with fecal material of human or even domestic animal origin, thus serving to amplify the numbers of the original contaminating isolate (257).
In view of the frequency of human-pathogenic Giardia genotypes, the public health risk of giardiasis from domestic animals appears to be small. This is the case at least for cattle in North America, Australia, and Europe, where G. duodenalis assemblage E predominates (149, 231, 249, 271, 272). The human-pathogenic assemblages (assemblages A and, occasionally, B) in cattle may have to compete with the more common assemblage E. Similar conclusions can probably be drawn for other livestock such as sheep, goats, and pigs, where assemblage E is also the predominant genotype (Table (Table3),3), or domestic pets such as dogs and cats, which have other host-adapted G. duodenalis genotypes (assemblages C and D in dogs and assemblage F in cats) as the dominant parasites (Table (Table44).
Most interests in the zoonotic transmission of G. duodenalis focus on assemblage A, as it is the most common non-host-specific assemblage in animals (Tables (Tables3,3, ,4,4, and and6).6). Subtyping data accumulated so far do not support a widespread occurrence of zoonotic transmission. The two most common subassemblages of assemblage A, subgroups AI and AII, appear to differ in host preference (247, 296). Humans are more commonly infected with subgroup AII, although subgroup AI has also been seen in some areas or studies (Table (Table99 ). In contrast, animals are commonly infected with subgroup AI, although subgroup AII is sometimes seen (Table (Table10).10). More systematic characterizations of the less common subtypes within subgroups AI and AII, however, are needed to improve our understanding of the host specificity of subgroups AI and AII. The other subgroup, subgroup AIII, has thus far been found only in animals, mostly in wildlife (Table (Table10).10). There are no clear geographic differences in the distribution of the two common assemblage A subgroups in humans; although subgroup AI was detected in humans at a high frequency in two studies in South America, two other studies conducted in the same countries found only subgroup AII in humans (Table (Table9).9). Increasing the typing resolution and detailed epidemiological data are needed to determine what portion of human subgroup AI infections is the result of zoonotic transmission.
MLG tools have recently been used in assessments of the zoonotic transmission of giardiasis because of their highly discriminatory power. When 978 human and 1,440 animal isolates were characterized using sequence analysis of four loci (the SSU rRNA, bg, tpi, and gdh genes), the zoonotic potential of both assemblages A and B was apparent at the levels of assemblages and subassemblages at each locus. However, when isolates were defined using an MLG scheme, only two MLG types of assemblage A and none of assemblage B appeared to have occurred in both humans and animals (247). In contrast, three assemblage A MLG types that were previously detected in humans were present among 114 Giardia isolates from various animals (pets, livestock, wildlife, and captive nonhuman primates) in Sweden when they were analyzed by an MLG technique targeting three genes (bg, gdh, and tpi) (154).
The finding of the same genotypes or MLG types in humans and animals is not by itself conclusive evidence that zoonotic transmission has taken place. A better assessment of zoonotic transmission can only come from studies that examine the dynamics of Giardia transmission between humans and animals living in the same household or localized focus of endemicity. In Aboriginal communities in Australia, SSU rRNA gene sequences from 13 humans and nine dogs identified assemblages A and B in humans and assemblages C and D in dogs. Only one dog had an assemblage B isolate, suggesting that zoonotic transmission between humans and dogs, if present, was infrequent, and the dog could have acquired the assemblage B infection from a human source (265). Similarly, in a community in which giardiasis is highly endemic in Peru, multiple examinations of dogs and humans in 22 households showed the presence of assemblages A and B in 167 human specimens genotyped and assemblages C and D in 67 canine specimens genotyped (57a). In contrast, in a study conducted in 20 temples and surrounding communities in Bangkok, Thailand, of 13 Giardia isolates from dogs and 3 from humans, 1 dog and 2 monks in the same temple had assemblage A isolates (129). In another similar study in the same area, 42 and 35 Giardia-positive specimens from dogs and humans, respectively, were chosen for genotyping. Assemblage A was shown to be the most common genotype in dogs (79%), followed by assemblages D (31%), B (21%), and C (12%). Likewise, 73% of humans were infected with assemblage A, followed by assemblage B and assemblage C. Based on the high level of occurrence of assemblage A in both dogs and humans, dogs were considered reservoirs for human giardiasis in temple communities in Bangkok (264).
