Health Stream Issue 16 - December 1999

National Workshop on Drinking Water Quality Management

A workshop was convened by the NHMRC in Adelaide on 8 October to discuss the development of national guidelines for drinking water quality management. Representatives from water authorities, state and territory health departments, catchment and environmental organisations, the Consumers' Health Forum, and the NHMRC/ ARMCANZ Drinking Water Review Coordinating Group attended the workshop together with specialists in environmental risk management and public health.

Professor Don Bursill, CRCWQT Director and Chair of the NHMRC/ARMCANZ Drinking Water Review Coordinating Group, opened the meeting with a background to the workshop including an overview of the Australian Drinking Water Guidelines, and the rolling revision process that was adopted in 1996 to ensure that the guidelines are continually kept up to date with evolving scientific knowledge.

There is some concern that the Australian Drinking Water Guidelines are used in the industry mainly as a focus for compliance-based management strategies without sufficient recognition of the importance of overall system management for assuring safe drinking water. Furthermore, with the rapidly changing environment of the water industry in Australia and the trend toward increasing fragmentation and disaggregation, the Guidelines need to reflect this to allow for greater divergence of responsibilities and accountabilities across multiple agencies.

As part of the second round (1999/2000) of the ongoing review process of the Australian Drinking Water Guidelines, increased emphasis will be given to the preventive nature of the Guidelines by expanding and reorganising the relevant information as a practical and comprehensive risk management approach to drinking water quality management from catchment to tap. In addition to emphasising prevention rather than corrective action (reaction), this approach will allow the numerical limits of the Guidelines to be viewed in the proper perspective as providing verification that the management systems are effective rather than being viewed as the primary means of protecting public health.

CRCWQT researchers, in cooperation with the Coordinating Group, have developed a discussion paper detailing a suggested scope and structure of a framework for drinking water quality management. The framework was derived by supplementing the information on preventive system management already provided in the current Australian Drinking Water Guidelines with principles from existing systems for quality management and risk assessment/management including:

This framework and paper formed the basis for discussions at the workshop.

Following Professor Bursill's introduction, Professor Steve Hrudey from the University of Alberta, noted expert in environmental risk management and recent Visiting Professor at the CRCWQT, presented an overview of essential risk management principles for effective drinking water quality management. He emphasised the particular importance of anticipating and preventing harm in water quality management rather than just reacting to problems after they arise. Risk assessment can be a useful tool in this process; however the degree of scientific uncertainty inherent in the process must be recognised with due regard for the contributions of direct evidence, scientific inference and policy-based assumptions. Risk management decisions must be informed decisions, but should include recognition of how much uncertainty exists.

Research Fellow Samantha Rizak from Monash University reviewed the proposed framework for drinking water quality management and the benefits offered by its adoption. The framework is intended to provide the directional guidance on a comprehensive preventive strategy to drinking water quality management from source to tap. Key features of the framework include:

The perspective of health authorities was summarised by Dr. David Cunliffe of the South Australian Department of Human Services. All Australian jurisdictions face the challenges of increasing requirements for surveillance of water quality and the trends to increase legislative controls. Dr Cunliffe also noted that the proper management of water quality includes management of all components of water supplies from catchment to tap and requires involvement of all stakeholders. Health authorities have a key role in supporting this process. The proposed framework provides the opportunity for all stakeholders to become involved and offers the outcome of a cooperative and coordinated approach with improved understanding of the roles and responsibilities of all parties.

The Water Corporation of Western Australia has recently undertaken work to improve its activities relating to water quality management. Mr Richard Walker, Water Quality Manager, presented the process and findings of the risk assessment workshop that the Water Corporation recently conducted. This has involved systematic risk assessment of several water systems representing different types of catchments, and the identification of potential risk management strategies. The process particularly highlighted the need for improved information flow between water authorities and other bodies with responsibility for land management in catchments.

