In August 2003 an outbreak of cryptosporidiosis was identified in children who had recently visited an adventure park in southwest England. The adventure park contained a number of activities involving contact with water (boats, log flume, interactive water features) and contact with farm animals. Following an earlier complaint from a visitor about the water quality of one of the interactive water features designed for water play, water sampling had revealed a high coliform count (2100 coliforms, 40 E.coli per cu mm). A cohort study was implemented to check whether there was any epidemiological evidence for a particular source within the adventure park.

Methods
The cohort population included all children (aged less than 16 years) among household members or friends of a probable or confirmed case who had visited the park with a case during August 2003. A probable case was defined as any child who had visited the park during August 2003 and who subsequently had gastro-intestinal symptoms including diarrhoea, blood in stools, vomiting, nausea, or abdominal pain. A confirmed case was defined as a probable case with a faecal sample positive for cryptosporidium. Children who had travelled abroad in the two weeks before the onset of symptoms were excluded from the study.

Cases were identified from laboratory reports to the Health Protection Agency (HPA). A standardised questionnaire was administered over the telephone with an adult in the family of a case. Exposure data included water exposure (contact duration, type of contact, type of water source), animal contact and food consumption. Data were analysed using Epi Info 6.04 [7]. A univariable analysis was run to assess the association between exposures investigated and onset of disease. As only one variable showed an association and as the numbers were small, multivariable analysis was not performed.

Ten-litre grab samples were taken from the various water features within the park for cryptosporidium oocyst detection by South West Water Ltd. Faecal samples from the farmyard animals were also submitted. Oocysts were detected by light microscopy. Positive specimens were sent to the HPA Cryptosporidum Reference Laboratory for genotyping.

Results
Ninety one children were identified in the cohort, of whom 71 were contacted, giving a 78% response rate. Sixty three children (89%) met the case definition (27 confirmed and 36 probable cases). The sex distribution was even. Median age was 6 years (range 1-15). The most common symptom was diarrhoea (94%), followed by vomiting (64%), abdominal pain (62%), and nausea (51%). None of the children reported blood in stools. The median duration of illness was 8 days (range 1-18) and more than 30% of the children were still ill at the time of interview. Ten children (16%) required hospital admission.

Forty-six of the children who were cases (73%) had visited the park on 8 August, the date of symptom onset for the first case. Of the 51 children whose date of illness onset was known, 45 (88%) had a date of onset within one incubation period (1-10 days) of visiting the park [FIGURE].

Dates of onset were between 8 and 29 August, and the outbreak peaked on 13 and 14 August. For two of the four cases with date of onset more than 10 days after visiting the park, other household members had had gastrointestinal symptoms in the 10 days before onset. The two probable cases with onset date on date of visit became ill during the evening after leaving the adventure park.

The exposure yielding the strongest association with illness was contact with the interactive water feature [TABLE] (RR= 1.8, CI 95% 0.45 to 7.31, p=0.06). No specific type of contact with this source of water was significantly associated with illness. This feature involved being sprayed with recirculated water. Children often entered the feature fully clothed and with their shoes on. Nineteen children drank the recycled water and one parent reported that the water ‘smelt like drains’. The filtration and disinfection systems were not adequate to cope with high levels of contamination, and the water feature was closed on 21 August, soon after the start of this investigation.

Samples from 23 of the 27 confirmed cases were sent for genotyping. Sixteen yielded a result and 14 of these were Cryptosporidium parvum genotype 2. The initial sample from the interactive water feature contained a single oocyst that could not be genotyped. Although a subsequent sample from this feature when not in operation was positive and identified as Cryptosporidium parvum genotype 2, there was insufficient DNA for subtyping. Due to a failure of communication, faecal samples taken from animals resident in the park were not tested for cryptosporidium.

Discussion
This outbreak of cryptosporidiosis was characterised by a high attack rate (89% in the cohort studied), long duration of illness (median 8 days) and high proportion admitted to hospital (16%). The dates of onset were consistent with a common source of infection from an exposure in the adventure park. The analytical study showed an association between exposure to water in the interactive water feature and illness. Although the strength of the evidence was reduced due to the small numbers in the unexposed group, the finding was supported by the microbiological results and environmental observations. No association with other water sources or animal contact was detected. It seems likely that water in the interactive water feature became contaminated with faeces containing cryptosporidium oocysts, either from the footwear of users or from an unidentified primary case. These oocysts then continued to circulate in a viable condition as a result of ineffective filtration and disinfection.

In response to the outbreak, the park reviewed and revised health and safety risk assessments to manage and control the risk from protozoan parasites. The design of the water treatment and disinfection system was improved. The park also provided additional drinking fountains around the park and asked children to remove footwear before entering the interactive water feature. They improved signage, instructing visitors at all water-related attractions not to drink the water.

This outbreak has similarities to two others reported in England in 2003 involving public water features. The first, which also occurred in southwest England, involved four cases of cryptosporidiosis in children who had played in a fountain. The water feature comprised two separate water bodies with separate holding tanks and water treatment systems using bromide and sand filtration. A large pool with water to a depth of 20cm was used as a paddling pool, although it was not intended for this purpose. Cryptosporidium oocysts were isolated from all four cases and detected in water samples taken from the fountain.

The second outbreak, which occurred in central England, was linked to a newly opened purpose-built interactive water feature, and involved 122 cases. More than 80% (102) of those infected were under 15 years old. Thirty five (85%) of 41 cases tested for cryptosporidium were positive. Indicator organisms of faecal contamination were identified from the water but no cryptosporidium oocysts were recovered.

These outbreaks raised issues about the lack of national guidance on operation and maintenance of water-based recreational attractions, which have now been addressed by the United Kingdom Pool Water Treatment Advisory Group [8]. The principal public health measure for preventing infections and outbreaks associated with these devices is risk assessment and management. The principal microbiological risks are cryptosporidiosis from inadequate filtration, and bacterial and viral infections, including legionella, from inadequate disinfection. This guidance proposes design and operational standards for filtration, chlorination and reducing contamination hazards.

Acknowledgements
We would like to acknowledge the contributions of R Chalmers, Cryptosporidium Reference Unit, Swansea, Wales; B Madhu, Coventry Health Protection Unit, UK; G Makin, Coventry City Council, UK; A Colville, Royal Devon and Exeter Hospital, UK; G Charles, East Devon District Council, UK; E Rodford, Somerset Health Protection Unit, UK; E Duffel, Somerset Health Protection Unit, UK; E Thomas, Health Protection Agency South West, UK.