More recently, subtyping studies of isolates from humans and animals living in the same community or area were conducted to further assess the likelihood of the zoonotic transmission of giardiasis. In a socially deprived small Rom community in Italy where dogs roamed freely, G. duodenalis specimens from children and dogs both had only subassemblage AI (168). Likewise, one child and a dog living in the same household were found to be infected with a subassemblage AI strain in a Brazilian study (287). In contrast, in Assam, India, subassemblage AII was found to be the dominant G. duodenalis assemblage in both humans and a dog living on a tea estate (265, 266), and in one household, two isolates from humans and one isolate from a dog all belonged to subassemblage AII. Evidence for zoonotic transmission was further supported by epidemiological data showing a highly significant association between Giardia infection of humans and the presence of a Giardia-positive dog in the same household. Unfortunately, it was not clear how many subassemblage AII cases in dogs in the community represented active infection. It was noticed that over 30% of dogs in the community had Ascaris lumbricoides eggs in high intensities in their feces (265, 266). As A. lumbricoides is a human-specific pathogen and dogs were shedding A. lumbricoides eggs after coprophage of human feces, some dogs might have acted as mechanical disseminators of Giardia subtype AII rather than being actively infected with this more human-adapted parasite.
Most cattle, sheep, and pigs are infected with G. duodenalis assemblage E (Table (Table3).3). Among 562 studied samples from cattle in Europe, 422 (75%) were found to be positive for assemblage E (247). A longitudinal study of dairy herds in Australia over several months found that 100% of calves became infected during the first 12 weeks of life, and all the animals sampled had assemblage E isolates (26). This livestock genotype was found to predominate in cattle in North America, Europe, and Australia (26, 149, 231, 271, 272). Although assemblage E appears to occur most frequently in cattle, studies in Belgium, the United States, and Europe have shown that a small proportion (<20%) of cattle in a herd may harbor assemblage A, the most common zoonotic genotype (231, 247, 271, 272) (Table (Table3).3). In contrast, assemblage B was found in only a small number of cattle in a few studies, and other assemblages have never been convincingly found in cattle (Table (Table3).3). The only exception is New Zealand, where limited studies indicated that assemblages A and B appeared to be common in cattle and that assemblage E was largely absent (127, 152, 294).
The actual infection rate of assemblage A in cattle may be higher than previously believed. A longitudinal study of 30 calves from birth to 24 months of age on a dairy farm in Maryland showed that the cumulative infection rate of assemblage E reached 100% by 7 weeks of age and that of assemblage A reached 70% by 15 months of age (231). Another longitudinal study of adult cattle also reported that 43% of positive isolates were assemblage A and 57% were assemblage E isolates (275). The findings of the two recent longitudinal studies were somewhat unexpected considering the low levels of assemblage A previously reported for cattle (17, 125, 192). In addition, a previous longitudinal study of dairy herds in Australia (26) indicated that zoonotic genotypes might be present only transiently in cattle when the frequency of transmission with assemblage E was high and competition between genotypes was likely. Additional longitudinal studies are necessary to determine if the data mentioned above are representative of dairy farms in general.
An age-associated difference in the distribution of assemblages A and E was reported for cattle. Four multistate pooled point-prevalence studies of G. duodenalis on 14 dairy farms in the eastern United States involving nearly 2,000 cattle from birth to adulthood reported that assemblage E was found in 34% of preweaned calves, 45% of postweaned calves, 33% of heifers, and 25% of cows, whereas assemblage A was detected in 6%, 7%, 3%, and 2% of the animals, respectively (270-273). However, in a longitudinal study of 30 calves from birth to 24 months of age on a dairy farm in Maryland conducted by that same group, assemblage E was detected in 61%, 25%, and 6% of preweaned calves, postweaned calves, and heifers, respectively, whereas assemblage A was detected in 0%, 7%, and 5% of the animals, respectively (231). Differences in sample sizes, housing conditions, and management practices were believed to be responsible for the differences in the distribution of G. duodenalis assemblages in dairy cattle among studies (231). Some studies reported that assemblage A was less commonly found in beef than in dairy cattle (17, 66a, 91).
There are only a few studies of subtypes of G. duodenalis in cattle, and subassemblage AI was found to be the major subassemblage (Table (Table10).10). In one study conducted in Europe, among 113 samples tested, 70 belonged to subassemblage AI, 39 belonged to subassemblage AII, and 4 belonged to subassemblage AIII (247). In one small-scale study conducted on three dairy farms, the intragenotypic diversity of G. duodenalis was studied by analyzing 58 bovine specimens harboring assemblages E (48) and A (8) and a mixture of them (2). Eleven distinct subtypes were identified in assemblage E; all farms had multiple subtypes of assemblage E, and concurrent infection with mixed subtypes occurred in 24% of animals. In addition, both subassemblages AI and AII were identified in assemblage A isolates. Therefore, the high intragenotypic diversity and occurrence of mixed infections reflect the high intensity of G. duodenalis transmission in cattle (81). More subtyping studies are needed to understand the transmission dynamics of Giardia infection in cattle.