In the discussions that followed, the benefits of adopting a national approach to drinking water quality management were evident. It was agreed that a comprehensive and preventive framework jointly developed by major stakeholders could provide a flexible and effective means of assuring the protection of public health, as well as increasing communication and defining responsibilities of the various agencies and stakeholders involved in the supply of water. The framework will also provide a common and unified approach throughout the industry which can establish due diligence and credibility.

However, it was stressed that whatever approach is developed must remain flexible and adaptable to local conditions to reflect the wide range of differences in Australian water supplies and the varying organisational and jurisdictional arrangements. Further, the importance that the system be industry-supported was highlighted and it was suggested that external auditing protocols for drinking water quality management will be helpful for establishing credibility and maintaining consumer confidence. The preference for industry-specific and industry-developed external auditing protocols was emphasised.

The Coordinating Group is now developing a work plan to develop the Drinking Water Quality Management Framework. This will commence with a series of desktop trials of the draft framework by teams from several water authorities. The first set of trials will involve Melbourne Water, Sydney Water, the Power and Water Authority of the Northern Territory and the Water Corporation of Western Australia. The framework will be refined following feedback from these trials and then tested by a larger number of water authorities. Both rounds of trials are expected to be completed by mid-2000.

Updates on the progress of the Rolling Revision of the Australian Drinking Water Guidelines can be found on the web site of the CRCWQT.

Cryptosporidium and water - the numbers game

Over the last decade the threat of waterborne disease from Cryptosporidium parvum has become a major influence on drinking water regulation in the developed world. This protozoan pathogen is common in surface water supplies and is resistant to the levels of chlorination that can be applied to drinking water. While unfiltered water supplies are generally believed to be at highest risk of outbreaks, the Milwaukee outbreak clearly demonstrated that even filtered supplies can be endangered if the pathogen load is high and the filtration process is not efficient.

The presence of Cryptosporidium oocysts in virtually all surface water supplies has raised the question of whether exposure to oocysts in drinking water could be causing low levels of illness in communities (endemic disease) in the absence of recognised outbreaks. This question is very difficult to answer since we do not yet have reliable tests to determine whether oocysts are capable of causing human infection, and cases of illness arising from waterborne exposure cannot be distinguished from illness transmitted by other routes except in the context of outbreaks.

The US Environment Protection Agency has adopted a quantitative risk analysis process to estimate the current magnitude of endemic waterborne cryptosporidiosis and derive predictions for the reduction in illnesses which would be achieved by improved water treatment. These calculations are the main driving force behind the Interim Enhanced Surface Water Treatment Rule ⁽¹⁾, which must be implemented in most US surface water supplies serving more than 10,000 people by December 2001.

Recently, an independent estimate of the occurrence of cryptosporidiosis in the community has been made by the Centres for Disease Control and Prevention as part of the process for deriving a national estimate of foodborne disease ⁽²⁾. Comparison of these two sets of calculations makes for some interesting reading.

The risk analysis considered a number of factors influencing the risk of infection in consumers, and made assumptions about their likely values based on published literature and other data.

Dose response slope parameter

Information from the only human feeding trial published at that time ⁽³⁾ was used to derive a slope constant k, which is part of the equation used to relate ingested dose with probability of infection.

Daily water consumption

It was assumed that water consumption in a population was lognormally distributed with a mean of 1.948 litres per day and the maximum capped at 3 litres per day.

Cryptosporidium in raw water sources

Data from a 1996 survey of 69 sites were used to estimate the occurrence of oocysts. This relatively limited data suggested that mean oocyst concentrations could be described by a lognormal distribution. Two sources of error in the data were assumed to cancel each other; firstly the sampled sites were in areas which might have higher concentrations than the US as a whole, secondly the recovery efficiency of the method was poor and thus the measured concentrations would have been underestimates.

Effect of current water treatment (filtration)

Information from studies at pilot and full scale water treatment plants was assessed together with consideration of more recent trends towards improved filtration efficiency.

