Limiting swimming pool outbreaks of cryptosporidiosis – the roles of regulations, staff, patrons and research

Death due to diarrhoea is one of the top five leading causes of mortality in children under 5 years of age (WHO 2013; GBD Mortality Causes of Death Collaborators 2015). Cryptosporidium is one of the four major contributors to moderate-to-severe diarrhoeal disease during the first 5 years of life in low-to-middle income countries (Kotloff et al. 2013). Indeed, it was second only to rotavirus as a cause of moderate-to-severe diarrhoea in children under 2 years and associated with a two to three times higher risk of mortality among children aged 12–23 months with moderate-to-severe diarrhoea, than in controls without diarrhoea (Kotloff et al. 2013). The most recent Global Burden of Disease study listed Cryptosporidium as an important cause of diarrhoea and death in children under 5 years of age (especially under two) in sub-Saharan Africa (GBD Mortality Causes of Death Collaborators 2015).

In Australia, there are an estimated 15.9 M cases of gastroenteritis each year (Kirk et al. 2014), resulting in economic costs estimated at over one billion dollars per year (through visits to medical practitioners and time taken off work) (Hellard et al. 2003; Anon 2006; Heyworth et al. 2006; Kirk et al. 2014). There have also been reports of an association between acute gastroenteritis (AGI) and the development of irritable bowel syndrome and reactive arthritis, and the costs of these chronic sequelae account for another $75 million (Anon 2006; Ford et al. 2014). Important agents that cause diarrhoea include viruses (e.g., rotavirus and norovirus), bacteria (e.g., Shigella species, Escherichia coli and Campylobacter jejuni) and parasites (Giardia duodenalis, Entamoeba histolytica and Cryptosporidium species (Shah et al. 2009).

Waterborne parasitic protozoa such as Cryptosporidium that are spread through contaminated drinking water and recreational water are a major contributor to gastroenteritis as they have a worldwide distribution (Cotruva et al. 2004; Shields et al. 2008a; CDC 2011), result in more than four billion cases of diarrhoea and 1.6 million deaths annually (www.who.int) and 62.5 million disability-adjusted life years worldwide (Cotruva et al. 2004; Baldursson & Karanis 2011; Burnet et al. 2014). From 2004 to 2010, 199 outbreaks of human gastroenteritis due to the waterborne transmission of enteric parasitic protozoa were reported worldwide and, of these, Cryptosporidium spp. were the aetiological agent in 60.3% of the outbreaks (Baldursson & Karanis 2011). Bathing in contaminated swimming and therapeutic pools is a major mode of waterborne transmission of Cryptosporidium and other pathogens (Dale et al. 2010; Soller et al. 2010; Sanborn & Takaro 2013).

Given the prominence of Cryptosporidium as a significant source of gastroenteritis globally, this review was undertaken to identify studies that describe swimming pool-associated outbreaks of cryptosporidiosis and to identify key barriers to limiting swimming pool-associated outbreaks of cryptosporidiosis. We hoped that such information can serve as a resource to health authorities worldwide.

To identify relevant studies, we searched PubMed (http://www.ncbi.nlm.nih.gov/pubmed/) as well as ScienceDirect (http://www.sciencedirect.com/) and Google Scholar (http://scholar.google.com/). The following search terms were included: Cryptosporidium, cryptosporidiosis, outbreak, diagnosis, causes, swimming pool, recreational water, disinfection, regulation, staff and patrons. A total of 1,400 potentially relevant studies were identified. Studies were considered relevant if they identified specific recreational water-associated outbreaks of cryptosporidiosis and/or causes, limits of detection, regulation and roles of staff and patrons in outbreaks, resulting in more in-depth analysis of 157 studies.

Currently, 31 Cryptosporidium species have been recognized as valid and of these more than 17 species have been identified in humans (Holubová et al. 2016; Kváč et al. 2016; Zahedi et al. 2016). By far the most common species reported in humans worldwide are C. parvum and C. hominis (Xiao 2010; Ryan & Xiao 2014) with C. hominis the main cause of human cryptosporidiosis in most countries (Xiao 2010; Ryan & Power 2012). Symptoms of cryptosporidiosis include diarrhoea, dehydration, weight loss, abdominal pain, fever, nausea and vomiting. In immunocompromised people, cryptosporidiosis can become chronic and result in death (Chalmers & Davies 2010; Xiao 2010). Although hundreds of drugs have been tested for prophylaxis and treatment of cryptosporidiosis in animals and humans, only one, nitazoxanide (Alinia^®), has been approved for use in humans by the US Food and Drug Administration (FDA). This drug, however, exhibits only moderate clinical efficacy in children and immunocompetent people and none in people with HIV (Abubakar et al. 2007; Amadi et al. 2009; Pankiewicz et al. 2015).