Like in cattle, sheep and goats are infected predominantly with assemblage E, with assemblage A being identified infrequently (Table (Table3).3). In studies with reasonable sample sizes, assemblage E was the most common genotype identified in preweaned lambs (36/52 lambs in one study and 74/75 lambs in another), juvenile and adult sheep (33/46 sheep), goats (39/39 goats), or both animal species (170/207 animals) (99, 225, 226, 247, 298). Multiple bg subtypes of assemblage E were also found in lambs, with 1 to 6 subtypes on each farm (99, 225, 226, 298). Although assemblage A was also found at a high frequency in some studies, subtyping showed that the isolates belonged mostly to subassemblage AI instead of subassemblage AII (Table (Table10).10). In contrast, assemblage B is rarely found in sheep (Table (Table3).3). One outbreak of giardiasis in lambs with severe weight loss and some mortality was attributed to assemblage B, although the conclusion was based on a PCR analysis of only two specimens (13).
Although assemblage E is the predominant genotype, assemblage A has also been found frequently in pigs (Table (Table3).3). In Australia, assemblage E was the most common genotype and was detected in 64% and 67% of positive specimens from preweaned and postweaned pigs, respectively, with the balance as assemblage A (20). In Denmark, assemblage E was also the most common genotype, being identified in 62% of specimens from postweaned pigs, while assemblage A was detected in only 12% of specimens (149). In Europe, assemblage E was the most common genotype, being detected in 78% of 140 positive pigs studied, followed by assemblage A in 21% of pigs. Surprisingly, assemblages B and D were also found in pigs, although they accounted for less than 1.5% of samples from the positive pigs (247). The majority of the assemblage A strains in pigs belonged to subassemblage AI (Table (Table1010).
Earlier studies reported that dogs were infected mostly (but not exclusively) with host-specific G. duodenalis, with assemblages C and D being the major parasites detected. Several recent studies, however, have shown modest to high infection rates of assemblage A (Table (Table4).4). In contrast, assemblage B was found only occasionally in dogs (53, 174, 214, 247, 264, 265), some of which might have resulted from the ingestion of human feces (266). Among assemblage A isolates from dogs subtyped, the majority of them belonged to subassemblage AI, with subassemblage AII being occasionally noticed (Table (Table10).10). One study, however, reported that subgroup AII was the dominant subgroup (53).
Differences in social and environmental conditions might have contributed to the variations in the distribution of G. duodenalis assemblages in dogs. It was suggested that there are probably two transmission cycles in domestic urban environments, with the transmission of dog-specific genotypes among dogs and the possible transmission of assemblage A between pets and humans. The transmission of dog-specific genotypes may be favored by intensive contact among large numbers of dogs living together and may outcompete the transmission of other genotypes. In household dogs, the frequency of dog-to-dog transmission may be lower, and consequently, infections with assemblage A in dogs are likely to persist (260). This was supported by the finding of high percentages of assemblage A isolates in household dogs (147, 155) and high percentages of assemblages C and D in kennel dogs (30, 251). This was further confirmed by a study from northern Belgium, where 80.5% of 41 Giardia-positive household dogs excreted assemblage A, while dogs in breeding kennels and clinically affected dogs had mainly dog-specific assemblages C and D (93.9% of 33 dogs and 80.0% of 45 dogs, respectively) (53). In some other surveys, however, household dogs were shown to be infected mainly with assemblages C and D (130, 195), and kennel dogs were shown to be infected predominantly with assemblage A (130, 155). Because only one sample was collected from each dog in these studies, it was hard to know whether the genotype detected was from new infection or from stable transmission. The percentage of zoonotic and dog-specific assemblages may change over time, depending on the competition of the two transmission cycles. In a study of domestic dogs living in urban environments (155), mixed infections of assemblages A and C in both group dogs (38%) and individual dogs (21%) were common, indicating the existence of the two transmission cycles in domestic urban environments. In addition, although assemblage A was prevalent in both dogs kept individually (91%) and dogs kept in groups (81%), the higher infection rates of assemblages C and D in group dogs (57%) than in individual dogs (30%) suggests that under conditions of intensive dog-to-dog transmissions of giardiasis, the host-adapted genotype is likely to outcompete the zoonotic genotype and become the dominant genotype.
