This paper summarizes relevant information and current research issues on Cryptosporidium relevant to its public health significance in water supplied. From this assessment, suggestions are made for actions to be taken by the San Francisco Public Utilities Commission

Background

To safeguard public health, regulations have been promulgated regarding drinking water authorized by the 1986 Amendments to the Safe Drinking Water Act. A pertinent piece of legislation is the Surface Water Treatment Rule, which was enacted by both the United States Environmental Protection Agency and the State Department of Health Services (DHS) in June 1989 and 1991, respectively. San Francisco is under a compliance order to meet this regulation by 1999. The purpose of the Safe Drinking Water Treatment Rule is to minimize microbial risk due to parasites (i.e., Giardia), bacteria (i.e., Legionella) and viruses from surface water supplies. While this rule does not place limits on Cryptosporidium due to lack of information, it has, however, become of great interest due to recent outbreaks in England and in North America. Cryptosporidium will be regulated under the Enhanced Surface Water Treatment Rule (interim rule to be promulgated in 1998 with a long-term version to be promulgated in 2002); watershed monitoring will be an essential element as well as removal or inactivation requirements.

History

Cryptosporidium was first described in 1907 by Ernest Edward Tizzer. His work was not regarded as important at the time, and half a century passed before Cryptosporidium became of minor interest in association with the incidence of cryptosporidiosis in turkeys. Interest in Cryptosporidium heightened in 1971 when Cryptosporidium was found to be associated with diarrhea in cows.⁵⁴ In 1976, the first cases of human cryptosporidiosis were reported. After that, relatively few cases were reported until 1982, when cryptosporidiosis was associated with protracted diarrhea in patients with acquired immune deficiency syndrome (AIDS).^7,18,41 This finding stimulated intense medical and veterinary interest in the epidemiology, diagnosis, treatment, and prevention of cryptosporidiosis.

The first reported human outbreak of cryptosporidiosis due to water supply occurred in Texas in 1984 concurrently with an outbreak of Norwalk virus. This was followed by the second largest North American outbreak in Carrolton, Georgia, in 1987, where over 13,000 people were affected.²⁸ Two outbreaks of Cryptosporidium occurred in the United Kingdom in 1988; the larger in December 1988 affected approximately 5,000 people.⁴² Since this time, several smaller outbreaks have occurred in the United Kingdom.^27,28 In April 1993, the largest North American outbreak affecting almost 400,000 people occurred in Milwaukee, Wisconsin.³⁹ This outbreak has attracted much national attention and the effects of the outbreak were still experienced even a year later.^33,50,58 In early 1994, an outbreak occurred in the Las Vegas area affecting approximately 78 people. Water was speculated to be the cause.²³ In 1996, several outbreaks occurred in Canada including approximately 14,500 cases in Kelowna and British Columbia.¹⁵ In 1997, an outbreak affecting four suburbs in London, England infecting at least 345 people. The cases were strongly associated with drinking water.⁵⁹ Most water supply related incidents of Cryptosporidium have occurred during the spring and in filtered supplies.

The Organism

Description

Cryptosporidium is an oval-shaped protozoan parasite found in man, mammals, birds, fish, and reptiles. Cryptosporidium has a complicated life cycle (Figure 1) which goes through many forms, the most relevant form being a 4 to 6 Fm diameter oocyst, which contains the infective sporozoites.^16,20 Although the number of species of Cryptosporidium is open to questions, only one, C. parvum, appears responsible for significant human health concerns.

Cryptosporidium oocysts are resistant to adverse environmental factors and can survive for months under optimum environmental conditions.⁴⁶ The released infective sporozoites do not survive well.

The Disease

In humans, cryptosporidiosis results in a self-limiting but unpleasant diarrhea in immunocompetent individuals with an incubation period of 2 to 10 days. Some of the associated symptoms include anorexia, weight loss, dehydration, abdominal cramping, and vomiting (i.e., headache, aching muscles, fever). On average, the symptoms last for 12 days with rare instances lasting as long as 4 weeks.^4,7,14 In patients with depressed immunity due to disease (i.e., HIV infection, chemotherapy, etc.) or congenitally depressed immunity (e.g., hypogammaglobuinemia), similar symptoms are observed. The duration, however, can be much longer and some individuals never clear Cryptosporidium from their systems. In cases where suppression of the immune system cannot be reversed (e.g., by stopping immune suppressant therapy) these symptoms may persist until death.^6,11,24

