Treatment requirements for Australian source waters to meet health-based target. WaterRA Project 1036

Transcription

1 Treatment requirements for Australian source waters to meet health-based target WaterRA Project 1036 Appendix 5 Review on Cryptosporidium species and shedding rates in animals in Australian catchments September 2014 Una Ryan and Paul Monis

2 EXECUTIVE SUMMARY Managing Cryptosporidium represents the major public health concern of water utilities in developed nations. Of the waterborne protozoan parasitic outbreaks that have been reported worldwide between 2004 and 2010, Cryptosporidium was the identified etiological agent in 60.3% (120) of the outbreaks. Risk assessment and management is essential because every water quality incident that is avoided through better management saves $100 million, and improved levels of community trust can avoid $300 million in unnecessary expenditure. The current Australian Drinking Water Guidelines (ADWG) do not have numerical targets for microbial safety of drinking water but it is anticipated that the target level for waterborne microbial risk (Cryptosporidium and Giardia) will be set at 1 microdaly per person/year, which equates to a log 10 reduction of 6.15 required to meet annual 1 microdaly health targets. As a result of this, accurate modelling and risk assessments will become even more important to the water industry To date, six Cryptosporidium species have been identified in humans in Australia; C. hominis, C. parvum, C. meleagridis, C. fayeri (from marsupials) and C. bovis and C. andersoni (from cattle). The most prevalent species in humans across Australia is C. hominis (~70-80% across different states), followed by C. parvum (~20-30%) and C. meleagridis (~1%). The remaining species have been identified in one or two individuals in New South Wales (NSW). This information is, however, based on very limited sampling data and more human cases need to be screened to determine a more accurate understanding of the prevalence and human-infectivity of other Cryptosporidium species. Key animal hosts in Australian catchments include various wildlife such as rodents, rabbits, marsupials, feral pigs and deer and livestock such as cattle (dairy, beef, intensive feedlot) sheep and pigs (intensive piggeries). Limited studies have been conducted in Australia on the prevalence of Cryptosporidium species and genotypes in these hosts and data indicates that wildlife and livestock are host to a wide range of Cryptosporidium species and genotypes, only some of which are known to be infectious to humans. Of these, cattle and sheep (and to a lesser extent marsupials and rabbits) are likely to constitute the main public health threat due to their prevalence in catchments, the prevalence of potentially human-infectious species in these hosts, the volume of their faecal output and their oocyst shedding rates. There are limited reports on the environmental loading of Cryptosporidium as a result of faecal contamination by wildlife and domestic animals. Both adult and juvenile animals shed Cryptosporidium oocysts with highest levels usually produced by juvenile animals. There is, however, insufficient data on Cryptosporidium oocyst loads in faecal samples from animals in catchments in Australia (particularly from wildlife) to allow for the estimation of human pathogen loads or human exposure levels based on host source, genotype or infectivity of oocysts. Other known or highly probable variables include catchment/site features (e.g. reservoir bathymetry, slope, vegetation and distance to water), host factors (including sample condition, host species, host numbers, breed of animal, age of animal, density and stocking rates, stock management practices) and meteorological parameters (e.g., hydrodynamics, solar UV index, season, rainfall, temperature etc.). There is a lack of understanding on the range of Cryptosporidium species infecting various hosts. As can be seen from Table 2, a wide variety of Cryptosporidium species and genotypes infect vertebrate hosts, only some of which are infectious to humans. Host specificity in Cryptosporidium is not absolute and many species have major hosts in which they are commonly found and minor hosts in which they are occasionally detected. The human-infectivity potential of many Cryptosporidium species is currently unknown but should not be ignored. Sub-typing analysis has revealed that not all C. parvum detected in humans and catchments comes from animal sources, as some C. parvum subtypes are exclusively transmitted between humans. Specific knowledge gaps include the following: Which species infect humans in Australia? o Existing studies are based on small sample sizes and are likely to have only detected the most common causes of clinical infection. ii Water Research Australia Ltd Final Report 1036 Appendix 5

3 o o APPENDIX 5 REVIEW ON CRYPTOSPORIDIUM SPECIES AND SHEDDING RATES More intensive sampling, including for different seasons, is required to determine which species infect humans in Australia (apart from C. parvum and C. hominis) and whether there is any difference in epidemiology for the different human-infective species. An alternative or adjunct to screening clinical cases is to analyse wastewater samples for oocyst loads and genotype to determine general changes in the infection rate in a population and to determine the presence of different species of Cryptosporidium. What is the prevalence and oocyst shedding rates of human-infectious Cryptosporidium sp. in young calves (both dairy and beef) and lambs across Australia? What is the prevalence and oocyst shedding rates of C. cuniculus and other Cryptosporidium genotypes in wild rabbits in Australia? o C. cuniculus is the only other species of Cryptosporidium to have caused a waterborne outbreak (in the UK) and has been detected in Australian surface waters What is the contribution of marsupials and other wildlife to oocyst loads in catchments, are these animals hosts to human-infectious Cryptosporidium? What are the true host ranges of Cryptosporidium species infecting domesticated and wild animals not widely studied to date (i.e. deer, goats, feral pigs etc.)? What is the infectivity of Cryptosporidium oocysts washed into drinking water sources following different scenarios (e.g. storm events following high temperature, low rainfall etc.) and can this be modelled? Do different species of Cryptosporidium have different sensitivities (inactivation rates) to environmental conditions? Are there differences in infectious dose for different humaninfectious Cryptosporidium species, o Two sub-type families of C. cuniculus have been reported to cause infection in humans in the UK, although the Va sub-type caused the outbreak reported there. Both sub-types families (Va and Vb) have been detected in Australian surface waters. Infectious C. meleagridis, responsible for approximately 1% of human infections in Australia, has also been detected in surface waters. Are fate and transport models generally applicable? o Accurate fate and transport predictions during wet weather events including accounting for preceding catchment loading (deposition less inactivation during dieoff) and transport from catchment to waterway are essential to estimate worst-case scenarios for oocyst loads All the above are essential for accurate quantitative system modelling and QMRA analysis to determine the possible challenges to water treatment plants and the relative pros and cons of catchment management initiatives can be evaluated. Water Research Australia Ltd Final Report 1036 Appendix 5 iii

