Abstract

• People with recent diarrhoea should avoid public swimming pools; and
• Non-toilet-trained and faecally incontinent swimmers should be provided with alternative swimming facilities with separate water and filtration systems.
  

• Doctors and pathology laboratories should consider cryptosporidiosis in patients with diarrhoea lasting longer than three days; and
  
• Laboratory reporting of cryptosporidia to local health departments should be mandatory in all States and Territories.

Cryptosporidia (Figure 1) are common in the environment and are excreted in the faeces of those infected. They can be transmitted through contact with infected cattle, sheep and other animals, ^5-7 person-to-person contact, ⁸ contaminated water supplies ^4,9 and swimming. ^10-12 Cryptosporidial infection was not a notifiable disease in New South Wales at the time of the study, and many laboratories do not routinely screen stool specimens for crypto sporidia. ³ Thus, we do not know the incidence of this disease in the community, nor the relative importance of different modes of transmission.

In January 1995, a general practitioner in the Sydney suburb of Sutherland reported to the Southern Sydney Public Health Unit increases in numbers of patients presenting with watery diarrhoea and of stool specimens positive for cryptosporidia at a local laboratory. We therefore investigated the extent and source of this outbreak of crypto sporidiosis.

Figure 1: Cryptosporidium oocysts from a faecal specimen. Modified Ziehl-Neelson stain; original magnification x 1000 (slide courtesy of Dr Stephen A Neville).

Methods

Methods were similar to those used by McAnulty et al. ¹⁰ and included a case survey and investigation of likely sources of the cryptosporidia, by comparison of cases and matched controls and an environmental study.

Case survey

Two of the pathology laboratories examined stool specimens microscopically for cryptosporidial oocysts with a modified Ziehl-Neelsen acid-fast stain (modified by decolorising with hydrochloric acid and ethanol and counterstaining with malachite green) if structures suggestive of protozoa were seen on wet-mount examination. ¹³ The other three laboratories did not screen for cryptosporidia unless requested by the referring medical practitioner.

Case patients were defined as people with a stool specimen positive for cryptosporidia between 1 September 1994 and 20 January 1995. Because of the high rate of person-to-person spread, cases were classified as primary (first to report diarrhoea in a household) and secondary (subsequent household cases).

All people diagnosed as case patients after 1 December 1994 (and reported before 20 January 1995) and aged over 18 years were interviewed by telephone in January 1995 with a 34-item structured questionnaire. Interview was with parents if the patient was less than 18 years old. The questionnaire asked about the illness and about potential risk factors in the two weeks before onset of illness, including travel outside Sydney; childcare attendance; contact with other people with diarrhoea, or with pets and domestic animals; swimming; and sources of drinking water. The period of two weeks was chosen to include the incubation period for cryptosporidiosis of 1-12 days (average, about seven days). ^14,15 People diagnosed before 1 December 1994 were not interviewed in detail because of the potential for poor recall given the time since infection.

Case-control study

Control subjects were administered a 23-item questionnaire on potential risk factors, referring to the two weeks before Christmas 1994. We chose this period as it was relatively easily identified by subjects and the season was similar or identical to that during the outbreak, so that conditions for swimming (a potential risk factor) were comparable. Differences in characteristics and risk factors between primary and secondary cases were tested for significance with either the c ² test or, if expected cell size was less than five, with a two-tailed Fisher exact test. Differences between matched pairs of cases and controls were tested for significance with the probability of the maximum likelihood estimate of the odds ratio. ¹⁶

Swimming pool investigation

Filtration systems used at local government-owned swimming pools in the inner, eastern and southern areas of Sydney were surveyed. Operators were asked over the telephone about the type of filtration system used and whether general-use pools shared their water and filtration systems with pools for infants.

Results
Case survey

Characteristics and potential risk factors of the 43 case patients reported between 1 December 1994 and 20 January 1995 are shown in Box 1. (Another five case patients were diagnosed in the week to 20 January but not reported until later and so were not surveyed.) Symptoms included diarrhoea (42 of the 43 patients, 98%), cramping abdominal pain (23, 53%), vomiting (21, 49%), nausea (16, 37%) and fever (16, 37%). None of the people affected reported being immunocompromised; more than half were children aged under 4.5 years. Twenty-eight cases were primary and 15 secondary. The only significant differences between the primary and secondary cases were that secondary-case patients were more likely to be male (80% of secondary cases versus 36% of the primary cases; P < 0.01) and, as expected, were more likely to have had contact with people with diarrhoea (100% of secondary cases versus 25% of primary cases; P < 0.001).
Figure 2: Cases of cryptosporidiosis diagnosed by laboratories in the Sutherland area of Sydney, 1 September 1194 - 20 January 1995 (specimen date was unknown for two cases).