Drinking water-associated outbreaks of diarrhoeal illness due to Cryptosporidium and other pathogens are declining, particularly in high-income countries, but the number of outbreaks of cryptosporidiosis and other diarrhoeal diseases associated with swimming pools, water parks and other ‘disinfected venues’ has increased dramatically (Pond 2005; Beach et al. 2009; IOM 2009; Valderrama et al. 2009; Yoder & Beach 2010; Yoder et al. 2010; Brunkard et al. 2011; Hlavsa et al. 2011, 2015; Sanborn & Takaro 2013). Outbreaks of cryptosporidiosis have been associated with recreational water in the USA (e.g., Sorvillo et al. 1992; Moore et al. 1993; CDC 1994, 2008; McAnulty et al. 1994; MacKenzie et al. 1995a; Lim et al. 2004; Mathieu et al. 2004; Craun et al. 2005; Causer et al. 2006; Wheeler et al. 2007; Shields et al. 2008b; Boehmer et al. 2009; Cantey et al. 2012; Cope et al. 2015; Hlavsa et al. 2015; DeSilva et al. 2016), Canada (Bell et al. 1993; Louie et al. 2004; Ong et al. 2008; Hopkins et al. 2013), Australia (e.g., Lemmon et al. 1998; Hellard et al. 2000; Stafford et al. 2000; Puesch et al. 2001; Black & McAnulty 2006; Paterson & Goldthorpe 2006; Dale et al. 2010; Mayne et al. 2011; Waldron et al. 2011; Ng-Hublin et al. 2015), Japan (Ichinohe et al. 2005; Yokoi et al. 2005; Takagi et al. 2008), Sweden (Insulander et al. 2005; Mattsson et al. 2008) and the UK (e.g., Joce et al. 1991; Hunt et al. 1994; Furtado et al. 1998; Chalmers 2000; McLauchlin et al. 2000; Smith et al. 2006; Coetzee et al. 2008; Briggs et al. 2014). In the USA, treated recreational water was the source for 77% of all waterborne outbreaks between 2011 and 2012 (Hlavsa et al. 2015). In Australia, swimming in public pools was shown to be the main risk factor associated with sporadic cryptosporidiosis (Robertson et al. 2002). Swimming and therapeutic pools are, therefore, considered a more important transmission route for gastrointestinal illness than drinking water in high-income countries (Pond 2005; IOM 2009; Dale et al. 2010; Chalmers 2012).

Cryptosporidium oocysts can survive for 3.5–10.6 days in swimming pool water where free chlorine levels are maintained at Centers for Disease Control (CDC)-recommended levels (1–3 mg/L) (Shields et al. 2008b). Bromine, another chemical disinfectant used in swimming pools, is also ineffective for inactivation of Cryptosporidium at the doses used in normal pool water treatment (Korich et al. 1990). Viruses and bacteria can also cause swimming pool outbreaks, however most of them are more easily controlled by the proper application of chlorine and other disinfectants (WHO 2000). For example, Shigella spp. are quite susceptible to low levels of chlorine disinfectants and can be inactivated at levels less than 1.0 mg/L (Kebabjian 1995). Some viruses such as norovirus are reportedly resistant to levels of chlorination used in swimming pools (Keswick et al. 1985), but more recent studies report that adequate chlorination of pools may be sufficient to prevent norovirus transmission (Podewils et al. 2007).

Cryptosporidium is particularly suited to waterborne transmission as the environmental stage, the oocyst, is highly resistant to disinfection (Korich et al. 1990; Chauret et al. 2001; Painter et al. 2015) and is excreted in large quantities (10⁸–10⁹ oocysts in a single bowel movement) (Cordell 2001; Yoder & Beach 2007; Yoder et al. 2012) for up to 60 days after cessation of gastrointestinal symptoms (Jokipii & Jokipii 1986; Stehr-Green et al. 1987). In addition, the low infectious dose (10–100 oocysts) (DuPont et al. 1995; Chappell et al. 2006), means that ingestion of a relatively small amount of contaminated water is sufficient to initiate infection in a susceptible individual. The long incubation period of cryptosporidiosis (averaging 7 days) (Chalmers & Davies 2010), delays the identification of the source and implementation of interventions to decontaminate the source thus perpetuating transmission.