4 TABLE OF CONTENTS Executive Summary... ii 1. Introduction Importance of managing pathogen risks Australian catchments What Cryptosporidium species are found in humans in Australia? What Cryptosporidium species occur in catchments and what are their host species? Animal population density and volume of manure Wildlife Population density Population density of domestic livestock Volume of wildlife manure Volume of domestic livestock manure Cryptosporidium prevalence and oocyst shedding rates Factors affecting the load of human-infective oocysts in source waters Common misconceptions Research gaps References FIGURES Figure 1. Cryptosporidium transmission... 4 Figure 2. kg manure produced/day by various animals... 8 TABLES Table 1. Valid species within the genus Cryptosporidium... 2 Table 2. Prevalence of Cryptosporidium species and genotypes in animals in Australian catchments. 5 Table 3. Estimates of wildlife animal density for native vegetation land use areas... 7 Table 4. Livestock numbers by State Table 5. Manure production rates for wildlife animals (wet weight) Table 6. Oocyst shedding rates for various hosts Table 7. Oocyst inactivation times for different temperatures measured using a cell culture infectivity model (King et al., 2005 and Keegan et al., 2008) Table 8. Time (hours) required to achieve 90% reduction in oocyst infectivity for different UV index days in different water matrices (data from King et al., 2008) iv Water Research Australia Ltd Final Report 1036 Appendix 5

5 1. INTRODUCTION Cryptosporidium spp. are protozoan parasites that infect a wide range of vertebrate hosts including humans (Xiao, 2010). The parasite causes self-limiting diarrhoea in immunocompetent individuals but may be chronic and life-threatening to those that are immunocompromised (Hunter et al., 2007). The incubation period for illness symptoms is approximately 7 days (range 1-14 days) and illness is usually self-limiting, with a mean duration of 6-9 days (Hunter et al., 2007). Relapses are common; reports indicate 1-5 additional episodes in 40-70% of patients (Hunter et al., 2004). The predominant symptom is watery diarrhoea, sometimes profuse (1-2 L/day in a small minority of cases, usually very young or old), sometimes mucous but rarely bloody. Humans can acquire Cryptosporidium infections through various transmission routes, such as direct contact with infected persons (person-to-person transmission) or animals (zoonotic transmission) or ingestion of contaminated food (foodborne transmission) or water (waterborne transmission) (Xiao, 2010). To date a total of 26 species of Cryptosporidium have been identified, including the recently described C. viatorum in humans, C. scrofarum (previously pig genotype II) in pigs and C. erinacei in hedgehogs (Table 1). There are also over 40 genotypes, with a high probability that many of these will eventually be given species status with increased molecular characterisation. Understanding the transmission dynamics of Cryptosporidium has traditionally been difficult because most species of Cryptosporidium are morphologically identical. Therefore, alternative molecular characterisation tools such as PCR and DNA sequence analysis have been required to reliably differentiate / identify species and genotypes of Cryptosporidium. The 18S ribosomal RNA (rrna) gene and the hypervariable 60-kDa glycoprotein (gp60) gene have been widely used as targets to identify species and track transmission (Xiao, 2010; Ng et al., 2008; Ng et al., 2010a). A major mode of transmission of Cryptosporidium is via contaminated water, including drinking water and recreational waters. Oocysts, which are the infectious life-cycle stage of Cryptosporidium, are environmentally stable, resistant to inactivation by commonly used disinfectants such as chlorine and able to survive and penetrate water treatment and wastewater treatment processes (Fayer et al., 2000). Currently, cryptosporidiosis represents a major public health concern of water utilities in developed nations (Fayer et al., 2000). This is because, compared to other protozoans that are commonly detected in water (Giardia, Entamoeba, Toxoplasma, Balantidium and Cyclospora), Cryptosporidium is the most persistent in the environment, most resistant to chemical disinfection, and smallest in size, so the most difficult to remove by filtration. Cryptosporidium was the etiological agent in 60.3% (120) of the waterborne protozoan parasitic outbreaks that have been reported worldwide between 2004 and 2010, (Baldursson and Karanis, 2011). Oocyst transport to surface water can occur by direct deposition of faeces into the water or from runoff from contaminated surfaces. Humans, wildlife and domestic livestock all potentially contribute Cryptosporidium to surface waters. Current standard methods for monitoring Cryptosporidium in water measure total Cryptosporidium and do not differentiate species or determine if the detected oocysts are infectious or have been inactivated by exposure to environmental conditions. There is therefore a significant knowledge gap regarding the load of human infectious Cryptosporidium in source waters and the contribution of animal hosts to this load. Identification of the sources/carriers of human pathogenic species of Cryptosporidium is essential for accurate risk assessment and catchment management. The mechanisms of host specificity remain unknown, but recent research indicates that many Cryptosporidium spp. are less host-specific than initially thought. This milestone examines the current information from Australia (and where necessary, with data from studies performed in other regions, primarily Europe and the USA) on faecal output, oocyst shedding rates, prevalence and genotypes of Cryptosporidium identified in wildlife and domestic livestock. Water Research Australia Ltd Final Report 1036 Appendix 5 1