Prevalence

Methods for determining the prevalence of infection with Cryptosporidium have been undergoing rapid change. Interpretation of the results is not without controversy. Nevertheless, human cryptosporidiosis has been identified on all six continents.3,5,14,55 Among people with gastrointestinal complaints, the prevalence of Cryptosporidium oocysts in patient's stools range from 1 to 4 percent in developed countries and up to 16 percent in developing countries. Specific North American surveys indicated similar levels. For example, in British Columbia, Canada, the prevalence was 0.6 percent, in Massachusetts it was 2.8 percent, and in South Carolina 4.3 percent. In England, Cryptosporidium is the fourth most common cause of diarrhea.53

In the immunocompromised population, particularly those infected with HIV, the prevalence is higher. Estimates of the percent of HIV-infected patients with cryptosporidiosis range from 1 to 2 percent up to 10 percent. Of AIDS patients with diarrhea, Cryptosporidium has been identified as an agent in 10 to 15 percent of the cases.6,55

Immunity

Exposure to Cryptosporidium does not necessarily lead to clinical disease. There is some indication that prior exposure results in protective immunity from cryptosporidiosis, though the duration of this immunity is unknown. Serological testing has found Cryptosporidium-associated antibodies in 25 to 35 percent of people tested in North America, indicative of a moderate level of Cryptosporidium exposure.⁵⁵ With the more widespread use of cocktail therapy, the health of AIDS patients has been improving. It is unclear what impact this will have on their susceptibility to cryptosporidiosis.

Treatment

Over 90 antimicrobial agents have been used against Cryptosporidium in animals and man, but no specific effective treatment for cryptosporidiosis has yet been found.²⁰ While this is not of great importance among immunocompetent individuals (except for days of work lost) where the infection is self-limiting, it is vital for immunosuppressed patients. Several clinical trials are currently evaluating some promising agents including letrazuril, azithromycin, paramomycin, and a hyperimmune bovine colostral immunoglobulin. The last mentioned agent shows great promise; it is being evaluated at San Francisco General Hospital.⁴¹

Transmission

Accurate identification of the modes of transmission is of critical importance in evaluating the risks associated with cryptosporidiosis.

Sources

Reservoirs for Cryptosporidium include mammals, birds, fish, and reptiles. Species shed by one host can infect other hosts (e.g., cattle can infect humans). Animals that can carry Cryptosporidium include cattle, pigs, cats, deer, guinea pigs, mice, rats, and sheep. Young animals that are most likely to develop symptomatic infections and shed high quantities of infective oocysts.

Exposure Routes

There are a variety of exposure routes. Cryptosporidiosis is normally transmitted by the fecal-oral route, when oocysts excreted by an infected animal or human are ingested by a susceptible person. A number of transmission routes exist because oocysts are capable of infecting other hosts immediately when released into the environment and after surviving in the environment for a period of time. These routes are:

• From Animals - Transmission from animals to man has been documented by veterinarians and others working closely with sheep and cows. Gulls have been shown to be carriers.⁵² Pet animals such as dogs and cats have also been implicated in human cases.⁴⁴
  
  
• Person-to-Person - Person-to-person transmission is important. Cryptosporidiosis transmission can occur easily within families, play groups, nursery schools, day care centers, hospitals, and other institutions where precautions are not taken. Sexual transmission is also suspected, particularly in the gay community.²⁶
  
  
• From Water - Several major outbreaks in the last 10 years have shown that cryptosporidiosis can be contracted from contaminated water. Wastewater plants are not wholly effective in removing Cryptosporidium.⁵⁶ Recreational contact with contaminated water in reservoirs, rivers, and other waters has been documented as a transmission source. Included in this category is an incident where 17 people contracted cryptosporidiosis while at a local wave pool.³⁵
  
  
• Airborne - Some indications exist that airborne spread of Cryptosporidium occurs, although these have not been well documented.
  
  
• Other - There have been a few reports of infection with Cryptosporidium following consumption of raw milk and raw sausages. In 1993, 160 people contracted cryptosporidiosis due to drinking contaminated apple cider at an agricultural fair in Maine.³⁷ This has not been previously thought to be a significant route.

Infectivity

Uncertainty exists concerning the dose required to induce infection (not even considering virulence differences for different strains). While most indications suggest that the dose required to induce infection is between 1 to 100 oocysts,^3,21,41 one study indicated that doses of 10,000 oocysts were not capable of inducing infections in adult monkeys. Recently, Dupont and his colleagues completed a human feeding study which determined that the dose at which 20 percent of the subjects were infected was 30 oocysts while a 40 percent infection rate of the subjects was 100 oocysts (Figure 2). From this information, Haas and Rose propose that the minimum infectious dose is 1 oocyst.²⁷

The immunocompromised apparently are the most susceptible population. There is some indication that deficiencies in the immune system during pregnancy make pregnant women more susceptible to a prolonged bout of cryptosporidiosis. It is reported that when CD4 counts (CD4 counts indicate the level of T-helper cells in the immune system) are less than 180, the host will usually be unable to clear the disease.^34,41 The CD4 counts for normal individuals typically are between 800 and 900.