Cats are infected with assemblages A and F, with the cat-specific assemblage F being found more frequently (Table (Table4).4). Among the limited numbers of assemblage A isolates from cats subtyped, subgroup AI was the dominant one, although subgroups AII and even AIII were also reported (Table (Table10).10). Surprisingly, in a multicountry study, assemblages A to F were all found in cats, although assemblages A and F were the dominant ones (247). Similar results were found in another study where almost all known G. duodenalis assemblages (assemblages A, B, C, D, and E) except F were identified in cats (214). However, there was a high level of discrepancy in the genotyping results of animal specimens between the two genotyping tools used in the study.
To date, there are few data on genotypes of G. duodenalis in horses. Two assemblage A isolates established in suckling mice, Ad-159 and Ad-162, were initially obtained from Australian horses (263). In a study of the molecular epidemiology of giardiasis in horses, 10 isolates of G. duodenalis recovered from horses in Ithaca, NY, and Perth, Western Australia, were characterized at the SSU rRNA and tpi gene loci. The results indicated that 3, 1, and 6 horses were infected with assemblages AI, AII, and B, respectively (263). In a more recent study from Italy, however, all 20 isolates characterized belonged to assemblage E (285). In other domestic pets, assemblage B was found in one rabbit each in China and Sweden (154, 249), and assemblage A was found in three ferrets in Japan (3, 4). MLG analysis of the three ferret isolates suggested that they might belong to two host-adapted assemblage A subtypes (4).
It was the association between infected beavers and waterborne outbreaks of human giardiasis that led the World Health Organization to classify Giardia as a zoonotic parasite (257). Results of recent studies have shown that beavers are frequently infected with human-pathogenic G. duodenalis genotypes. In two studies, all 11 isolates from beavers in the United States belonged to assemblage B (78, 249). In another study, 12 of 113 (10.6%) beaver fecal specimens from southern Alberta, Canada, had assemblage A (16). Assemblage B was found in some muskrats in the United States, in addition to G. microti (249). In other rodents, assemblage A was found in a chinchilla in Germany (138), assemblage G was found in rats in Australia and Sweden (154, 178), assemblage E was found in one ash-gray mouse, and assemblages F and C were found in one bush rat in Australia (262). The finding of assemblages other than rodent-specific assemblage G in rodents may be due to the habit sharing of wildlife with domestic animals and humans (262).
Both assemblages A and B are commonly found in other wild mammals (Table (Table6).6). In wildlife in Europe, 74% of 172 samples belonged to assemblages A and B (247). In nonhuman primates, assemblage A was found in 16 southern brown howler monkeys in Brazil (288) and two gorillas in Uganda (104), whereas assemblage B was found in many other nonhuman primates, such as macaques, chimpanzees, and mandrills (20a, 37, 130). In ruminants, assemblage A was found in moose, fallow deer, reindeer, white-tailed deer, and roe deer (146, 154, 176, 219, 269, 282). Genotyping results indicated that G. duodenalis isolates infecting moose, reindeer, and fallow deer were either of assemblage AI or AIII. Surprisingly, assemblage E is largely absent in wild ruminants (Table (Table6).6). In carnivores, among Norwegian wild red foxes shot during a hunting season, assemblages A and B were found in five and two foxes, respectively (109). Among seven coyotes positive for G. duodenalis in the United States, one had assemblage B, three had assemblage C, and three had assemblage D isolates (268). In another study of Giardia infection in coyotes in Canada, assemblages A, C, and D were reported (258). Most of the 30 specimens from 30 captive and wild African painted dogs had assemblage B or A and B, with assemblages C and D rarely detected (20a). Since host-specific genotypes are not frequently seen in wildlife, it is unclear whether dogs were the likely contamination sources for assemblage C and D infection in some wild carnivores. The sources of assemblages A and B in wild mammals are also not clear.
In marine animals, assemblage B was found in a thresher shark, and mixed infection of assemblages A and B was found in a Mako shark in the United States (150, 151). Although assemblage A was found in most dolphins and porpoises studied, mixed infections of assemblages A and B were also found in some of them (150, 151). Assemblages A, B, D, and H were found in seals (18, 86, 150, 151). In shore birds in the same marine system, almost all six gulls and three eiders studied had mixed infection of assemblages A and B (150), and one gull had mixed infection of assemblages B and H (151). Subtyping results indicate that either subassemblage AI or AII or mixed infection of both was found for most marine animals (37) (Table (Table1010).
In one of the few studies conducted with marsupials, 41 isolates from various animal species had assemblage A and 8 had assemblage B found (255).