6 Table 1. Valid species within the genus Cryptosporidium Species Major host Minor host Site of infection Reference C. parvum Cattle/sheep/humans Deer/mice/pigs Small intestine Tyzzer, 1912 C. hominis Humans Dugong/sheep Small intestine Morgan-Ryan et al., 2002 C. viatorum Humans - Small intestine Elwin et al., 2012 C. meleagridis Turkey/humans Parrots, chickens Small intestine Slavin, 1955 C. wrairi Guinea pigs - Small intestine Vetterling et al., 1971 C. canis Dogs Humans Small intestine Fayer et al., 2001 C. felis Cats Humans/Cattle Small intestine Iseki, 1979 C. macropodum Marsupials - Small intestine Power and Ryan, 2008 C. fayeri Marsupials Humans Small intestine Ryan et al., 2008 C. ubiquitum Sheep Cattle/humans/wildlife. Small intestine Fayer et al., 2010a. C. ryanae Cattle - Small Intestine Fayer et al., 2008 C. bovis Cattle Sheep Small intestine Fayer et al., 2005 C. suis Pigs Humans Small intestine Ryan et al., 2004 C. scrofarum Pigs Humans Small Intestine Kváč et al., 2012 C. xiaoi Sheep Yak, goat Small intestine Fayer and Santin, 2009 C. cuniculus Rabbits Humans Small intestine Robinson et al., 2010 C. tyzzeri Mice - Small intestine Ren et al., 2012 C. erinacei European hedgehog Humans Small intestine Kváč et al., 2014 C. molnari Fish - Stomach/ Alvarez-Pellitero and intestine Sitja-Bobadilla, 2002 C. fragile Frogs - Stomach Jirku et al C. varanii Lizards Snakes Stomach/ Pavlasek et al., 1995 intestine C. serpentis Lizards/snakes - Stomach Levine, 1980 C. muris Rodents Ungulates/humans/marsupials Stomach Tyzzer, 1907 C. andersoni Cattle Sheep Abomasum Lindsay et al., 2000 C. galli Birds - Proventriculus Pavlasek, 1999; Ryan et al., 2003a C. baileyi Poultry Quails/Ostriches Bursa Current et al., Importance of managing pathogen risks Every water quality incident that is avoided through better management potentially saves at least $100 million; and improved levels of community trust can avoid $300 million in unnecessary expenditure (CRC, 2008). For example, in 1993 in Milwaukie, WI, USA, ~ 403,000 individuals became infected due to contamination of the public water supply and 100 people died at an estimated illnessassociated cost of $96.2M (Corso et al., 2003). The Sydney water contamination incident in 1998, where high concentrations of Cryptosporidium were detected in the city's water supply and distribution systems, cost Sydney Water, $75M (Cox et al., 2003). The incident was a highly publicised event, which even made the front page of the New York Times. However, the direct costs borne by Sydney Water were only a fraction of the cost borne by the community in what has been defined as averting behaviours i.e. household and business costs directly linked to behaviours associated with losing confidence in the water supply. A study done for the Department of Human Services in Victoria estimated these costs for the Sydney event to be $308.1 M (CRC, 2008). Clearly, water quality incidents are the major risk facing any operating water authority and are at least equivalent to the crash of a jumbo jet for a major airline in terms of the severity of the incident on reducing public confidence. The Australian Drinking Water Guidelines (ADWG), provide a framework for good management of drinking water supplies based on a preventive risk management approach (NHMRC-NRMMC, 2011). Risk management is based on assessing risks and reducing them to acceptable levels to assure safety. The ADWG currently do not have numerical targets for microbial safety of drinking water but it is anticipated that this discrepancy will soon be addressed and that a combination of quantitative 2 Water Research Australia Ltd Final Report 1036 Appendix 5

7 microbial risk assessment (QMRA) and the metric of Disability Adjusted Life Years (DALYs) will be used to define microbial safety. QMRA is applied to determine the likelihood of infection and illness occurring from exposure to specific pathogens contained in water, while DALYs are used to convert the likelihood of illness into impacts or burdens of disease. It is anticipated that the target level for waterborne microbial risk (Cryptosporidium and Giardia) will be 1 microdaly per person/year. A highly impacted catchment, with a source water containing 20 oocysts/l, would require a log 10 reduction of 6.15 required to meet the annual 1 microdaly health target (Sinclair, 2012). Bearing this in mind, accurate monitoring, modelling and risk assessments will become even more important to utilities to ensure that sufficient data are available to characterise the human health risk from Cryptosporidium and that adequate removal targets are selected for each source water. 1.2 Australian catchments Australian catchments consist of so-called protected catchments that have various wildlife including rodents, rabbits, marsupials, feral pigs and deer, emus etc. and agricultural catchments that contain dairy cattle, sheep, beef cattle, intensive feedlot cattle and intensive piggeries. In these catchments, beef cattle and sheep are often free-ranging with calves and lambs wandering around with the adults. For dairy cattle, calves may or may not be free-ranging and are sometimes kept in pens or paddocks of their own when young. The seasonality of calving varies across different states to one or more specific seasons. Septic systems from rural living areas in agricultural areas can also contribute sources of Cryptosporidium. Some catchments, particularly those in the Adelaide hills area, also have urban development in addition to agriculture, adding potential sources of Cryptosporidium from septic systems or wastewater treatment plants. A further potential source of contamination is recreational activity, which is permitted in some catchments / water storages. For example in Western Australia (WA), the Water Corporation is the current principal supplier of water. Its operations cover an area of 2.5 million square kilometres (km 2 ), or about one third of the Australian continent, making it one of the world s broadest operating areas for this industry. Across the Water Corporation s 245 localities, pathogen risks to the water supply can come from human activity (illegal activities and approved recreation in the outer catchment), native animals (predominantly marsupials and rabbits), as well as domestic animals of which sheep and cattle (mostly dairy) are the predominant grazing animals in the South Western Agricultural region. Of these, sheep account for ~95% of the animals (Cameron Gordon, Water Corporation). In New South Wales (NSW), more than four million people in Sydney, the Illawarra and the Blue Mountains rely on several individual catchments that cover 16,000 square kilometres (km 2 ) and include 21 storage dams (11 major dams) holding 2,500 billion litres of water. The total estimated human population of the entire catchment is ~150,000. High sheep stocking rates are present in the Wollondilly catchment near Goulburn (approximately 550,000 sheep) and in the Shoalhaven catchment near Braidwood (approximately 109,000 sheep). Beef cattle grazing numbers are consistent throughout the catchments at a stocking density of approximately 25% of the sheep population. Dairy farms are also present, predominantly within the Wingecarribee catchment (approximately 10,500 adult cattle). The Warragamba catchment contains mostly wildlife (marsupials, particularly kangaroos and rabbits) (Cox et al., 2005). WA is experiencing high levels of drought, whilst NSW is subject to highly variable levels of rainfall and storms. As these are key conditions predicted for climate change in the future (Garnaut review, 2008), measuring the infectivity of different Cryptosporidium species under different climatic conditions is crucial for accurate risk assessment. 1.3 What Cryptosporidium species are found in humans in Australia? Humans can acquire Cryptosporidium infections through various transmission routes, such as direct contact with infected persons (person-to-person transmission) or animals (zoonotic transmission) and ingestion of contaminated food (foodborne transmission) or water (waterborne transmission) (Karanis et al., 2007; Xiao, 2010). World-wide, C. hominis and C. parvum are responsible for the majority of clinical infections in humans. However, more than eight other species/genotypes of Cryptosporidium have been identified in human cases of cryptosporidiosis, including C. meleagridis (from birds and humans), C. felis (from cats), C. canis (from dogs), C. ubiquitum (from sheep/deer), C. cuniculus (from rabbits) C. fayeri (from marsupials), C. andersoni and C. bovis (from cattle), C. muris (from rodents) and C. suis and C. scrofarum (from pigs) (Fig. 1) (Xiao, 2010; Chalmers et al., 2011; Ryan and Power, 2012). Prevalence data from the UK suggests that C. meleagridis and C. cuniculus are Water Research Australia Ltd Final Report 1036 Appendix 5 3