Two sets of assumptions were used to assess the impact of different treatment efficiencies. The first set assumed that the performance of current treatment plants was normally distributed with a mean removal of 2.5 logs and a standard deviation of 0.63 logs. The second set of assumptions assumed the current mean removal was 3 logs, again with a standard deviation of 0.63 logs.

Viability of oocysts

Current detection methods cannot reliably determine whether Cryptosporidium oocysts detected in water supplies belong to the single species known to infect humans, nor whether they are viable and capable of causing infection. For the analysis it was assumed that viability had a mean of 10% with a uniform distribution ranging between 5% and 15%.

Based on the observations from the only human feeding trial published at that time, it was assumed that 39% of infected people would become ill (with 95% confidence limits of 19 to 62%).

The probability of mortality among ill people was estimated at 0.000125 based on data from the Milwaukee outbreak.

It was estimated that 139.2 million people are supplied by the 1,381 water systems subject to the IESWTR. This represents 51% of the total US population.

Using the assumptions outlined above, a Monte Carlo simulation was carried out to estimate the current number of illnesses and deaths predicted to be due to Cryptosporidium exposure from water supplies subject to the IESWTR. This type of simulation treats each parameter as a probability distribution and carries out repeated calculations selecting a value for each parameter in turn based on their individual distribution. This results in the generation of another probability distribution of the estimated number of infections and illnesses. From this distribution the mean (expected value) and range of risk is calculated.

These simulations resulted in the following estimates for current infection and illness rates due to Cryptosporidium in drinking water:

Assumed effect of water filtration	2.5 log removal	3.0 log removal
Estimated illnesses	643,000	208,500
Estimated deaths *	80	26

* not stated in reference - estimated from data supplied

The higher figure of 643,000 illnesses attributable to Cryptosporidium infection from current water supplies is most commonly used in EPA documents.

It should be noted that these figures apply to the 139.2 million people served by the water supplies which are subject to the IESWTR (this corresponds to about 52% of the total US population). The remainder of the population is served by surface water supplies not covered by the Rule (ie less than 10,000 people, or larger surface supplies not required to filter their water) or by groundwater supplies or private supplies.

The CDC paper on foodborne disease estimates the annual number of cases of gastrointestinal illness and death in the US from all sources including known pathogens, unknown pathogens and noninfectious causes, and then estimates the proportion that are attributable to food. A variety of data sources are used including the FoodNet active surveillance system, passive surveillance systems, outbreak investigation and specific surveys. For each known pathogen, the data sources and assumptions made to derive the estimate are summarised in a detailed Appendix.

The authors comment that the number of cases of illness reported to any surveillance system is considerably less than the number which occur in the community, as many people with mild illness do not seek medical attention, most who see a doctor do not give a faecal specimen for analysis, the relevant tests may not be performed on faecal specimens, and test results may not be reported to surveillance.

Some information on the degree of under reporting has been gained from specific studies on Salmonella and E.coli O157:H7. Salmonella (which typically causes non-bloody diarrhoea) is under reported by a factor of about 38-fold in the US, while E.coli O157:H7 (which typically causes bloody diarrhoea) is under reported by a factor of about 20-fold.

For Cryptosporidium parvum the number of total cases in the US was estimated from the observation that about 2% of stool specimens tested for Cryptosporidium are positive for this organism. It was assumed that 2% of the 15 million people who see a doctor for acute gastroenteritis would have a Cryptosporidium infection - giving an estimate of 300,000 cases per year. This number is vastly in excess of the number of cases reported to active surveillance (6,630) or to passive surveillance (2,788).

The ratio of cryptosporidiosis cases reported to active surveillance to the estimated total cases is about 45-fold, which may perhaps be explained by the fact that tests for this organism are less commonly performed that those for common bacterial pathogens such as Salmonella. The total number of deaths attributable to C ryptosporidium infection from the estimated 300,000 cases was estimated at 66 per year based on the fatality rate in cases reported to FoodNet.