It appears that wild animals are commonly infected with zoonotic assemblages A and B (Table (Table10).10). Unlike Cryptosporidium spp., where many host-specific Cryptosporidium genotypes were found in wild animals (80), the numbers of host-specific Giardia assemblages or genotypes found in wild animals are quite limited, with assemblage H being found only in seals and gulls and the quenda genotype being found only in quendas. This is obviously due to the wide host range of assemblages A and B (Table (Table5)5) and highlights the potential role of wildlife in maintaining the zoonotic transmission cycle of giardiasis. Nevertheless, a recent MLG study indicated that assemblages A and B from captive mammals in a Croatian zoo are genetically different from isolates of human and domestic animal origins (26a).
An analysis of over 4,000 human isolates from different geographical locations by PCR amplification of DNA extracted directly from feces demonstrates that almost exclusively, only G. duodenalis assemblages A and B are associated with human infections (Tables (Tables11 and and2).2). The distribution of these two assemblages in humans varies among studies, sometimes within the same country (42). Assemblage B seems to be slightly more common in both developing (708 cases) and developed (1,589 cases) countries than assemblage A (482 and 1,096 cases, respectively) (Tables (Tables11 and and2).2). The number of molecular epidemiological studies of giardiasis in humans is still small and does not allow a clear detection of geographic or socioeconomic differences in the distribution of assemblages A and B or an assessment of the role of anthroponotic and zoonotic infections in human giardiasis.
Canine-specific assemblages C and D were reported for humans in Thailand in one study (264), but this might be the result of a poor genetic resolution of the SSU rRNA gene PCR products. A previous study by that same group identified assemblages C and D in humans in India using sequence analysis of the SSU rRNA gene. However, only assemblages A and B were detected by sequence analyses of the tpi and ef1α genes (265). Mixed infections of assemblages A and F were recently identified for seven humans in Ethiopia (87), but the presence of assemblage F could not be confirmed by sequence analyses of the SSU rRNA, gdh, or tpi gene. Likewise, the recent finding of assemblage E in three human cases by tpi gene sequencing needs confirmation by analyses of other loci (83). In a more recent study conducted in Europe, assemblage C was found in two human specimens, and assemblages D to F were each found in four specimens using sequence analysis of the bg, gdh, or tpi gene (247).
Subtyping results indicated that both subgroups AI and AII were found in humans, with subgroup AII being more common (Table (Table9).9). Many of these studies targeted only one or two loci; thus, it is not possible to accurately assign some of the less common sequence types to subassemblages in the context of MLGs. There are many more subtypes in assemblage B, and multiple subtypes are usually present in humans in one study area, although the significance of this is not clear (247).
Few studies have investigated the association between assemblage occurrence and the age of patients. In one study of 321 persons between 2 and 76 years old, children ≤12 years of age were at a higher risk of infection with assemblage B (165). The association between assemblage B infection and young age raises the question of whether the distribution of Giardia assemblages is affected by age. A comparison of data from studies in the Philippines, Ethiopia, Australia, and Spain showed that children were susceptible to both assemblages, with variable assemblage distributions among countries (87, 213, 228). The reasons behind the geographic variations in the distribution of the G. duodenalis assemblages are still unclear. It may be explained by the difference in transmission routes and infection sources. It is possible that assemblage A, with a wide range of animals as reservoir hosts, is more likely responsible for zoonotic transmission.
Several molecular epidemiological studies carried out in Ethiopia and the Philippines showed no significant difference between assemblage distribution and gender (87). However, a recent study in Malaysia showed that females were at a 2-fold-higher risk of acquiring giardiasis caused by assemblage B than males (165). Female gender as a risk factor for giardiasis could be related to the women's role as caretakers of children and direct contact with infected children and changing diapers. Nursing has been identified as a risk factor for giardiasis in communities where the infection rate is high in children (76). Thus, there might be some differences in the transmission routes of human giardiasis between males and females living in the same area.
The transmission of G. duodenalis through drinking and recreational waters is well documented (139). Despite this, little is known about the Giardia spp. or G. duodenalis assemblages present in water (240). As expected, assemblages A and E were detected in wastewater samples taken from a slaughterhouse (32). In contrast, only assemblages A and B have been found in urban wastewater samples examined to date, supporting a contamination of human origin (13a, 32, 38, 222, 223, 250). Not surprisingly, assemblage A was recently found in surface water in Portugal, Hungary, and Malaysia (157, 159, 204). There are no data available to make epidemiological linkages between Giardia genotypes/subtypes in source and/or drinking water and human giardiasis.
Data on G. duodenalis genotypes and subtypes in outbreaks of human giardiasis are very limited. Only one person-to-person outbreak of giardiasis was investigated by genotyping. In April 2000, a nursery outbreak of giardiasis occurred in North Wales, United Kingdom, where children, child care workers, and parents had confirmed giardiasis. An analysis of 21 tpi PCR-positive specimens from the outbreak showed the presence of only G. duodenalis assemblage B. In contrast, an analysis of 35 specimens from sporadic giardiasis cases in England and Wales identified subgroup AII in nine specimens, assemblage B in 21 specimens, and both subgroup AII and assemblage B in three specimens (14).