8 each responsible for approximately 1% of human clinical cases, whereas the other species combined account for less than 1% of human cases of infection. Figure 1. Cryptosporidium transmission Data from the National Notifiable Diseases Surveillance System (NNDSS) in Australia shows that outbreaks are common, particularly in summer, and that cryptosporidiosis infections are increasing across Australia. For example, in NSW in 2009, 1163 cryptosporidiosis cases were reported, representing a 313% increase in incidence compared to Contaminated public swimming pools were identified as the source of the outbreak and up to 19 pools were superchlorinated as a result (Waldron et al., 2011a). In 2007, there was a four-fold increase in cases in SA (n=448 cases) (Ng et al., 2010a). In WA, there was a three-fold increase (n=608) in 2007 compared to 2006 (Ng et al., 2010a), and another three-fold increase (n=448) in 2011 compared to 2010 (NNDSC). In 2012, Queensland (Qld) received 1,371 notifications compared to 465 the previous year (NNDSC) and the Northern Territory (NT) received 235 notifications compared to 94 the previous year. The incidence of cases in Qld and the NT appears to be independent of season, whereas peak infections occur during summer / early autumn in the other states. In Australia, C. hominis, C. parvum and C. meleagridis have been identified in humans in Western Australia (WA), New South Wales (NSW), Victoria (VIC) and South Australia (SA), with C. hominis being the most frequently identified species of the three (Robertson et al., 2002; Chalmers et al. 2005; Jex et al., 2007; Ng et al. 2008; O Brien et al., 2008, Jex et al., 2008; Alagappan et al. 2008; Waldron et al. 2009a; 2009b; Ng et al., 2010a; 2010b; 2012; Waldron and Power, 2011; Waldron et al., 2011a; 2011b). Recently, C. fayeri, C. andersoni and C. bovis have also been identified in human cases in NSW (Waldron et al., 2010; 2011a; Ng et al., 2012). In 2009, C. fayeri (from marsupials), was identified in a woman in Sydney and identical subtypes were found in marsupials in the area and there have also been several reports of C. parvum and C. hominis in marsupials (Waldron et al., 2011; cf. Ryan and Power, 2012). The most prevalent species in humans across Australia is C. hominis (~70-80%), followed by C. parvum (~20-30%) and C. meleagridis (~ 1%) (Jex et al., 2007; Ng et al., 2008; O Brien et al., 2008, Jex et al., 2008; Alagappan et al. 2008; Waldron et al. 2009a; 2009b; Ng et al., 2010a; 2010b; 2012; Waldron and Power, 2009; 2011; Waldron et al., 2011a; 2011b). A recent study of human clinical samples from South Australia found a similar incidence of C. hominis and C. parvum, failed to detect any C. meleagridis and most interestingly identified an infection caused by C. cuniculus (Ika Sari, PhD thesis unpublished). This finding represents the first detection of C. cuniculus in a human clinical sample outside of the UK. However this result and the 4 Water Research Australia Ltd Final Report 1036 Appendix 5