A quantitative risk assessment has estimated that ingestion of a single oocyst of the C. parvum IOWA isolate will result in clinical disease in 2.79% of immunologically normal persons (Pouillot et al. 2004). Given that the 50% infective dose (ID₅₀) for C. hominis is less than one-tenth that of the C. parvum IOWA isolate (DuPont et al. 1995), ingestion of only one oocyst of a more infectious isolate may lead to a higher incidence of infection in the general immunocompetent population. The following example shows how a single accidental faecal release (AFR) from a person infected with Cryptosporidium in a swimming pool can result in an infective dose for a human child or adult. An infected individual can excrete up to one billion oocysts during an infectious period and 1 mL of faeces can contain as many as 5 × 10⁷Cryptosporidium oocysts (Polgreen et al. 2012). If a child has a loose bowel movement of 150 mL into a typical 25 m × 12 m municipal pool containing about 450 m³ of water, then there would be an average concentration of about 20,000 oocysts/litre (20/mL) (Gregory 2002). Oocysts would be dispersed throughout the pool due to the currents created by other swimmers, especially in a heavily used venue. The average swimmer ingests ∼10–150 mL per hour of pool water (Dufour et al. 2006; Soller et al. 2010; Schets et al. 2011), therefore, a swimmer swallowing just 10 mL of water would ingest an average of 200 oocysts (Kebabjian 1995), which is well above a dose capable of causing infection (<100 oocysts) (DuPont et al. 1995; Chappell et al. 2006). A recent US study reported the risk of Cryptosporidium infection in all swimmers at 25 infections per 1,000 swimmers/year; 22 infections per 1,000 adult swimmers/year, and 29 infections per 1,000 child swimmers/year (Suppes et al. 2016).

Studies worldwide have shown that there is at least one AFR every week in most swimming pools throughout summer, with many pools reporting daily contaminations, especially hydrotherapy pools or those highly utilized by non-toilet-trained infants and toddlers (CDC 2001; Castor & Beach 2004; Schets et al. 2004; McManus et al. 2009; Chalmers et al. 2011). Evidence also indicates that commercially available swim diapers are of limited value in preventing the release of Cryptosporidium into pool water (Maas et al. 2004; Amburgey & Anderson 2011).

Despite this, the majority of gastroenteritis outbreaks due to cryptosporidiosis and other enteric diseases are never identified, as: (1) it is estimated that less than 10% of individuals with gastroenteritis visit their local doctor, and of these, less than 10% have a faecal specimen collected (Garthright et al. 1988; Hellard & Fairley 1997; Majowicz et al. 2005; Tam et al. 2012); (2) not all individuals presenting with gastroenteritis will have faecal samples tested for microorganisms (Majowicz et al. 2005; Tam et al. 2012); (3) identification of Cryptosporidium and other pathogens via microscopy lacks specificity and sensitivity (Checkley et al. 2015); (4) microbiological confirmation of a swimming pool contamination is limited by the often transient nature of the contamination; and (5) current outbreak detection methods lack sensitivity, specificity and timeliness, i.e., there is a considerable lapse of time before an outbreak is recognized (Hellard et al. 2000; Dale et al. 2010). For example, the largest Cryptosporidium outbreak occurred in Milwaukee in 1993, where an estimated 403,000 individuals contracted cryptosporidiosis via contaminated drinking water (MacKenzie et al. 1995b) with an estimated illness-associated cost of $96.2 M and 100 deaths (Corso et al. 2003). Despite the high number of infected individuals, only a small number of cases were investigated prior to the identification of the outbreak and confirmation of its cause. Another outbreak of cryptosporidiosis in California involving more than 250 people, whose common exposure was a waterpark, went unreported to the local public health department for weeks, despite ∼75% of the waterpark's 26 employees developing a gastrointestinal illness (Wheeler et al. 2007). In fact, most incident cases of cryptosporidiosis are not confirmed to be due to Cryptosporidium (Chalmers 2008; Smith et al. 2010; Tam et al. 2012; Briggs et al. 2014), and many apparently sporadic cases may be part of small outbreaks where the remainder of the cases go undiagnosed.