This estimate does not take into account the number of people who become ill with cryptosporidiosis but do not seek medical attention. A recent British study estimated that about 1 in 2 people who have symptomatic cryptosporidiosis see a doctor because of their illness ⁽⁴⁾. If a similar figure applies in the US, then the total number of illnesses due to Cryptosporidium would be around 600,000 cases.

Differences in the healthcare systems between the two countries may mean that a smallerproportion of people in the US seek medical attention, and this would have the effect of increasing the estimate of the total number. However the British study also found that testing of stool specimens was related to severity of symptoms, and that people who saw a doctor with cryptosporidiosis were more likely to have a specimen tested than the average gastroenteritis patient. Thus the US assumption that the 2% positivity rate for Cryptosporidium would apply to all people who visited a doctor for gastroenteritis is likely to be an overestimate. This factor would tend to reduce the estimate of total cases of illness.

How do the numbers compare?

The EPA figure relates only to waterborne illness in that portion of the population which is served by supplies subject to the IESWTR - it does not incorporate illness due to other routes of transmission in this population or in the remainder of the US population. At this time there is no numerical data available on exposures to Cryptosporidium oocysts by non-water routes (for example in foodstuffs and drinks which are consumed without heating, from recreational water use, by animal to person contact, or person to person contact). Therefore it is not possible to apply quantitative risk analysis techniques to these exposures. However examination of serological data may perhaps give some insight in to the contribution of non-water sources.

Antibodies against C. parvum have been found in 15% to 60% of US adults in various studies. These antibodies are believed to decline below detectable levels within about 18 to 24 months if exposure does not recur in this period. Therefore the presence of antibodies is taken to represent fairly recent exposure to C. parvum .

If it is assumed that infections resulting in illness occur in the same proportion as exposures from different sources which lead to antibody formation, then comparison of antibody prevalence rates in populations with differential waterborne exposure may indicate the relative importance of water as a route of transmission. It should be noted that this assumption has not been proven, although it is one of the arguments used to support the view that water is a major transmission pathway for endemic disease due to Cryptosporidium.

Comparison of antibody prevalence rates in populations served by groundwater sources (assumed to have virtually zero waterborne transmission) with those in populations served by surface waters of varying quality, suggests that non-water sources are responsible for at least 50% of exposures. If we assume that the same holds true for illnesses then we would predict that communities subject to the IESWTR are experiencing at least 643,000 cases of non-waterborne cryptosporidiosis per year in addition to the 643,000 cases attributable to water.

If it is assumed that the population served by sources not affected by the IESWTR has zero waterborne risk, then we would predict that they experience about 617,000 cases of illness from other sources (48% of the US population x 643,000 cases). This would give an overall estimate of at least 1,903,000 symptomatic cases of Cryptosporidium infection each year in the US. By the same logic, the total number of deaths predicted from EPA estimates of Cryptosporidium infection would be about 237 per year in the US. In contrast, the corresponding figures from the CDC estimate would be around 600,000 illnesses (using the multiplier factor from the British study) and 66 deaths per year.

Clearly the large discrepancy between these estimates raises some concerns - particularly as the EPA figures have been used in cost-benefit analyses to justify the costs entailed in implementing the IESWTR (approximately $307 million annually). If the current number of waterborne Cryptosporidium illnesses has been substantially overestimated, then the number of cases prevented by improvements in drinking water filtration will also have been overestimated. The CDC estimate for cryptosporidiosis also involves a number of assumptions and uncertainties, although fewer than the EPA estimate, and these are arguably more amenable to direct investigation and resolution with currently available methodologies, including epidemiological studies of community gastroenteritis to determine the true rate of disease.

(1) Regli S, Odom R, Cromwell J, Lustic M and Blank M (1999) Benefits and costs of the IESWTR. J American Water Works Association 91 (4) p148-158.

(2) Mead, P. S., L. Slutsker, et al. (1999). Food-Related Illness and Death in the United States. Emerging Infectious Diseases 5(5): 607-25.

(3) See From the Literature section for the most recent publication on Cryptosporidium infectious dose.