9 results of previous Australian studies are based on limited sample numbers and more human isolates need to be screened to fully understand the prevalence of various Cryptosporidium species in humans and obtain more accurate infection rates. 1.4 What Cryptosporidium species and hosts occur in catchments? There is no evidence linking contamination of water supplies by animals in catchments to outbreaks of cryptosporidiosis in Australian human populations and there is limited conclusive molecular evidence for outbreaks reported in other parts of the world. However, several studies have strongly linked outbreaks of cryptosporidiosis with sheep and cattle grazing near the implicated reservoirs, catchments or rivers (eg. Yang et al., 2008; Ruecker et al., 2007). One recent waterborne outbreak in the UK was caused by C. cuniculus (from rabbits), although in this instance the contamination was at the water treatment plant rather than as a result of faecal contamination from the catchment (Chalmers et al., 2009; 2011). A recent survey of rainfall-runoff in a multi-use catchment in the Adelaide Hills area detected predominantly animal-associated species, including C. cuniculus (both Va and Vb sub-types), C. fayeri, C. bovis, C. muris, C. ryanae, C. meleagridis, C. tyzzeri and several genotypes (rat genotype, mouse genotype II, novel genotypes) (Swaffer et al. in press). The dominant species detected was C. muris (detected in 20 of 41 samples), followed by C. ryanae and C. bovis. Cryptosporidium parvum was detected in only 3 of 41 samples. Infectivity was also measured using a cell culture infectivity assay and found overall to be 3% for the oocysts analysed. Characterisation of the infectious Cryptosporidium identified C. cuniculus (Vb), C. fayeri, C. meleagridis and C. sp mouse genotype II. No infectious C. parvum was detected. Cryptosporidium hominis was not detected in any of these samples. A survey of faecal samples collected over a 2-year period from 9 catchment locations around a Melbourne reservoir identified multiple species including C. hominis, C. parvum, C. cuniculus (Vb), C. canis, C. fayeri, C. macropodum and C. ubiquitum, as well as 6 novel genotypes distinct from named species or genotypes of Cryptosporidium (Nolan et al., 2012). The following table summarises what is known about the prevalence of Cryptosporidium species and genotypes in animals in Australian catchments. Table 2. Prevalence of Cryptosporidium species and genotypes in animals in Australian catchments. Host Cryptosporidium Species Prevalence in primary host Human infectious Yes/No/Unknown Bilby (Macrotis lagotis) C. muris Unknown** Occasional human pathogen Kangaroo (various C. fayeri 0.32% to Yes - Human infection species) 28.5% in a woman in NSW C. macropodum 0.32% to 28.5% Reference Warren et al., 2003 Morgan et al., 1997; Power et al., 2003; 2004; McCarthy et al., 2008; Waldron et al., 2010; Ng et al. 2011a Yang et al., 2011; Nolan et al., Unknown Power et al., 2003; 2004; Ng et al. 2011a Yang et al., 2011; Nolan et al., C. parvum/c. hominis 15.6% Yes Ng et al. 2011a; Nolan et al., Kangaroo ~4% Unknown Yang et al., 2011 genotype I C. xiaoi ~2% Unknown Yang et al., 2011 Wallaby (no species ID C. parvum 5.9% Yes Ng et al., 2011a Swamp wallaby (Wallabia C. macropodum Unknown** Unknown Ryan et al. unpublished bicolour) Yellow-footed rock wallaby C. fayeri Unknown** Yes Power et al., 2009 (Petrogale xanthopus), Brush tail possum (Trichosuris vulpecula) 11.3% (urban) 5.6% (wild) Hill et al., 2008 Quokka (Setonix brachyurus) Other marsupials Western-barred bandioot Brushtail possum genotype I, C. parvum/hominislike. Brushtail possum genotype I Unknown C. parvum/hominis-like - Yes Kangaroo 4.7% Unknown Austin et al., unpublished genotype I C. fayeri Unknown** Yes Weilinga et al. unpublished Water Research Australia Ltd Final Report 1036 Appendix 5 5

10 Host (Peremeles bougainville) Cryptosporidium Species Prevalence in primary host Human infectious Yes/No/Unknown Reference Long-nosed bandicoot (Perameles nasuta) Southern Brown Bandicoot (Isoodon obesulus) Other Birds Adult dairy cattle C. parvum / hominis like C. galli, avian genotypes I to IV, C. meleagridis, C. hominis C. bovis, C. andersoni Juvenile dairy cattle* C. parvum, C. bovis, C. ryanae, C. ubiquitum, C. suis % Yes Dowle et al % C. meleagridis and C. hominis are known human pathogens 15% Both have been reported in humans in Australia % C. parvum is the second most common species in humans. C. bovis, C. ubiquitum and C. suis are occasional human pathogens. Adult Beef cattle C. andersoni 0-26% Has been reported in humans in Australia Adult sheep C. ubiquitum, C % C. ubiquitum and C. xiaoi, C. parvum + parvum are known low amounts of human pathogens other species Juvenile sheep C. parvum, C. xiaoi, C. ubiquitum, % C. ubiquitum and C. parvum are known human pathogens Pigs C. suis % Occasional human pathogen Deer Rodents Mice and rats C. scrofarum (previously pig genotype II) % Occasional human pathogen Ng et al., 2006; McCarthy et al., 2008; Ferguson, 2010; Ralston et al., Becher et al. 2004; Nolan et al. 2009; Ferguson, 2010; Ng et al. 2011b; Ng et al., 2012 Ralston et al., Ryan et al., 2005; Ferguson, 2010; Sweeney et al., 2011 Yang et al., 2009; Sweeney et al., 2011 Ryan et al., 2003b; Ryan et al., 2004; Johnson et al., 2008 Ryan et al., 2003b; Ryan et al., 2004; Johnson et al., 2008 C. parvum ~25% Yes Morgan et al., 1999a C. ubiquitum C. parvum Similar to environmental sequence AY C. tyzzeri, mouse genotype II and 9% 0-50% 0.7% rat-like genotypes Rabbits C. cuniculus 3-17% ~33% Foxes C. canis and C. macropodum-like genotype Dingoes and wild dogs C. canis + C. hominis-like genotype Feral cats Unknown but C. felis and C. muris reported in domestic cats in Australia Yes Unknown Cinque et al., 2008; Ng et al., 2011a; Nolan et al., % Unknown Morgan et al., 1999b; 1999c; Foo et al., 2007; Paparini et al., % Occasional human pathogen Unknown 22.7% Occasional human pathogen Yes 0% to 29.4% in domestic cats Yes Nolan et al., 2010; Ferguson, 2010; Nolan et al., Occasional human pathogens *C. hominis has been reported in cattle in New Zealand (Abeywardena et al., 2012). Ng et al., 2011a; Nolan et al., Ng et al., 2011a; Morgan et al., 2000 Sargent et al., 1998; Palmer et al., 2008; FitzGerald et al., 2011; Ryan and Power, **Unknown: In these cases only positive samples were referred for analysis and the actual prevalence is not known. 6 Water Research Australia Ltd Final Report 1036 Appendix 5

