Protozoa in drinking water: is legislation the best answer?

A cooperative approach is a better option to protect public health

MJA 1998; 169: 296-297

In late July, Australians were shocked to learn that the water supply in their largest city had been declared unsafe to drink because of protozoal contamination.1 In late August Sydney's water was again declared unsafe. The publicity surrounding these events in Sydney doubtless left most of the public, and perhaps many health professionals, with the impression that swallowing a mouthful of the affected water virtually guaranteed an episode of severe gastroenteritis. However, the magnitude of the risk to public health is far from clear as there are many limitations in our understanding of the biology of these organisms, and in the methods used for their detection in drinking water. Public and political concern has resulted in calls for legislative enforcement of water quality standards for Giardia and Cryptosporidium, but we believe that this approach is both unwise and unworkable at present. Such measures would require the setting of permissible levels for protozoa in drinking water, specification of sampling programs and approved testing methods. Uncertainties in all these aspects make it extremely difficult to define appropriate parameters to protect public health. Human infection: Information on levels of protozoa associated with illness is available from a small number of human experiments and limited data collected from outbreaks. For Giardia lamblia, a study of adult male prison inmates showed as few as 10 cysts could establish infection (determined by cysts in stools). However, none of the 40 subjects reportedly developed symptomatic giardiasis despite ingesting up to one million cysts, although 21 became infected.2 More recent human experiments,3 in which 50 000 Giardia lamblia trophozoites were inoculated into the duodenum, showed that with one strain all 10 subjects became infected and four developed clinical giardiasis. With another strain none of five subjects became infected or showed clinical disease. In human infection experiments with Cryptosporidium parvum, the minimum dose tested was 30 oocysts. Of five seronegative subjects receiving this dose, one became infected (oocysts in stools) but experienced no symptoms. At the next dose level of 100 oocysts, three of eight subjects became infected and developed symptoms.4 Waterborne outbreaks: Information derived from investigation of waterborne outbreaks is limited and difficult to interpret because of the time lag between the contamination event, the onset of symptoms in the exposed population and subsequent investigation to identify the source. For the Milwaukee cryptosporidiosis outbreak, which affected an estimated 400 000 people, the only data on oocyst levels came from samples of stored ice.5 A concentration of 13.2 oocysts per 100 L was found in ice made eight days before the outbreak was recognised. Maximum exposure to cryptosporidia probably occurred three days later (five days before the outbreak was recognised), when water turbidity rose suddenly to about seven times normal levels, signalling a failure of the water filtration plant. The process used to recover oocysts is noted for its variability,6 and this figure may be a substantial underestimate. Swimming pool outbreaks of cryptosporidiosis illustrate that illness can result from ingestion of a small volume of contaminated water, but, again, the time lag between a contamination event and examination of water samples makes it difficult to estimate actual exposure levels.7 Infected people can shed millions of oocysts per gram of faeces, so ingestion of tiny fragments of faecal matter may be sufficient to cause infection in other pool users.8 Safe drinking water: Overall, data presently available are insufficient to allow a "safe" drinking water level to be defined for these protozoa. It is theoretically possible that ingestion of even a single cyst or oocyst may carry a low risk of developing illness, but it is not feasible to test this hypothesis experimentally. Cryptosporidiosis may cause diarrhoeal illness lasting several days in healthy people, but in AIDS patients inability to clear the infection may result in severe and intractable diarrhoea which ultimately contributes to premature death. It is notable that during the Milwaukee outbreak people with HIV were not more likely to become ill than those in the general population.9 This suggests that the infectious dose for Cryptosporidium parvum is similar in both immunocompromised and immunocompetent people, although the consequences of infection are clearly different. There is evidence that the coagulation step used in water treatment to remove particulate matter causes clumping of coliform bacteria and spores, and similar effects may occur with protozoa. This would result in exposure of fewer people to larger numbers of protozoa than would be predicted from assumptions of uniform distribution in drinking water.10 Water testing: The formulation of a meaningful water sampling program is also a problem. Waterborne outbreaks are rare, and are believed to result from short term "spikes" of contamination from increases in protozoa numbers in the source water, or failure in normal water treatment processes, or a combination of both factors. A program based on spot sampling would be unlikely to detect contamination spikes, and could not provide statistically meaningful information on the very low numbers of