One travel-associated giardiasis outbreak was investigated by genotyping. During 28 September to 4 October 2008, 19 travelers from Illinois, New Mexico, South Dakota, and Wisconsin embarked on a barge trip down the Rhone River in south France. Eighteen of the passengers (97%) reported symptoms consistent with giardiasis. Of the cases reported to the four states, 10 tested positive for Giardia. Specimens from two patients were genotyped as belonging to assemblage B. They produced tpi sequences identical to a subtype commonly found in several countries (strains reported under GenBank accession numbers AY368171 [raw wastewater from Milwaukee, WI], EU518582 [eider in Nantucket Island, MA], EU518572 and EU518561 [dolphins in Cape Cod, MA], and EU518566 [seal in Cape Cod, MA] and strains BAH34, AD-19, and M68, originating from Australian and Belgian residents) (L. Xiao, unpublished data).
Numerous food-borne outbreaks of giardiasis have been reported over the years. Unfortunately, only one food-borne outbreak was investigated by genotyping. The outbreak occurred in >30 school and church staff members who had lunch in the same restaurant in San Francisco, CA, in September 2001. Stool testing of restaurant employees identified three asymptomatically infected food handlers, including a cook, a waiter, and a counter assistant. Only five formalin-fixed stool specimens from school employees were available for molecular diagnosis, two of which were identified as having G. duodenalis assemblage B. They belonged to a common tpi subtype occurring in sporadic cases of giardiasis in India and Peru (249).
Only several waterborne outbreaks of giardiasis were investigated with genotyping tools. One was a large drinking water-associated outbreak during autumn and winter 2004 in Bergen, Norway, in which over 1,500 patients were diagnosed with giardiasis. One particular septic tank was thought to be the possible source of contamination. Analysis of specimens from the outbreak showed that Giardia cysts in the suspected septic tank belonged to assemblage A, whereas those in 21 patients belonged to assemblage B. Thus, the septic tank was exonerated as the contamination source. Sewage leakage from a residential area was subsequently considered the probable source of contamination (223).
Assemblage B was responsible for a waterborne outbreak of giardiasis in California in 2007, which occurred among attendees of a private recreational camp, with 26 laboratory-confirmed cases and 24 probable cases. A retrospective cohort study determined that showering was associated with illness. The camp had installed a slow-sand water filtration system 2 days before the outbreak. A review of historical water quality data identified substantially elevated total coliform and turbidity levels in sand-filtered spring water used for showering during the suspected exposure period. Unfiltered spring water tested at the same time had acceptable coliform and turbidity levels, implicating the filtration system as the most likely contamination source. Two PCR-positive stool specimens produced identical tpi sequences of assemblage B. The sequences obtained from the two patients were mostly identical to the WB6, S7, and B6 subtypes (GenBank accession numbers AY368167, AY228634, and GU564284), with only one nucleotide substitution (T to C) in the 530-bp region under analysis (140).
Assemblage B was responsible for another outbreak of giardiasis associated with a community drinking water system in New Hampshire in 2007, with at least 31 ill persons (59). The assemblage B subtype found in three human cases from this outbreak had substantial sequence differences in the tpi gene (five or more nucleotide substitutions within the 530-bp PCR target) from other known subtypes in humans but was identical to a subtype previously found in beavers in Massachusetts (GenBank accession number DQ789113) and a Barbary macaque in Italy (accession number EU637589) (see Fig. S7 in the supplemental material). Because specimens from two adult beavers near the water source were negative for Giardia, whether the beaver was the source of Giardia cysts could not be confirmed.
The only known outbreak of giardiasis involving assemblage A was a waterborne outbreak that occurred in Temagami, Ontario, Canada, in the spring of 1994 and was characterized by water Giardia cyst concentrations 2 to 3 orders of magnitude higher than normal and a significant increase in the number of cases of giardiasis in the community. Of the 10 human specimens analyzed, 6 had assemblage A isolates, whereas the remainder could not be amplified (283). Human sewage contamination was thought to be the likely contamination source.
Clinical manifestations of giardiasis are quite variable, ranging from the absence of symptoms to acute or chronic diarrhea, dehydration, abdominal pain, nausea, vomiting, and weight loss (221). The severity of disease is probably determined by the interplay between the virulence of the parasite, the developmental, nutritional, and immunological status of the host, the nature of intestinal microflora, and the presence or absence of other copathogens. Although different G. duodenalis assemblages may produce different toxins or metabolic products that may contribute to their pathogenicity (274), studies of the possible association between G. duodenalis assemblages and virulence (indicated by the likelihood of causing diarrhea and other clinical symptoms, as many G. duodenalis infections are asymptomatic) have thus far produced inconsistent results.