Cryptosporidiosis is an important disease in Oceania with an annual incidence of 12.8 illnesses per 100,000 and 22 per 100,000 in Australia (Lal et al. 2015a) and New Zealand (Snel et al. 2009), respectively. These rates are significantly higher than the USA (2.9 per 100,000), Canada (2.7 per 100,000) and in England and Wales (8 per 100,000) (Snel et al. 2009). In Australia, swimming pools were identified as the source of the majority of outbreaks of cryptosporidiosis in urban areas with cases mostly reported in children under 5 years of age and adult females (Lal et al. 2015a). This trend is also observed worldwide with the number of cryptosporidiosis outbreaks associated with disinfected recreational water venues increasing over the past several decades. For example, outbreaks linked to swimming pools have increased in the UK, with the majority of cases reported in the last 10 years (Chalmers 2012). In the USA, the number of recreational water-associated outbreaks reported annually has increased substantially since reporting to CDC began in 1978 (Hlavsa et al. 2011). Between 1995 and 2012, there were >1,000 recreational water-associated outbreaks of cryptosporidiosis in the USA (Painter et al. 2015), and the percentage of outbreaks of AGI caused by Cryptosporidium and associated with treated recreational water was 82.9% for both 2007–2008 (58/70) and 2005–2006 (29/35) (Yoder et al. 2008). In contrast, it was 68.2% (15/22), 50.0% (9/18) and 55.6% (10/18) in 1999–2000, 2001–2002 and 2003–2004, respectively (Lee et al. 2002; Yoder et al. 2004; Dziuban et al. 2006; Hlavsa et al. 2011).

There is a distinct seasonal pattern in the incidence of cryptosporidiosis, with the peak risk period coinciding with the summer recreational water season, reflecting increased use of communal swimming venues (Lal et al. 2015b; Painter et al. 2015). The recent increase in the number of reported outbreaks of recreational water-associated cryptosporidiosis is also associated with changes to reporting requirements in the USA and improvements in detection, investigation and report­ing of waterborne disease outbreaks (Hlavsa et al. 2011).

In contrast to drinking water, there is a lack of uniform national and international standards for public treated recreational water venues (Chalmers 2012; Lewis et al. 2015; Painter et al. 2015). This is a major barrier to the prevention and control of illness associated with swimming pools as many swimming pool outbreaks are due to failures in policies and procedures, poorly designed sanitation systems and inadequate maintenance (Chalmers 2012; Lewis et al. 2015; Painter et al. 2015).

The World Health Organization published guidelines for swimming pools in 2006 (WHO 2006), which addresses a wide range of hazards, including water quality, physical hazards (leading to drowning and injury), contamination of associated facilities and air quality. It includes guidelines for the assessment of the health hazards associated with swimming pools, their monitoring and assessment and maximum permissible microbiological concentrations derived from the findings of numerous studies into the infective doses of Cryptosporidium and other pathogens (WHO 2006). This guidance was included in the revised guidelines published by the UK Pool Water Treatment Advisory Group (PWTAG) in 2009 (http://pwtag.org). However, neither of these documents is legally enforceable and therefore adherence to the guidance varies across the industry (Lewis et al. 2015). In most pools, monitoring for potential microbial hazards is done using faecal indicator microorganisms such as coliform bacteria and especially Escherichia coli (E. coli) (rather than specific pathogens such as Cryptosporidium), which are easier and more cost-effective to detect and enumerate but are not informative on the presence of Cryptosporidium (Harwood et al. 2005; Coupe et al. 2006) and have much shorter survival times when exposed to chlorine than viruses and protozoa (Krusé 1969; Dahling et al. 1972; Steenland & Savitz 1997). Furthermore, many pools have been shown to exceed national legal limits for microbiological quality (Tesauro et al. 2010; Blougoura et al. 2011) and, in some facilities such as splash parks, no consistent requirements for water treatment exist (de Man et al. 2014).

In the UK, key aspects of Cryptosporidium (AFR) management are covered by the general requirements of current UK legislation such as the Health and Safety at Work Act 1974 (HMSO 1974). However, there are no specific regulations designed specifically for the swimming pool industry. In the USA, treated recreational water and pool regulations are enacted, implemented and enforced by state or local governments (Painter et al. 2015). In 2014, the CDC published a new set of guidelines – the Model Aquatic Health Code (MAHC) (http://www.cdc.gov/mahc, accessed 10 May 2016), which identifies Cryptosporidium as the leading cause of pool-associated outbreaks in the USA. These guidelines provide support to state and local health departments looking to reduce the risk of recreational water-associated illness. However as with the UK, they are voluntary guidelines and are currently not enforceable. Similarly, in Australia, swimming pools are managed at a local government level and while there are guidelines designed to regulate and guide the operation of swimming pools, including the Australian Standards (AS) and the Guidelines for Safe Pool Operation (GSPO), they are a voluntary guide for best practice only. In stark contrast, there is an apparent absence of swimming pool-associated outbreaks of cryptosporidiosis in Japan (Nakamura-Uchiyama et al. 2003; Karanis et al. 2007), where, anecdotally, swimming pool regulation is tightly enforced including compulsory showering prior to bathing and bans on young swimmers (<3 years old) and swimmers with signs of respiratory infections. While it is not possible to ascribe cause and effect in this situation it is tempting to consider the impact that more proactive regulation could have.