Case-control study

Potential risk factors in the case and control groups are compared in Box 2. There was no association between illness and attendance at childcare, contact with people with diarrhoea, and drinking bottled water or city water. Although more of the case group than the control group reported swimming in the two weeks surveyed, the difference was not significant. However, significantly more in the case group than in the control group reported swimming in pool A. There was no significant association between cryptosporidiosis and swimming in any other pool.

Swimming pool investigation

Chlorine levels in pool A, documented every three hours between 6 am and 9 pm daily, ranged from 0.55 to 5.0 mg/L. The New South Wales Health Department recommends that chlorine levels for indoor pools be maintained at 1.5 mg/L for those heated to less than 26¡C, and at a minimum of 2.0 mg/L for warmer pools. ¹⁸ Pool A was heated to more than 26¡C, and chlorine levels were in the recommended range 28% of the time.

Eleven Cryptosporidium oocysts and 57 Giardia cysts were detected in a 55-L sample of the pool water taken in January 1995, but none in a 500-mL sample of filter backwash water.

Among the 20 local government-owned pools surveyed, filtration used sand at 19, and fine-grade diatomaceous earth (a porous form of silica, composed of the fossilised shells of a type of alga) was used in the other. At all these pool complexes, general purpose pools and infants' pools shared water and filtration systems.

As a result of the investigation, management at pool A erected signs warning patrons of the possibility of pool-water contamination and instructing that people who were not toilet-trained, who were faecally incontinent or who had had diarrhoea in the previous week should not enter the pool. The public was warned through a press release of the possible connection between the pool and cryptosporidiosis. On 20 January 1995, to allay public concern, pool A operators decided to replace the water in the pool and the outbreak subsequently abated.

Discussion

This outbreak probably began when the water in indoor swimming pool A was contaminated with cryptosporidia from an infected bather. The likelihood of contamination was increased by the pool's use by infants and children too young to be faecally toilet trained (crypto sporidia are present in the faeces of those infected). The infective dose of C. parvum causing illness in humans remains unclear, but recent evidence suggests it is very low (e.g., a dose of 30 oocysts has been reported to cause infection in a healthy volunteer). ¹⁹ While chlorine levels in pool A were not optimal, higher levels would have been unlikely to have prevented the outbreak, as cryptosporidial oocysts are extremely resistant to chlorine and can survive many days in chlorinated water. ²⁰ We do not know if protection from ultraviolet light in the indoor pool may have enhanced cryptosporidial survival.

While swimming at pool A was strongly associated with crypto sporidiosis, it explained only 49% of cases. No other single pool was significantly associated with illness and, as the epidemic subsided with draining of pool A, others were not tested. However, it is possible that others in the case group were infected at pools other than pool A (although not identified in our survey), or by direct contact with people with the disease, given the small infective dose. The localised nature of the outbreak and the age distribution of those affected (i.e., children or parents) indicated that drinking Sydney water was not a likely source of the outbreak.

Detection of cryptosporidial oocysts in water from pool A confirmed the epidemiological findings that C. parvum contaminated the pool. However, the viability of the organisms detected is unknown. Although Giardia cysts were also detected, they are relatively sensitive to chlorine and unlikely to be viable. ²⁰ Furthermore, local laboratories reported no increase in detection of Giardia cysts in stool specimens.

Sand filtration systems have been implicated in previous swimming pool-related outbreaks of cryptosporidiosis, ¹⁰ and almost all of the local government-operated pools surveyed relied on this type of filtration. An uneven sand surface in this type of filter -ay reduce its efficiency. ¹¹ Alternative filtration systems, such as those with diatomaceous earth, may be more effective in removing oocysts. However, swimming pools with malfunctioning diatomaceous earth filters have also been implicated in outbreaks of crypto sporidiosis. ¹² There is not enough information on the comparative effectiveness of different filtration systems to warrant recommending one form of filtration over another at present.

Because of the difficulty of eradicating cryptosporidia from swimming pools by either disinfection or filtration, preventing similar outbreaks depends on reducing contamination through the cooperation of swimmers and pool operators. ¹¹ Patrons who have had diarrhoea in the previous week, who are faecally incontinent or not toilet trained should be discouraged from using the pool by signs at the pool entrance and in the change rooms. Similar recommendations have been made for controlling swimming pool-associated crypto sporidiosis in the United States. ¹⁰ In addition, new swimming pool complexes should provide separate facilities with their own filtration systems for patrons who are not toilet trained or who are faecally incontinent. People using these facilities should avoid swallowing pool water. Screening swimming pools for crypto sporidia would be of little use at present, as current detection methods do not determine oocyst viability and consequent ability to cause disease.