Some studies reported that assemblage A isolates were more virulent than assemblage B isolates. Assemblage A is often associated with the presence of symptoms, while assemblage B is not. Thus, in one case-control study of 322 giardiasis patients in Dhaka, Bangladesh, assemblage A, particularly subgroup AII, was strongly associated with the occurrence of diarrhea (111). In 267 cases of giardiasis in southwest London studied, both assemblages A and B caused similar illnesses, but assemblage A was significantly more frequently associated with fever than assemblage B, and all assemblage A strains subtyped were of subgroup AII (35). In contrast, assemblage B was more often associated with asymptomatic infection. A similar association was found for children less than 5 years of age in Spain; subassemblage AII was associated with symptomatic infection, and assemblage B was associated with asymptomatic infection (228). In a longitudinal study of giardiasis in children under 5 years of age in Australia, children infected with assemblage A isolates were 26 times more likely to have diarrhea than those infected with assemblage B isolates (213). In agreement with these observations, in two studies of giardiasis in children less than 9 years old conducted in Peru and Turkey, assemblage A was preferentially found in symptomatic cases, whereas assemblage B was preferentially found in asymptomatic cases (21, 203).
The opposite conclusion was drawn from some other studies. In a study of giardiasis in persons of 8 to 60 years of age in the Netherlands, there was an association of assemblage A with intermittent diarrhea and an association of assemblage B with duodenal inflammation, nausea, and persistent symptoms (117). In a study of giardiasis in persons 2 to 76 years of age in Malaysia, assemblage B infection was significantly correlated with clinical symptoms of giardiasis (165). Likewise, in elementary school children in Saudi Arabia, all infections with assemblage B were symptomatic, while only subassemblages AI and AII were found in asymptomatic infections (12). It was also noticed that symptomatic giardiasis was more significantly associated with assemblage B in children and adults in Ethiopia (87) and elementary school children in Cuba (202).
One factor that was not considered was the intra-assemblage variation within assemblages, which could possibly account for the differences between the studies (180). Furthermore, it was suggested that in regions where a genotype is endemic, a new genotype might cause particularly severe symptoms when it first appears in the population, and dual infection with two different genotypes might produce a synergistic increase in pathology (221). Other parasite factors (such as the rate of multiplication, variable surface proteins expressed, resistance to pharmaceuticals, and ability to invade immune response) and their interplay with host factors has also been suggested to contribute to the pathophysiology observed for clinical giardiasis (221, 274). In addition, infection pressure may be one of the reasons for the different observations of the virulences of assemblages A and B in humans, as the infection dose and dominant G. duodenalis assemblage vary with different socioeconomic development and hygiene practices. On the other hand, host factors such as immune status and age might have played a role in the differences in observations of the virulences of assemblages A and B in humans. It was reported that symptomatic giardiasis was confined mainly to the young and the elderly (228, 274). The susceptibility of these age groups is believed to be the immaturity of the immune system in the former and the immune incompetence in the latter (100). Other potential host factors include history of exposure, diet, and concomitant intestinal microbiota (221).
The use of molecular diagnostic tools has significantly changed our understanding of the zoonotic potential of Giardia spp. in various animals. Evidence accumulated over the past 2 decades has firmly established giardiasis as a zoonotic disease. We are beginning to use these tools to answer questions regarding the transmission of giardiasis in humans and domestic animals. Despite these progresses, we are still facing daunting challenges, especially regarding the disease burden of zoonotic infections, roles of various farm and companion animals in the transmission of human giardiasis, source and human-infective potential of Giardia cysts in source water and drinking water, and differences in biology, clinical manifestations, and outbreak potentials among G. duodenalis genotypes and subtypes. These issues can be addressed effectively only through improvements in molecular diagnostic tools, the more systematic use of these tools in well-designed epidemiological investigations in both endemic and epidemic settings, and a better understanding of the population genetics of G. duodenalis in various hosts under different socioeconomic and environmental conditions.