protozoa which are normally present. Other parameters, such as turbidity or particle counts, may provide warning of abnormalities in water treatment processes (such warnings were unfortunately ignored in Milwaukee), but in some instances outbreaks have occurred without detectable changes in such measures or any identifiable breakdown in operating processes.11 Only Giardia and Cryptosporidium species of mammalian origin are believed to constitute a risk to human health, but current tests do not indicate the type of animal the protozoa originated from or the viability of cysts and oocysts. Several techniques to determine viability and species have recently been developed but are not yet fully validated. Considerable variability exists in the recovery efficiency of concentration techniques for isolating protozoa from water, making it difficult to compare levels between different studies and different laboratories.6 False positive results from algae and other particles of similar size and appearance to protozoa may also be a significant problem.12 Because of uncertainties about testing methods and the public health significance of low levels of protozoa in water, the National Health and Medical Research Council decided not to set guideline levels for protozoa in the 1996 Australian Drinking Water Guidelines, or to recommend testing for these organisms. Considerable progress has been made in detection techniques since then, but many problems are still to be resolved before we can accurately and reliably measure the number of viable protozoa of the relevant species. Nevertheless, major water companies are testing for protozoa with increasingly sensitive methods in an effort to improve the quality of their supplies, but water and health authorities are faced with a dilemma over what to do when positive results are found. Solutions: While legislation may appear to be the answer to this problem, we believe this issue is far too complex to be resolved in this way. The interests of public health would be better served by an open, cooperative approach bringing together the expertise of government, public health and the water industry. Australia would benefit from the development of best-practice programs, appropriate for the circumstances of different water supplies, and covering water quality from source to tap. Such programs are already being implemented in other countries.13 There is also a need to develop consensus protocols for graded responses to contamination incidents, and improved communication with the public and interest groups. Current surveillance mechanisms for communicable diseases are fragmentary, relatively insensitive and slow,14 and should be improved and integrated with water quality data to provide more sensitive and rapid detection of outbreaks. Cooperative research efforts are required to improve water monitoring techniques and confirmatory tests, together with appropriate measures for quality assurance. Only then will we be in a position to assess whether protozoa in drinking water are causing illness in the community, and determine appropriate measures to protect public health. Martha I Sinclair Senior Reseach Fellow Christopher K Fairley Associate Professor Margaret E Hellard NHMRC PhD Scholar Department of Epidemiology and Preventive Medicine and Cooperative Research Centre for Water Quality and Treatment Monash University, Melbourne, VIC Most of Sydney told: boil drinking water. The Sydney Morning Herald, 1998; 30 Jul: 1. Rendtorff RC. The experimental transmission of human intestinal protozoan parasites. II Giardia lamblia cysts given in capsules. Am J Hyg 1954; 59: 209-220. Nash TE, Herrington DA, Losonsky GA, Levine MM. Experimental human infection with Giardia lamblia. J Infect Dis 1987; 156: 974-984. DuPont HL, Chappell CL, Sterling CR, et al. The infectivity of Cryptosporidium parvum in healthy volunteers. N Engl J Med 1995; 332: 855-859. MacKenzie W, Hoxie N, 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. LeChevallier MW, Norton WD, Siegel JE, Abbaszadegan M. Evaluation of the immunofluorescence procedure for detection of Giardia cysts and Cryptosporidium oocysts in water. App Environ Microbiol 1995; 61: 690-697. Lemmon JM, McAnulty J, Bawden-Smith J. Outbreak of cryptosporidiosis linked to an indoor swimming pool. Med J Aust 1996; 165: 613-616. Chappell CL, Okhuysen PC, Sterling CR, DuPont HL. Cryptosporidium parvum: intensity of infection and oocyst excretion patterns in healthy volunteers. J Infect Dis 1996; 173: 232-236. Frisby HR, Addiss DG, Reiser WJ, et al. Clinical and epidemiologic features of a massive waterborne outbreak of cryptosporidioisis in persons with HIV infection. J Acquir Immune Defic Syndr Hum Retrovirol 1997; 16: 367-373. Gale P, van Dijk PAH, Stanfield G. Drinking water treatment increases micro- organism clustering; the implications for microbiological risk assessment. J Water Services Res Technol -- Aqua 1997; 46: 117-126. Goldstein ST, Juranek DD, Ravenholt O, et al. Cryptosporidiosis: an outbreak associated with drinking water despite state-of-the-art water treatment. Ann Intern Med 1996; 124: 459-468. Clancy JL, Gollnitz WD, Tabib Z. Commercial labs: how accurate are they? J Am Water Works Assoc 1994; 5: 89-97. The partnership for safe water. American Water Works Association. <URL http://www.awwa.org/partner1.htm> Padiglione AP, Fairley CK. The early detection of outbreaks of waterborne gastroenteritis -- a feasibility study. WaterTECH Conference, Brisbane, April 1998.