11 1.5 Animal population density and volume of manure Excretion of faecal material from wildlife and domestic animals is a major source of pathogen contamination of catchments. In order to understand the potential for Cryptosporidium contamination in catchments by wildlife, it is essential to have some understanding of animal density, as animal density by area is an important determinant of pathogen loadings. Higher animal density results in a larger volume of manure excreted per unit area. Thus there is an increase in pathogen source material that may be transported in runoff to surface waters and/or deposited directly into streams (Ferguson, 2010) Wildlife Population density It is difficult to quantify animal densities for wildlife because animal movement is uncontrolled and animal populations vary with season and environmental conditions, with many species being migratory. Published values and estimates are summarised in Table 3. The term wildlife includes both native and feral free-living species of animals. Native animal species are defined as those free-living fauna that are indigenous to a specific geographical area, usually a continent or island. Frequently these species are protected and cannot be kept by humans as companion animals. Feral species can be defined as those free-living fauna species that are not native to a specific geographical area and not confined or managed as domestic animals. These species are frequently regarded as pest species that damage native flora and fauna and can be the subject of hunting and culling activities. Within Australia the legal culling of native animals is also allowed for some species, such as kangaroos, due to their over-abundance in certain regions (Ferguson, 2010). Table 3. Estimates of wildlife animal density for native vegetation land use areas Animal Country Density (per km 2 ) Reference Kangaroos Australia Fletcher, 2007 Kangaroos Australia, Sydney Fergusen, 2010 Kangaroos Australia 245 Coulson, 1985 Kangaroos Australia, Melbourne 178 Ramp, 2002 Marsupials Australia 20 Fergusen, 2010 Pigs (feral) Western Australia 2-3 Dr. Peter Adams, pers comm. Deer Australia 0.5 Fergusen, 2010 Rodents UK Similar to human population O'Keefe, D'Arcy et al., 2003 Rabbits Australia Williams et al., 1995 Foxes Australia Saunders et al., Dogs (feral) Australia 0.25 Fergusen, 2010 Cats (feral) Australia 1-57 Dickman, Population density of domestic livestock An assessment of stock numbers within Australia was obtained from the Australian Bureau of Statistics Agricultural Commodities, Australia, 2009/2010 (Table 4). In 2009/2010, NSW had the highest number of sheep (24.3 million), followed by WA (14.6 million) and Victoria (Vic) (14.3 million). Dairy cattle were reported at 2.5 million for 2009/2010, with Victoria continuing to dominate the dairy industry with 62% of Australia's total dairy herd at 1.5 million. Meat cattle were reported as 24 million in 2009/2010 with the highest number in Qld (11.1 million), followed by NSW (5.1 million), and WA (2.2 million). Pigs were reported as 2.2 million in 2009/2010, with the highest density in NSW at 0.58 million. Water Research Australia Ltd Final Report 1036 Appendix 5 7

12 Table 4. Livestock numbers by State. Animal Total no. NSW Vic Qld SA WA Tas NT ACT Sheep & lambs 68,085,497 24,366,338 14,377,696 3,622,141 8,989,472 14,691,553 1,991, ,010 Dairy cattle Meat cattle 2,542, ,318 1,588, , , ,023 1,916, ,007,730 5,107,062 2,079,529 11,193, ,861 2,206, ,751 2,065,746 6,250 Pigs 2,289, , , , , ,393 11, Volume of wildlife manure Manure excretion rates and volumes for wildlife are less well documented than for domestic animals. An estimate of the volume of manure produced by wildlife, however, is important to assess the impact of wildlife manure on catchments. Estimates of manure production rates for wildlife are shown in Table 5 and Figure 2. Table 5. Manure production rates for wildlife animals (wet weight). Animal Country Kg manure/animal -1 /d -1 Reference Kangaroos Australia, Canberra 0.3 Dr. Peter Adams, pers comm. Pigs (feral) Western Australia 4-5 Dr. Peter Adams, pers comm; Dexter, Deer USA 1-2 Anon, 2003 Rabbits Netherlands Medema, 1999; 2001 Dogs (feral) Australia Anon, Figure 2. Example kg manure produced/day by various animals 8 Water Research Australia Ltd Final Report 1036 Appendix 5