(4) Wheeler, J. G., D. Sethi, et al. (1999). Study of infectious intestinal disease in England: rates in the community, presenting to general practice, and reported to national surveillance. British Medical Journal 318(17 April): 1046-50.

US EPA sued over chloroform

Legal action has been taken against the US Environment Protection Agency by the Chlorine Chemistry Council, an organisation representing manufacturers of chlorine and chlorinated products. The action stems from withdrawal of the EPA proposal to relax the Maximum Contaminant Level Goal (MCLG) for chloroform in drinking water. The MCLG is not an enforceable standard but rather provides a target for water utilities to aim for. MCLGs are defined under the US Safe Drinking Water Act as levels "at which no known or anticipated adverse health effects occur, allowing for an adequate margin of safety".

The proposal, released for public comment in March 1998, would have raised the MCLG for chloroform from zero to 300 micrograms /litre, while setting the Maximum Contaminant Level (an enforceable level) at 80 micrograms /litre for total trihalomethanes (including chloroform). However in its final version of the Disinfection /Disinfection Byproduct Rule released in November 1998, the EPA withdrew the proposed change and retained the previous chloroform MCLG of zero.

The proposal to raise the MCLG was based on advice from an expert panel convened by the EPA to examine evidence on the carcinogenicity of low level chloroform exposure. The panel concluded that early rodent experiments which employed single high daily doses of chloroform in corn oil were flawed in design and did not provide useful information on the risks of exposure to chloroform in drinking water.

More recent experiments comparing high doses of chloroform delivered in corn oil and the same doses in drinking water clearly showed liver cell toxicity and stimulation of cell growth only with the corn oil vehicle. These effects on cell function are responsible for increased rates of liver cancer development when corn oil is used for dosing. In contrast, no impact on cancer rates was seen in mice with drinking water exposures as high as 1800 milligrams /litre.

The second major factor in the expert panel's decision was the accumulation of evidence that chloroform does not act directly on genetic material (DNA). Substances which are able to directly damage DNA are by convention deemed to have no threshold for carcinogenic action, since a single mutation induced by such damage could potentially be inherited by daughter cells and perhaps lead ultimately to cancer. The EPA's risk assessment procedure for such substances is based on the assumption that any exposure results in some finite risk of cancer (modelled as a linear no-threshold dose response curve).

There is now a substantial body of evidence that chloroform does not directly damage DNA, thus the appropriate method to assess the cancer risk is to adopt a model which allows for a no-effect threshold level (modelled as a non-linear dose response curve). However while the EPA issued a statement that it "believed the non-linear cancer extrapolation approach is the most appropriate means to establish an MCLG for chloroform based on carcinogen risk" , it nevertheless did not accept the outcome of this analysis as a basis for setting the MCLG. The reason for this decision appears to be the strong negative opinions expressed by environmental groups during the public comment period on the proposed revision of the chloroform MCLG.

The EPA decision to disregard expert advice has been cited as an example of how the rigid regulatory structures which exist in the US may render any "relaxation" of environmental standards virtually impossible even when such action is clearly justified by rigorously assessed scientific evidence.

The Chlorine Chemistry Council has launched legal action on the grounds that the EPA has failed to follow the statutory requirements of the Safe Drinking Water Act to use the "best available peer-reviewed science" in setting the MCLG for chloroform. The case is due to begin before a panel of judges in the Washington DC Circuit Court of Appeals on 11 February 2000.

Cryptosporidium detection methods

The lack of a rapid reliable assay to determine whether Cryptosporidium oocysts detected in water are infectious for humans has been a major problem in risk assessment and development of water treatment techniques. This has been compounded by problems with the efficiency and reliability of oocyst recovery from water samples. Several recently published papers suggest that rapid improvements are occurring in both these areas of research.