Currently, on the diagnosis front, MLG tools are used increasingly in characterizing G. duodenalis infections of humans and animals and comparisons of giardiasis transmission between them. This is especially important for resolving the question of the infectivity of some host-adapted assemblages (such as assemblages C, D, E, and F) in humans and in assessing the human-infective potential of assemblage A from animals and the disease burden of zoonotic infections. The latter is hampered by the lack of standardized subtype nomenclature and subtyping procedures, which would require coordinated efforts in thorough characterizations of a large number of reference strains and field isolates using multiple genetic markers and primer sets. New polymorphic markers are probably needed for subtyping isolates within assemblage A, as current tools can differentiate only a few subtypes within subassemblages AI, AII, and AIII. In contrast, other typing strategies may be needed for assemblage B because of the high genetic heterogeneity among isolates in most markers and the presence of apparent heterozygosity at some genetic loci (84, 144a, 154, 247, 254). This can include the use of whole-genome analysis employing the next-generation sequence technologies and single-nucleotide polymorphism (SNP) arrays.
A thorough understanding of the transmission of giardiasis in humans and animals also requires better characterizations of the population genetics of G. duodenalis. Thus far, studies of the population genetics of G. duodenalis are restricted largely to determinations of the presence of genetic recombination (57, 151a, 247, 254). Many other aspects of the population genetics of the parasites have not been addressed, such as population substructure, geographic and host segregations, and selection and dispersal of genetic lineages. Although urgently required, discriminatory subtyping tools based on microsatellites and microsatellite targets that are not under selection pressure are not currently available for Giardia. Sequences of the whole genome are now available for assemblages A, B, and E of G. duodenalis (6, 84, 135a, 276). It is now possible to “mine” the genomic databases to identify targets that are likely to show high-level sequence polymorphism. These population genetic tools would also be useful in addressing epidemiological issues related to the infection sources in humans and wildlife; the genetic basis for virulence and drug resistance; the temporal and geographic dispersal of assemblages A and B in humans, domestic animals, and wildlife; and the absence of assemblage E in some hosts (such as wild ruminants) and geographic areas (such as New Zealand).
Our understanding of giardiasis transmission can be improved through the systematic use of molecular diagnostic tools in well-designed epidemiological studies. The assessment of the disease burden attributable to zoonotic parasites requires the case-control study design, extensive collection of epidemiological data using questionnaires, subtyping of specimens from both humans and animals, and more sophisticated statistical analyses (such as logistic regression and modeling) of research data. Likewise, the identification of the transmission of giardiasis between animals and humans requires longitudinal follow-up and subtyping of both humans and animals in the same focus of endemicity, which allows the establishment of the sequence of infections in humans and animals. Only through the integration of molecular diagnostic and epidemiological tools can we improve our understanding of the importance of zoonotic transmission in the epidemiology of human giardiasis and the transmission dynamics of giardiasis in various socioeconomic and geographic settings.
This work was supported in part by the National Natural Science Foundation of China (grant no. 30771881, 30928019, and 81041078); fundamental research funds for the Central Universities, China (grant no. WB0914044); the Shanghai Science and Technology Committee (grant no. 09540704400); and the State Key Laboratory of Veterinary Etiological Biology at the Lanzhou Veterinary Research Institute.
The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
Yaoyu Feng obtained her B.S. and M.S. at Nankai University, China. After working at Tianjin University, China, as a lecturer for 6 years, she obtained her Ph.D. degree in 1999. She received postdoctoral training and worked as a research fellow at the National University of Singapore. During her work at Tongji University, China, as a professor between 2004 and 2007, she was a guest researcher at the Centers for Disease Control and Prevention as an International EID Fellow for 1 year. She is currently a professor at the East China University of Science and Technology, China. Dr. Feng's earlier research interest was mainly on applied microbiology. Since 2003, she has been working on the molecular epidemiology, pathogenesis, transmission, and environmental ecology of waterborne and food-borne pathogens, such as Cryptosporidium, Giardia, and microsporidia. Dr. Feng has published over 50 original papers, invited reviews, and book chapters.
Lihua Xiao received his veterinary education in China. After his M.S. in veterinary parasitology and 2 years of teaching at the Northeast Agricultural University in China, he did his Ph.D. work in veterinary parasitology at the University of Maine and received postdoctoral training at The Ohio State University College of Veterinary Medicine. In 1993, he moved to the Centers for Disease Control and Prevention, first as a guest researcher and then a senior staff scientist. Dr. Xiao's earlier research interests were in the epidemiology and pathogenesis of gastrointestinal nematodes, cryptosporidiosis and giardiasis of farm animals, immunopathogenesis of and vaccine development for malaria, and immunopathogenesis of HIV. For the last 15 years, he has largely focused on the taxonomy, molecular epidemiology, and environmental biology of Cryptosporidium, Giardia, microsporidia, and other zoonotic protists in humans and animals. Dr. Xiao has published over 200 original papers, invited reviews, and book chapters.
†Supplemental material for this article may be found at http://cmr.asm.org/.