13 1.5.4 Volume of domestic livestock manure Livestock excretion rates and volumes are reasonably well documented compared to those for wildlife. Estimates for manure production rates for domestic livestock were obtained from a 2003 revision of the Manure Production and Characteristics produced by the American Society of Agricultural Engineers (Anon, 2003). The data were combined from a wide base of published and unpublished information on livestock manure production and characterization (Anon, 2003). It has been estimated that a 400-kg adult beef cow will produce on average 23 kg of faeces per day and a 400 kg dairy cow 34.4 kg of faeces per day. A 45 kg adult sheep will produce on average 1.8 kg of faeces per day and a 40 kg pig will produce on average 3.4 kg of faeces per day (Anon, 2003) (Figure 2). 1.6 Cryptosporidium prevalence and oocyst shedding rates Accurate quantification of Cryptosporidium oocysts in animal faecal deposits on land is an essential starting point for estimating catchment Cryptosporidium loads (Davies et al., 2003). There are limited reports, however, on the concentration and environmental loading of waterborne protozoan pathogens as a result of faecal contamination by wildlife and domestic animals. It is also important to note that oocyst recovery rates from faecal samples and across animal types can be highly variable. For example, recovery rates of 14 to 70% for cattle faeces (adults), from 0 to 83% for calf faeces, 4 to 48% for sheep faeces, 40 to 73% for kangaroo faeces, and from 3 to 24% for pig faeces have been reported (Davies et al., 2003). Thus, oocyst shedding rates reported in various studies may underestimate the number of oocysts unless recovery efficiency is factored into the analysis. Research has indicated that both adult and juvenile animals shed Cryptosporidium and the data presented in Table 6 is derived from available literature from Australia, the USA and Europe (Bukhari and Smith, 1997; Hoar et al., 2000; Atwill et al., 2003; Davies et al., 2003; Power et al., 2005; Hill et al., 2008; Nolan et al., 2009; Klein et al., 2010; Ng et al., 2011b; Izzo et al., 2001; Lalancette et al., 2012; Oates et al., 2012; Yang et al., 2013). The prevalence of Cryptosporidium in marsupials varies, as does the oocyst shedding rate. In NSW, the prevalence of Cryptosporidium in faecal samples from eastern grey kangaroos (Macropus giganteus) was 6.3% (239/3,557) (Power et al., 2005). Another study in NSW on common brushtail possums (Trichosurus vulpecula) reported that Cryptosporidium occurred with a higher prevalence in possums from urban habitats (11.3%) than in possums from woodland habitats (5.6%) (Hill et al., 2008). In WA, the prevalence of Cryptosporidium in wild western grey kangaroos (Macropus fuliginosus), was 9.3% (Yang et al., 2011). Little information is available on Cryptosporidium oocyst shedding rates in marsupials. One study on marsupials in New South Wales (NSW) reported that the number of oocysts detected in kangaroo faeces ranged from 1-39,423 oocysts/g faeces (dry weight), with a median of 0.5 oocysts in the majority of samples (Davies et al., 2003). Another study on marsupials in NSW reported that the number of oocysts detected in kangaroo faeces ranged from below 20 to oocysts/g faeces (wet weight) with ~10-50 oocysts in the majority of samples (Power et al., 2005). More information is available for cattle and sheep and although oocyst shedding is higher in young animals, adult cattle and sheep can still shed Cryptosporidium oocysts in significant amounts. In Australia, the prevalence of Cryptosporidium in cattle ranges from % (Becher et al., 2004; Nolan et al., 2009; Ng et al., 2011b; 2012; Izzo et al., 2011). A recent study reported that the total prevalence of Cryptosporidium in calves from 84 dairy and dairy beef properties across Australia was 58.5% (Izzo et al., 2011). In Victoria, the prevalence of Cryptosporidium in faecal samples from 268 individual calves on pasture-based dairy farms in three regions (Northern Victoria, South Gippsland and Western District) was 46.3% (124/268) (Nolan et al., 2009). In newborn calves, excretion of oocysts usually occurs after 7 days and peaks around 14 days. At the peak of the infection, oocysts per gram faeces can be excreted. A study in NSW reported 1-5,988 oocysts/g faeces (dry weight) with a median of 0.5 oocysts for adult cattle and 1-17,467 oocysts/g faeces (dry weight) with a median of 23 oocysts for juvenile cattle (Davies et al., 2003). Other studies have estimated the environmental loading rate of Cryptosporidium in beef cattle at between 3,900 to oocysts cow 1 day 1 (Hoar et al., 2000; Atwill et al., 2003). In eastern Australian cattle feedlot manures, the occurrence of Cryptosporidium and other pathogens was quantitated using quantitative PCR. High counts of Cryptosporidium (>10 5 g-1) were sporadically identified in all manures (Klein et al., 2010). A more recent study reported that the maximum shedding was 3 x 10 7 oocysts beef cow 1 day 1 Water Research Australia Ltd Final Report 1036 Appendix 5 9

14 with a mean shedding of 1,884.8/gram faeces day 1 (Oates et al., 2012). Beef cattle can therefore potentially contribute significantly to contamination of drinking water catchments with Cryptosporidium. For dairy cattle, a recent study has reported that the median concentrations for total oocysts were higher in calves (333 oocysts g 1 ) than in cows (52 oocysts g 1 ) (Lalancette et al., 2012). Of these, infectious oocysts were found in only 7% of the samples as determined by cell culture (Lalancette et al., 2012). Cryptosporidium has been reported in sheep worldwide, however most studies have been based on microscopy and have reported prevalences ranging from 2.6 to 82% for Cryptosporidium (cf. Ryan et al., 2005; Yang et al., 2009, Wang et al., 2010; Sweeny et al., 2011). In Australia, reported prevalences for ewes in Western Australia ranged from % (Sweeney et al., 2011) and for lambs from % on different properties (Ryan et al., 2005; Yang et al., 2009; Sweeny et al., 2011). It has been reported that postparturient ewes may shed increased but low concentrations of Cryptosporidium oocysts (100 to 5,700 oocysts g 1 ) (Xiao et al., 1994). A study by Bukhari and Smith (1997) which examined oocyst excretion and viability patterns (measured by vital dye staining or in vitro excystation) in experimentally infected lambs reported that oocyst shedding ranged from x 10 7 g 1 with a median of median 1.4 x10 6. Oocyst viability ranged from 39-69% (Bukhari and Smith, 1997). Another study reported oocyst shedding rates of 1-152,474 oocysts/g faeces (dry weight), with a median of 275 for adult sheep and oocysts/g faeces (dry weight), with a median of 148 for juvenile sheep (Davies et al., 2003). Recent data generated at Murdoch University using quantitative PCR (qpcr) on total DNA extractions from lamb faecal samples reported similar oocysts excretion rates for both pre and post-weaned lambs with oocyst shedding of 1-50,000 oocysts/g faeces (dry weight) and a median of 15,000 for pre weaned lambs and 1-75,000 oocysts/g faeces (dry weight) and a median of 25,000 for post-weaned lambs (n=500) (Yang et al., 2013). Little is known about the prevalence of Cryptosporidium in feral pigs. A study in California reported that 12 (5.4%) of 221 feral pigs were shedding Cryptosporidium oocysts (Atwill et al., 1997). The authors also reported that younger pigs (< or = 8 months) and pigs from high-density populations (> 2.0 feral pigs/km2) were significantly more likely to shed oocysts compared to older pigs (> 8 months) and pigs from low-density populations (< or = 1.9 feral pigs/km2) (Atwill et al., 1997). This trend makes reduction of feral pig abundance in high-density catchment areas even more important to reduce the risk of waterborne feral pig pathogens being introduced to reservoirs. In Spain, a prevalence of 7.6% was reported for wild boar and infections were significantly higher in juvenile male wild boars (22%) than in adult males (6%) (Castro-Hermida et al., 2011). The mean intensity of infection by Cryptosporidium was 5 to 200 oocysts per gram of faeces (Castro-Hermida et al., 2011). A study in WA reported a prevalence of 0.3% (1/292) (Hampton et al., 2006). Genotyping attempts were unsuccessful. A more recent study of 237 wild pigs in WA did not identify Cryptosporidium by PCR (Pallant et al., unpublished). Studies in Australia in domestic pigs have identified prevalence rates of % (Ryan et al., 2003b; Ryan et al., 2004; Johnson et al., 2008). Oocyst shedding rates of 1-58 oocysts/g faeces (dry weight) with a median of 0.5 for adult pigs and oocysts/g faeces (dry weight) with a median of 0.5 for juvenile pigs have been reported (Davies et al., 2003). There is insufficient data on Cryptosporidium oocyst loads in faecal samples from animals in catchments in Australia (particularly from wildlife) to allow for the estimation of human pathogen loads based on host source, genotype or infectivity of oocysts. Table 6. Oocyst shedding rates for various hosts. Faecal source Concentration of oocysts/g -1 Median g -1 (wet weight) Range Dairy Cattle (adult) <1-5, Dairy Cattle (juvenile) <1-17, Beef Cattle (adult) <1-3 x ,870.7 Sheep (adult) <1-152, Sheep (juvenile) <1-1.5 x x10 6 Pig (adult) < Pig (juvenile) < Kangaroo (adult) <1-2 x Water Research Australia Ltd Final Report 1036 Appendix 5