In swimming pools, chlorine is the most commonly used disinfectant. Other disinfectants, such as chloramine, chlorine dioxide, ozone and UV irradiation are also applied for water disinfection (Chowdhury et al. 2014). The filtration and disinfection of water in a swimming pool typically occurs as the water is recirculated through a treatment area at regular intervals that could be up to 4–8 hr on average (via a system of interconnected inlets and outlets spaced around the pool). A target pH (e.g., 7.2 to 7.6) and free chlorine level (e.g., 2 to 4 mg/L) are typically maintained in pool water (often by an automatic controller) to achieve some level of residual disinfection between treatment cycles. However, as discussed above, Cryptosporidium can survive for up to 11 days in properly chlorinated water (Amburgey et al. 2012). In addition, when chlorine reacts with organic matter in water (e.g., constituents of sweat and urine, skin particles, hair, microorganisms, cosmetics and other personal care products), a variety of disinfection by-products (DBPs) can be formed. Among these, trihalomethanes (THMs), haloacetic acids (HAAs), haloacetonitriles (HANs) and chloral hydrate (CH) are the most prevalent chlorinated by-products (Chowdhury et al. 2014). Exposure to those DBPs is inevitable for any bather or trainer and can have elevated risks to human health as THMs and other DBPs are classified as probable carcinogens in humans (Chowdhury et al. 2014).

Currently AFRs are managed opportunistically, with high doses of chlorine added to the contamination area. In some instances, pools are closed to the public, then either emptied, disinfected and then refilled or the entire water body of the complex is superchlorinated (usually requiring a 12–24 hr closure). These procedures are a significant cost to the industry (∼US$12,000/pool in lost revenue, cleaning expenses and labour). However, the recommendations for responding to faecal accidents may not be effective. For example, 20 mg chlorine/L for 12.75 hr is required to achieve a Ct [chlorine concentration in mg/L × time in min] value of 15,300 (=3 log reduction in oocyst viability and infectivity) (Shields et al. 2009). Furthermore, a study that simulated an AFR in a swimming pool to test the effectiveness of superchlorination on faeces-associated oocyst infectivity reported that oocysts were still infectious for mice after 3 days of exposure, possibly because the organic material protected oocysts from inactivation (Carpenter et al. 1999). There is also a risk that oocysts that form aggregates could be protected from inactivation (Carpenter et al. 1999; Gregory 2002). In addition, cyanuric acid, a chlorine-stabilizing compound used in pools throughout the world, has been shown to decrease the inactivation rate of Cryptosporidium (Shields et al. 2009; Murphy et al. 2015). The pH of the water, age of the oocysts and geographical origin (i.e., isolate variation) have also been shown to result in large differences in the Ct values required for a 3 log reduction of oocysts (Shields et al. 2008b).

As chlorine is largely ineffective at inactivating Cryptosporidium, the burden of protecting the public from cryptosporidiosis typically falls on the filters. Most swimming pool filters are sand filters that are not designed to remove Cryptosporidium (Amburgey et al. 2012). Research has shown that even ‘high-rate’ (i.e., 25–49 m/hr filter loading rate) swimming pool sand filters can only consistently deliver 0.11 to 0.28 log (22 to 48%) removal of Cryptosporidium oocysts and/or a microsphere surrogate without coagulation (Croll et al. 2007; Amburgey et al. 2012), which is inadequate to prevent outbreaks of cryptosporidiosis. In addition, modelling has shown that ‘dead spots’ exist at either end of the pool where Cryptosporidium could remain, even with filtration (Lewis et al. 2015). Precoat (i.e., diatomaceous earth and perlite) filters have been reported to remove 2.3–4.4 log (99.5–99.996%) spiked Cryptosporidium from swimming pool water at a filtration rate of 6.1 m/hr (swimming pool volume was less than 1,000 L) (Amburgey et al. 2009a, 2009b, 2012). However the ability of precoat filters to remove Cryptosporidium in full-scale swimming pools has yet to be extensively researched.