Significant virulence differences exist between strains of Cryptosporidium. For example, comparing two different isolates showed a similar rate of infection but twice the incidence of diarrheal illness.¹²

Relative Public Health Significance

Assessing the relative public health significance of Cryptosporidium is complex because of the different responses between immunocompetent and immuno-compromised individuals. Sexually transmitted diseases typically account for the highest incidence of infectious diseases followed by gastrointestinal illnesses.

Major Facts

• Several points frame the question of public health significance:
  
  
• The prevalence of cryptosporidiosis is between 1 and 4 percent of the total population in North America.⁵⁵
  
  
• A number of waterborne disease outbreaks have been associated with Cryptosporidium.³⁸
  
  
• Cryptosporidiosis is usually self-limiting, except in immunocompromised individuals.⁴¹
  
  
• Groups at risk include:
  • Animal handlers
  • Health care workers
  • Day care center children/employees
  • Consumers of contaminated water
  • Travelers to developing countries
  • Immunodeficient and immunosuppressed persons
    
    - Congenital deficiency
    - Acquired deficiency
    - Immunosuppressive therapy
    - Malnourished

• Incidence of cryptosporidiosis amongst AIDS patients is not well known, but is estimated to be between 1 and 10 percent.^6,41 Of these infected patients, it is not known how many may die directly from cryptosporidiosis, but numbers as high as 20 percent have been speculated.⁵⁵
  
  
• No therapeutic agent has been found to treat cryptosporidiosis.
  
  
• In some cases where Cryptosporidium has been detected, other pathogenic organisms have also been detected (e.g., Giardia, rotavirus).¹⁴
  
  

Other Infectious Organisms

Cryptosporidium is not the only organism which causes diarrheal symptoms.^{5,11,14,26,36,42} Immunocompromised individuals, particularly with those with AIDS, can be infected by a variety of other diarrhea causing organisms, including:

• Cytomegalovirus
• Mycobacterium avium
• Salmonella
• Entomoeba histolytica
• Giardia lamblia
• Herpes simplex
• Campylobacter jejuni
• Isopora belli
• Campylobacter difficile
• Candida
• Strongyloides
• Enterocytozoon bieneusi

Some of these organisms (e.g., cytomegalovirus and Isopora belli) have been detected concurrently with Cryptosporidium during bouts of infectious diarrhea.^11,26,36,42

Summary

Cryptosporidium is one of several agents involved in infectious diarrhea and is particularly devastating for immunocompromised individuals who are unable to clear the disease. While the incidence of reported cryptosporidiosis appears to be low, this may be underestimated due to factors such as lack of reporting by doctors and lack of diagnosis. The treatment of cryptosporidiosis has been relatively unsuccessful, although there appears to be some promise in use of a hyperimmune bovine colostrum.

Water Treatment

Occurrence in Water

Cryptosporidium is shed from infected individuals in their stools. Concentrations of infective oocysts are very high in the stools with levels on the order of one million oocysts per day being estimated. Cryptosporidium can be transmitted directly from person to person through the fecal-oral route. Cryptosporidium can also find its way into the environment and, hence, into drinking water sources.

Detection

Because the conventional indicators of microbial water quality (e.g., coliforms and heterotrophic plate counts) do not necessarily correlate with the presence or concentrations of Cryptosporidium and because the minimum infective dose is thought to be very low, detection of low Cryptosporidium concentrations is necessary. To accomplish this, methods have been developed that rely on concentrating large volumes of water (i.e., 100 to 1000 gallons) into a small pellet (ASTM Method P229). The method is detailed in the proposed Information Collection Rule. The basis of the procedure is as follows:

• Sampling - Water is taken from the source by pumping into a filter housing containing a polypropylene yarn cartridge filter.⁵⁴ The volume of water passed through the filter is measured using a water meter.
  
  
• Concentration - After transport to the laboratory, the particles trapped on the filter are eluted using large volumes of detergent. The eluate is then centrifuged to concentrate and separate out particles denser than the oocysts.^48,54
  
  
• Identification - Microscopic examination of the concentrated sample relies on actual measurement of size and the use of fluorescent antibody stains to identify the oocysts.¹⁶ A distinguishing feature of Cryptosporidium is the fold in the oocysts.
  