In 1995, a Centers for Disease Control and Prevention workshop concluded that current knowledge about cryptosporidia, and about waterborne cryptosporidiosis in particular, is minimal and does not provide a scientifically sound basis for many essential decisions about the public health risks associated with infection. ²¹ In Australia, we do not know the incidence of cryptosporidiosis in the community or the extent to which swimming pools contribute to transmission. A first step in obtaining this information is to make cryptosporidiosis a notifiable disease and to encourage doctors and laboratories to consider cryptosporidiosis as a diagnosis in patients with diarrhoea lasting longer than three days.

In 1995, South Australia was the only Australian State to require routine reporting of cryptosporidial detection by laboratories and doctors to public health authorities. Since then, Victoria and Queensland have introduced the requirement and New South Wales followed suit on 1 December 1996. Introducing this requirement in all States and Territories would assist in early detection and control of future outbreaks of cryptosporidiosis.

Acknowledgements

1. Nime FA, Burek JD, Page DL, et al. Acute enterocolitis in a human being infected with the protozoan Crypto sporidium . Gastroenterology 1976; 70: 592-598.
2. Current WL. Cryptosporidiosis parvum : household transmission [editorial]. Ann Intern Med 1994; 120: 518-519.
3. Berkelman RL. Emerging infectious diseases in the United States, 1993. J Infect Dis 1994; 170: 272-277.
4. Gallaher MM, Herndon JL, Nims LJ, et al. Crypto sporidiosis and surface water. Am J Public Health 1989; 79: 39-42.
5. Miron D, Kenes J, Dagan R. Calves as a source of an outbreak of cryptosporidiosis among young children in an agricultural closed community. Pediatr Infect Dis J 1991; 10: 438-441.
6. Casemore DP. Sheep as a source of human cryptosporidiosis. J Infect 1989; 19: 101-104.
7. Reif JS, Wimmer L, Smith JA, Dargatz DA, Cheney JM. Human cryptosporidiosis associated with an epizootic in calves. Am J Public Health 1989; 79: 1528-1530.
8. Newman RD, Zu SX, Wuhib T, et al. Household epidemiology of Cryptosporidium parvum infection in an urban community in northeast Brazil. Ann Intern Med 1994; 120: 500-505.
9. MacKenzie WR, Hoxie NJ, Proctor ME, et al. A massive outbreak in Milwaukee of Cryptosporidium infection transmitted through the public water supply. N Engl J Med 1994; 331: 161-167.
10. McAnulty JM, Fleming DW, Gonzalez AH. A community-wide outbreak of cryptosporidiosis associated with swimming at a wave pool. JAMA 1994; 272: 1597-1600.
11. Joce RE, Bruce J, Kiely D, et al. An outbreak of cryptosporidiosis associated with a swimming pool. Epidemiol Infect 1991; 107: 497-508.
12. Sorvillo FJ, Fujioka K, Nahlen B, et al. Swimming- associated cryptosporidiosis. Am J Public Health 1992; 82: 742-744.
13. Rusnak J, Hadfield TL, Rhodes MM, Gaines JK. Detection of Cryptosporidium oocysts in human fecal specimens by an indirect immunofluorescence assay with monoclonal antibodies. J Clin Microbiol 1989; 27: 1135-1136.
14. Jokiph LJ, Jokiph DMM. Timing of symptoms and oocyst excretion in human cryptosporidiosis. N Engl J Med 1986; 315: 1643-1647.
15. Benenson AS. Control of communicable diseases manual. 16th ed. Washington: American Public Health Association, 1995.
16. Dean AG, Dean JA, Coulombier D. Epi Info Version 6. Atlanta (GA): Centers for Disease Control and Prevention, 1994.
17. Vesey G, Narai J, Ashbolt N, et al. Detection of specific microorganisms in environmental samples using flow cytometry. Methods Cell Biol 1994; 42: Pt B: 489-522.
18. Marsh C. Guidelines for disinfecting public swimming pools and spa pools. Sydney: New South Wales Health Department, 1991.
19. DuPont HL, Chappell CL, Sterling CR, et al. The infectivity of Cryptosporidium parvum in healthy volunteers. N Engl J Med 1995; 332: 13: 885-859.
20. Korich DG, Mead JR, Madore MS, et al. Effects of ozone, chlorine dioxide, chloride and monochloramine on Cryptosporidium parvum oocyst viability. Appl Environ Microbiol 1990; 56: 1423-1428.
21. Centers for Disease Control and Prevention. Assessing the public health threat associated with waterborne cryptosporidiosis: report of a workshop. MMWR Morb Mortal Wkly Rep 1995; 44 (no. RR-6): 15.