Chemical coagulation (creating clumps of oocysts) prior to filtration, greatly enhances the ability of the filters to remove Cryptosporidium oocysts, which normally pass through the filter (by reducing the natural electrostatic repulsion that exists between negatively charged oocysts and negatively charged filter media in water (Hall et al. 1994; Hendricks 2006; Amburgey et al. 2012). For example, addition of the coagulant polyaluminium chloride (PAC) consistently removed more than 90% of Cryptosporidium-sized microspheres at 30 m/hr filtration rate while the removals dropped to approximately 50% at a filtration rate of 37 m/hr with a 5.5 m³ pilot-scale swimming pool (Lu & Amburgey 2016). Coagulant efficiency is pH dependent (should be <7.5) and even with coagulation under the correct pH, filters can allow breakthrough of Cryptosporidium oocysts if they are not backwashed frequently enough (Hall et al. 1994). Filter backwashing is a cleaning process by which the water flow through the filter is reversed so that accumulated debris trapped in the filter is dislodged and directed to waste. However, filter backwashing can disrupt filter integrity and allow Cryptosporidium oocysts to re-contaminate the pool water and infect swimmers (Greinert et al. 2004; Schets et al. 2004; Chalmers et al. 2011; Ehsan et al. 2015). Other factors that affect the ability of filters to remove Cryptosporidium include filter loading rate, depth of the filter media, coagulant type, coagulant dosage and, potentially, even how and where the coagulant is dosed (Amburgey et al. 2007, 2012; Croll et al. 2007).

Other disinfection methods such as ozone followed by chlorine as a secondary disinfectant (ozone/chlorine), and electrochemically generated mixed oxidants are increasingly being substituted for chlorine to minimize the formation of DBPs. Ozone is a high potential oxidant and disinfectant but is usually followed by sequential disinfection with chlorine (ozone/chlorine) because of its instability and the relatively high doses required (Kleiser & Frimmel 2000). UV disinfection can reduce the overall chlorine consumption by 50%, thus treatments with UV light and/or ozone can increase the efficacy of chemical disinfectants for Cryptosporidium (Chalmers 2012; Donofrio et al. 2013; Lu et al. 2013) if the dose and contact time (CT) are sufficient (Driedger et al. 2000; Rennecker et al. 2000; Chalmers et al. 2011). However, as with ozone, UV only renders the water bactericidal at the point of use and therefore additional (residual) disinfection is required in swimming pools to deal with the effective inactivation of Cryptosporidium. In addition, bathers in the vicinity of an AFR from someone infected with Cryptosporidium may still be at risk if they ingest pool water because outbreaks have been documented even in pools disinfected with chlorine and ozone (Moore et al. 1993; McAnulty et al. 1994; Chalmers 2000; Boehmer et al. 2009; Chalmers et al. 2011). In addition, ozone is depleted more rapidly in the presence of organic materials and its efficacy is highly temperature dependent (i.e., decreasing at lower temperature) (Chalmers et al. 2011). Although widely used in the drinking water industry, there are relatively few data relating directly to UV light disinfection of Cryptosporidium in swimming pool water, but bench-scale studies from drinking water have shown that Cryptosporidium oocysts are irreversibly inactivated by UV treatment (at doses of ∼10 mJ/cm²) (Clancy et al. 2004; Rochelle et al. 2005; Hayes et al. 2013). Typically, doses of >60 mJ/cm² are used in swimming pools, which should be effective (Chalmers et al. 2011). While UV is effective at killing oocysts, such a barrier will only be effective if it is maintained and operated correctly. UV irradiation has been suspected to result in enhancement of THM concentrations (Judd et al. 1998); however, another study reported an absence of THM formation by UV and UV combined with chlorine (Spiliotopoulou et al. 2015). In addition, most of the research on ozone and UV disinfection of Cryptosporidium has involved small-scale laboratory bench studies, using very limited numbers of Cryptosporidium species. This is important because the robustness of the oocysts can vary substantially between sources and batches (Gregory 2002). More evidence of the efficacy of combined disinfection methods is needed under real-world conditions in continuous flow swimming pool waters.

Membrane filtration processes where water is forced by pressure (positive or vacuum) across a surface, excluding or rejecting particles greater than the effective pore size of the membrane material, such as microfiltration, ultrafiltration (UF), nanofiltration (NF), and reverse osmosis (RO) have been applied in swimming pools to remove DBPs (Barbot & Moulin 2008; Klüpfel et al. 2011) and Cryptosporidium-sized particles have been removed by NF membranes from a public drinking water source (Patterson et al. 2012). Largely because of the costs involved, this technology has yet to be routinely applied to swimming pools but full-scale application of a combined UF and RO treatment significantly reduced particulate matters from a swimming pool (Reißmann et al. 2005) and would prove useful in removing Cryptosporidium oocysts in response to an AFR.