Di Giovanni and coworkers ⁽¹⁾ have reported the use of an integrated cell culture-PCR technique to detect infectious oocysts in both seeded and unseeded water samples and filter backwash. Oocysts were recovered from 10 litre samples (seeded where appropriate with 1600 to 2900 oocysts) using immunomagnetic separation (IMS). The washed and resuspended oocysts were inoculated into cell cultures of human ileocecal adenocarcinoma cells and incubated for 72 hours at 37C. The cell monolayers were washed to remove non-excysted oocysts and then harvested. The presence of viable C parvum was detected by PCR of the hsp70 gene. The identity of PCR products was confirmed by DNA sequencing. For comparison, oocysts were also concentrated by conventional percoll-sucrose flotation. Aliquots from both methods were also examined by immunofluorescence microscopy (IFA).

The spiking experiments showed higher recoveries with immunomagnetic separation compared to flotation but this was not statistically significant. Recoveries were higher with both methods for seeded raw water than for seeded backwash water. The results of PCR detection (positive/negative) agreed with detection by microscopy following either IMS or flotation.

A total of 122 raw water samples and 121 filter backwash samples from 25 different sites were tested. Oocysts were detected in 16 (13.1%) of raw water samples and 7 (5.8%) of backwash samples by flotation - IFA, and in 6 (4.9%) of raw water samples and 9 (7.4%) of filter backwash samples by cell culture- PCR. However only 2 samples tested positive by both methods. The flotation-IFA method should detect all Cryptosporidium species whether dead or alive, while the cell culture - PCR technique should detect only viable C. parvum.

The cell culture technique (without PCR) has been adapted by other researchers to provide a Most Probable Number method for estimating oocyst numbers ⁽²⁾. In this case cell cultures were grown in 8 well slides prior to inoculation with oocysts, then incubated for 48 hours prior to fixing. Slides were stained with a 2-stage antibody procedure, then examined under a microscope for signs of sporozoite invasion (initial infection) and clustering (indicating secondary infection after completion of the first replication cycle). Wells were scored as negative unless both invasion and replication were present. The viability of lots of oocysts was also evaluated using vital dyes and mouse infection tests.

The results showed considerable variability in the infectivity of different lots of oocysts (all from the Iowa strain), suggesting that preparation techniques may be responsible. The cell culture infectivity of all tested lots decreased over time, but the vital dye and excystation assays showed little change. This reinforces observations from other studies that such markers do not accurately reflect infectivity. This type of cell culture assay may have advantages over those employing molecular techniques (eg PCR) since only oocysts capable of undergoing more than one round of replication will be scored positive. This will prevent over estimation of viable oocyst numbers as some oocysts in environmental samples have been shown to be able to invade cells but not proceed to the secondary infection stage.

Zuckerman and coworkers have tested a portable continuous flow centrifuge for concentrating oocysts from water samples, and suggest it may have advantages over current cartridge filtration and calcium carbonate flocculation methods ⁽³⁾. This centrifugation technique has been previously reported for oocyst recovery, but only with large non-portable equipment. The results of spiking experiments and field tests with natural water samples comparing the three methods by these authors suggest that continuous flow centrifugation shows the best recovery.

(1) Detection of infectious Cryptosporidium parvum oocysts in surface and filter backwash water samples by immunomagnetic separation and integrated cell culture-PCR. Di Giovanni, G. D., F. H. Hashemi, et al. (1999). Appl Environ Microbiol 65(8): 3427-32.

(2) A most-probable-number assay for enumeration of infectious Cryptosporidium parvum oocysts. Slifko, T. R., D. E. Huffman, et al. (1999). Appl Environ Microbiol 65(9): 3936-41.

(3) Evaluation of a portable differential continuous flow centrifuge for concentration of Cryptosporidium oocysts and Giardia cysts from water. Zuckerman, U., R. Armon, et al. (1999). J Appl Microbiol 86: 955-61.

News Items

UK Cryptosporidium outbreaks linked to pools

Health authorities in Britain have linked several recent small outbreaks of cryptosporidiosis to swimming pools. Over 90 confirmed cases have been recorded in four separate towns since August. In two outbreaks the pools were reported to be using ozonation as well as chlorination.