15 1.7 Factors affecting the load of human-infective oocysts in source waters The total load of oocysts in a catchment is initially determined by a number of host-related factors, including host species, host number, breed of animal, age of animal, density and stocking rates, excretion rate, etc as discussed in the preceding sections. Cattle breed is a known factor, with bivariate analyses indicating that breed (Holstein vs Jersey) and presence of diarrhoea (yes vs no) were significantly associated with the species of Cryptosporidium isolated (C. parvum vs C. bovis) (Starkey et al., 2006). For example, C. bovis was significantly more likely to be isolated from Jersey cattle than from Holstein cattle, but C. parvum was significantly more likely to be isolated from cattle with diarrhoea than from cattle without diarrhoea (Starkey et al., 2006). Seasonality is another important factor, however, insufficient data are available to determine the prevalence of Cryptosporidium in wildlife hosts across various seasons. For example, one study on eastern grey kangaroos in NSW reported that the highest prevalence rate was found to be in autumn (Power et al., 2004; Power, 2010), while a study in WA on western grey kangaroos reported that the highest prevalence occurred in summer (Yang et al., 2011). After oocysts are excreted by a host there are a number of environmental factors that affect oocyst persistence and retention of infectivity. The critical processes affecting oocysts survival in terrestrial and aquatic environments have been reviewed by King and Monis (2007). A major limitation of many survival studies has been the use of viability measures (vital dye staining, excystation) as a measure of potential infectivity, rather than using a direct infectivity measurement such as the mouse model or in vitro cell-based assays. The major processes affecting oocysts in a terrestrial environment are desiccation and temperature (King and Monis 2007). Oocysts have been shown to be especially sensitive to desiccation, with only 3% retaining viability (measured by vital dye staining) after 2 hours of air drying at room temperature (Robertson et al., 1992). The drying rate of faeces will depend on environmental conditions and so may be difficult to model, but biological processes can cause desiccation and oocyst inactivation, such as processing of manure by dung beetles (Ryan et al., 2011). Ambient temperatures of 4-25 C are unlikely to be a major factor contributing to inactivation of oocysts deposited in faeces on land because the inactivation rate at these temperatures are too slow (requiring 6 weeks at 25 C to achieve 99% inactivation) and other processes such as desiccation are more likely to cause inactivation if the oocysts are not otherwise transported into surface waters. However, in cold climates, oocysts in soil and water, have been shown to be inactivated or destroyed by freeze/thawing and abrasion from soil particles (Robertson and Gjerde, 2004; 2006). In addition, oocysts are quite sensitive to higher temperatures that might be encountered by manure exposed to the sun or from composting, with survival measured in days at 37 C (99.9% inactivation after 3 days) and minutes at temperatures above 42 C (King and Monis 2007). A summary of the impact of temperature on Cryptosporidium infectivity is provided in Table 7. Ammonia is another important factor affecting oocyst survival, particularly in manure storages or decomposing manure where the concentration of ammonia is high enough to achieve high levels of inactivation (99.999%) within 8 days of exposure. In the case of faecal material from wildlife, ammonia may play less of a role (since it is not being collected/managed), but there is little/no information on the survival of oocysts in wildlife scats. Table 7. Oocyst inactivation times for different temperatures measured using a cell culture infectivity model (King et al., 2005 and Keegan et al., 2008). Temperature ( C) Time for 90% inactivation Time for 99% inactivation weeks 18 weeks weeks 8-10 weeks weeks 5 weeks hours 180 hours hours 56 hours 45 3 ND 4 20 minutes 5 1 Keegan et al., King et al., 2005, 4 separate sets of experiments conducted using reagent water or Happy Valley reservoir water. 3 King unpublished 4 Not Done 5 greater than 99% (2 log 10) inactivation observed Water Research Australia Ltd Final Report 1036 Appendix 5 11