Staff at recreational facilities play a key role in controlling cryptosporidiosis outbreaks via an informed and timely response to AFRs and by adhering to strict hygiene standards themselves. Illness among lifeguards has been noted previously in reports of cryptosporidiosis outbreaks associated with recreational water. In the first reported outbreak of swimming pool-associated cryptosporidiosis in the USA, the authors noted that 75% of the lifeguards were ill (Sorvillo et al. 1992). Likewise, a swimming pool-associated cryptosporidiosis outbreak in Canada reported that 27% of the lifeguards had been ill (Bell et al. 1993). However, neither of these reports included the illness onset times of lifeguards relative to patrons, nor whether the lifeguards could have contributed to illness among patrons. In a Canadian swimming pool-associated cryptosporidiosis outbreak in 2004, a pool employee had a positive stool test for Cryptosporidium one month before the outbreak had been detected, and this person and other employees had been in the pools during their symptomatic periods (Louie et al. 2004). In 2012, in Western Australia, a swimming pool outbreak of cryptosporidiosis was linked to a swimming instructor, diagnosed with cryptosporidiosis, who continued to swim at the local pool, contrary to the doctor's advice (Ng-Hublin et al. 2015). Thus lifeguards and swimming instructors may, in some cases, initiate, amplify or prolong an outbreak by going into the water while ill. A clear policy to keep lifeguards and bathers with diarrhoea out of pools and waterparks should be standardized, implemented and enforced.

A rapid and adequate response by staff to known and suspected faecal contamination events is also essential to prevent outbreaks. Management of faecal and diarrhoeal episodes should adhere to recommended guidelines and should include hyperchlorination for a time period that equates to a CT inactivation value of 15,300 (e.g., free chlorine levels of 20 ppm for 12.5 hr) (CDC 2016). Pool management and staff should ensure the facility complies with current codes and standards of operation and insist on patrons’ adherence with regulations and practice of good hygiene when using the pool facilities. However, a recent survey of AFR responses in Australian pools revealed that many pools were not super-chlorinating adequately following an AFR (i.e., the majority of pools (>80%) were using less than half the recommended chlorine concentration and for only a fraction of time (10 min)). Many operators were also unclear as to how Cryptosporidium is transmitted. This finding is consistent with similar surveys conducted in the USA (CDC 2010). Studies of swimming pool-associated waterborne outbreaks have reported that a lack of coherent and standardized pool regulations are a major reason for outbreaks (Wilberschied 1995; Joce et al. 1991; Wheeler et al. 2007), which highlights the need for better education of staff and more effective and standardized management practices.

Swimming pool patrons also have an important role to play in preventing Cryptosporidium transmission. Prevention is a two-fold approach where a patron needs to prevent self-exposure, but also contaminaton of others (McClain et al. 2005). Current recommendations to prevent introducing Cryptospodirium into a pool are that swimmers refrain from entering a pool while ill with diarrhoea and, if diagnosed with cryptosporidiosis, for at least 2 weeks following complete symptom resolution (Painter et al. 2015). Yet, a recent US survey examining caregivers’ awareness of cryptosporidiosis in children found just over a quarter of respondents (26.9%) were aware of abstaining from swimming until at least 2 weeks following complete symptom resolution (Munoz et al. 2016).

High risk swimming activities or behaviours (e.g., pool water ingestion, frequency of splashes to the face, leisurely swimming) (Suppes et al. 2014) should be avoided; however, one study reported that only 15% of respondents were aware that swallowing recreational water while swimming can lead to Cryptosporidium transmission (Munoz et al. 2016). Outbreak investigations of cryptosporidiosis related to exposures in swimming pools show that questionnaires are typically administered 2–3 weeks after exposures occur due to illness latency and case identification complications (Hellard et al. 2000; Insulander et al. 2005; Causer et al. 2006; Boehmer et al. 2009). Extended time periods between exposure and questionnaire administration increases the likelihood that recall bias of swimming activities associated with outbreaks will occur, with a recent study reporting up to 28% difference in recall bias when a questionnaire was administered immediately and 1 week after study participants swam in pool water (Suppes & Reynolds 2014). The uncertainties associated with swimmer recall have important implications for statistical associations between outbreaks and patron swimming activities, and needs to be factored into conclusions formulated by swimming pool outbreak investigators.

Recreational water can amplify smaller outbreaks into community-wide transmission when persons who are ill introduce the parasite into multiple recreational water venues or other settings (e.g., childcare facilities) (Cordell 2001; Turabelidze et al. 2007). Thus, management of an individual cluster of cases of cryptosporidiosis associated with a particular venue may be confounded by the simultaneous transmission at other venues (Painter et al. 2015). For example, an investigation of a community-wide cryptosporidiosis outbreak in Utah indicated that 20% of case-patients swam while ill with diarrhoea and identified approximately 450 potentially contaminated recreational water venues (CDC 2008). Educational and health promotion efforts to improve awareness and change swimmer behaviour is therefore a critical component of outbreak prevention.