Australian water still low cost

A survey by the National Utility Service has shown that Australians still pay less for their water than most western countries. Australian water charges were rated the 5th cheapest among 15 countries, compared to 3rd cheapest in last years survey. The highest prices were reported for Germany ($2.73 per cubic metre) and Denmark ($2.43 per cubic metre). Countries with cheaper water prices than Australia were Canada (62 cents per cubic metre), Sth Africa (75 cents), the United States (80 cents), and Spain (82 cents). According to the survey, Australian prices rose by about 3% during the last year.

Hurricane provokes storm over pig farm waste

A political battle has erupted in North Carolina US over plans to rebuild waste lagoons at pig farms damaged by Hurricane Floyd. The eastern part of the state contains many intensive pig farms, established since the 1980s, which have been permitted to dispose of animal waste in open waste pits with little protection of local water supplies. Flooding from the September hurricane resulted in widespread pollution of surface water sources, and there are indications that seepage is also affecting groundwater supplies which serve up to half the residents in the worst affected areas.

Moves to tighten regulation of the pig farming industry began in the early 1990s but little progress has been made due to strong opposition from major producers, and the economic importance of the industry. Now, opponents say that federal aid should not be approved to rebuild the same inadequate waste storage facilities which threaten the environment.

Ulcers linked to Helicobacter

Researchers at the Pennsylvannia State University have reported evidence of a direct link between the presence of the bacterium Helicobacter pylori in unchlorinated well water and stomach ulcers in consumers. The research team had previously reported finding the microorganism in samples from many private wells in the state. According to a press release from the University they have now conducted interviews with residents and found a statistically significant association between cases of stomach ulcers among consumers and the presence of Helicobacter in water samples.

Scientists plan subterranean lake exploration

Scientists from 14 countries have begun preparations to study a huge lake more than 2 miles under the Antarctic ice cap. Lake Vostok is believed to be one of the oldest of about 80 lakes formed by heat from the earth's core and pressure from the icecap above. The lake is permanently pressurised, cold and dark, and may have formed as long as 40 million years ago. Despite these extreme conditions, some researchers speculate that microorganisms may have evolved there.

An expedition using a small robotic submarine to explore Lake Vostok is planned to occur within 5 years. Knowledge gained from this project will be used by NASA to develop an exploration program to search for alien life forms in the oceans of Europa, one of Jupiter's moons.

Legionnaires disease outbreak in Belgium

Health authorities in Belgium are investigating an outbreak of Legionnaires disease at a trade fair in Kapellan in November. More than 80 of the 60,000 visitors to the exhibition developed symptoms of the infection and there have been at least 4 fatalities. Investigators believe whirlpool spa baths on display at the exhibition are likely to be the source of the outbreak.

Spa baths at a similar trade fair were responsible for a Legionnaires disease outbreak in the town of Bovenkarspel in the Netherlands in March. In this outbreak 226 confirmed cases were diagnosed and at least 18 people died.

Pumping water is child's play

An enterprising South African company has solved the problem of how to supply energy for extracting ground water by incorporating a pump into a children's merry-go-round. Sixty of the novel water pumps have already been installed in rural areas of the country.

The low maintenance "Playpumps" are capable of pumping up to 1,400 litres per hour to a 5,000 litre overhead storage tank, whereas conventional handpumps can pump only 150 litres per hour to ground level. The higher cost of the playpumps (about $7,000 in contrast to $3,300) has been offset by selling advertising space on the storage tanks, which are a natural meeting place for villagers collecting water.

From the Literature

Contact Information

Editor - Martha Sinclair	email martha.sinclair@med.monash.edu.au
Assistant Editor - Pam Lightbody	email pam.lightbody@med.monash.edu.au

The printed version of Health Stream is available free of charge - to be added to our mailing list please contact Pam Lightbody (email above or fax + 61 3 9903 0576). Past issues can be found under Publications.