Cryptosporidium is a major cause of swimming pool-associated outbreaks and developing a safer swimming environment is essential to protect patrons against contracting cryptosporidiosis from pools. Further research is required to better understand the knowledge, attitudes and practices (KAP) of swimming pool operators with regards to the public health risks associated with AFR and their management in order to inform swimming pool regulations. In addition, a better understanding of the KAP of swimming pool patrons, especially parents of young children, would enable the development of evidence-based education programmes to reduce the likelihood of AFR and reduce the likelihood of an infected person swimming. As child swimmers have the highest risk of infection due to their immature immunity, they should be targeted by public awareness messages and in education campaigns.

Lack of uniform national and international standards for swimming pools is a major issue, and state and local governments need to establish and enforce regulations for protecting swimming pools from Cryptosporidium contamination, including legislation on minimum design standards, operation, disinfection, filtration and AFR management. In the absence of this, communication bridges between pool operators, local health departments and other aquatic facilities should be encouraged. Pool operators need to be better educated on the risks of Cryptosporidium contamination and best practice management procedures, including a written AFR response policy and keeping records of all faecal accidents, chlorine and pH level measurements, and any major equipment repairs or changes. A number of factors can contribute to inadequate pool treatment: poor operating procedures (including disinfection), inadequate or inappropriate filtration equipment, excessive numbers of bathers, a high contamination load or a combination of these factors (Chalmers 2000). Most swimming pool filters have have limited effectiveness for Cryptosporidium removal and therefore continuous application of coagulants such as polyaluminum chloride (PAC) to reduce oocyst concentrations that may be present from unreported AFRs is recommended. The Pool Water Treatment Advisory Group recommends dosing pool water with PAC continuously by peristaltic pump at 0.1 mL/m³ of the circulation flow in the UK and other European countries (Pool Water Treatment Advisory Group 2015). Very little is known about filter monitoring practices in most swimming pools but monitoring filter performance using parameters such as turbidity might provide an indication of filter performance, especially when coagulation is used. Pool operators should also be encouraged to separate filtration systems between childrens’ paddling pools and adult pools, as if they are connected, faecal contamination can be dispersed from the children's paddling pool to the other pools. Provision of clean toilets, stocked with toilet paper, with ample soap for hand washing as well as clean shower areas and child changing rooms will also promote better hygiene practices by patrons.

Understanding the efficacy of existing pool management practices is essential and requires sensitive molecular detection and typing of Cryptosporidium from pool filters and AFRs (in addition to standard microbial testing) to determine which pools have the highest prevalence (and if specific strains/subtypes are being spread to other pools), coupled with patron and staff questionnaires to analyse factors linked to high prevalence levels and transmission of the parasite. Molecular detection should also be coupled with trialling of improved accidental faecal release management guidelines (AFRM) to determine if their implementation will reduce the public health risk associated with Cryptosporidium contamination of public swimming pools. These would include development and distribution of healthy swimming education materials to season pass holders, daily patrons and children at schools, prior to the ‘pool season’, that focus on avoiding swimming for 2 weeks after experiencing diarrhoea, intentional pool water ingestion, and splashing others in the face, etc.

Outbreaks of cryptosporidiosis are difficult to detect because surveillance systems are relatively insensitive (Hellard et al. 2000). However, timely intervention in the event of a rapid increase in notifications of disease linked to swimming pools has been shown to reduce the probability and scale of a community-wide outbreak (Coetzee et al. 2008). The interventions are broad in nature and generally include communication with local medical services (community and acute care) to increase vigilance and rate of confirmatory testing and to give patients appropriate hygiene and exclusion advice (Coetzee et al. 2008). In addition, swimming pools can be encouraged to increase chlorination rates and enhance signage regarding exclusion after illness (Hellard et al. 2000).

Lastly, most disinfection studies have been conducted on only a few isolates of C. parvum and only one study has been conducted on UV inactivation of C. hominis (Johnson et al. 2005), despite this species being responsible for 70–90% of human cases of cryptosporidiosis (Xiao 2010). Although morphologically indistinguishable, there are distinct biological differences between the two species. Cryptospordium hominis is more virulent than C. parvum and has been associated with increased oocyst shedding (intensity and duration) and non-intestinal symptoms such as joint pain, eye pains, recurrent headache and fatigue whereas C. parvum is only associated with diarrhoea (Chalmers & Davies 2010). Therefore, full-scale swimming pool experiments to estimate the effectiveness of disinfection methods including superchlorination, filtration, ozone, UV and membrane filtration on a range of C. parvum and C. hominis isolates with and without the presence of faecal material need to be conducted.