  

There are several problems with current detection methods.^1,2,47,48

• The species of Cryptosporidium cannot be distinguished using conventional methods.
  
  
• Some of the antibodies used for detection may cross-react with other organisms (e.g., yeasts) so that enumeration of oocysts may include species (or other organisms) that are not infectious to humans.^14,48
  
  
• Detection of the oocysts does not indicate whether they are viable (i.e., capable of inducing infection).
  
  
• There is no method for assessing the virulence (i.e., the severity of the disease produced) of a particular strain of Cryptosporidium.
  
  
• Current concentration techniques result in recoveries in the range of only 20 to 70 percent with the efficiency being a function of the water matrix.⁴⁸
  
  
• Variations between laboratories for identical samples can be as high as 100 percent, even through improvements to standardized procedures used for the EPA's Information Collection Rule sampling, these wide variations have somewhat reduced.
  

Development work has been conducted on better concentration techniques (e.g., immunomagnetic separation), specific identifications with DNA probe technology (i.e., using polymerase chain reaction) , and assessment of viability (i.e., through ELISA/RT-PCR or PI/DAPI staining).

Due to the significant uncertainties in the methods, the numerical values produced are only rough estimates. This creates challenges for communicating data to the general public.

There is some indication that particle counting may be a useful surrogate for assessing efficiency of Cryptosporidium removal by water treatment processes.

North American Source Waters

Typical geometric average concentrations for various water types have been as follows:⁴⁸

• Lakes - 0.44 oocysts per liter
• Rivers - 0.43 oocysts per liter
• Springs - 0.04 oocysts per liter
• Groundwater - 0.003 oocysts per liter

In North American waters, the values of Cryptosporidium range from 0.002 to 5,800 oocysts per liter, depending on the source. These values will vary depending on the watershed characteristics of the water source. Concentrations of Cryptosporidium oocysts in source water tend to be higher than Giardia.^2,48

North American Treated Waters

Cryptosporidium oocysts have been detected in treated waters in the western United States. These values tend to be low, averaging 0.001 oocysts per liter in filtered waters and 0.006 oocysts per liter in non-filtered waters. In waterborne outbreaks Cryptosporidium oocyst concentrations in the treated water were much higher than these values. For example, in the Carrolton, Georgia outbreak, treated water oocyst levels were 2.2 per liter. After the outbreak in Milwaukee, treated water oocysts levels reached 0.16 per liter. However, there is significant uncertainty as to how high the levels were during the outbreak.

Outbreaks have been more often associated with filtered waters than unfiltered waters, and usually with agricultural (particularly animal wastes) contamination of drinking water sources. In a filtration process, Cryptosporidium oocysts, Giardia cysts, other pathogens and debris are concentrated in the filters so that breakthrough of the accumulated material can increase the risk of Cryptosporidium infections.

Effectiveness of Water Treatment Processes

The two basic mechanisms for eliminating pathogenic organisms during water treatment are: chemical inactivation and physical removal. The former is accomplished through disinfection, and the latter through coagulation and filtration.

Disinfection

Chlorine is not effective for inactivating Cryptosporidium oocysts. It has been reported that oocysts exposed to undiluted household bleach (5% NaOCl) for several hours were still capable of inducing infection.⁵¹ Chlorine dioxide appears to be effective, but at doses far higher than would be reasonable in water treatment, especially in light of the concern over the chlorine dioxide by-products chlorite and chlorate.²⁴ Ozone is effective for inactivating Cryptosporidium but requires Ct values 10 to 20 times greater than for Giardia.²² Ultraviolet light is effective at inactivating Cryptosporidium though at very high doses.³² Ct values for various disinfectants against Cryptosporidium and Giardia are presented in Table 1.^3,22,24,54 Sequential use of these disinfectants, however, results in synergestic effects.

1. Assume 2-logs inactivation and 10^oC.
2. Units in mg/l-min.

Even though ozone appears to be effective for Cryptosporidium inactivation, there is a major concern regarding its use. Due to its powerful oxidizing ability, it breaks down the naturally occurring organic matter in water, causing it to become a nutrient source for bacteria. It is possible, then, that the use of ozone, while inactivating Cryptosporidium, could stimulate the growth of bacteria in the distribution system. Some of these bacteria may cause disease, particularly in immunocompromised individuals.

Filtration

Removal of Cryptosporidium by filtration is approximately 99 percent or 2 logs of reduction. Although levels as high as 99.9 percent or 3 logs of reduction have been reported. At times oocysts can be detected in filtered water; their breakthrough can be attributed to a variety of factors, including:

• Increases in source water concentrations of Cryptosporidium.
  
  
• Recycling filter washwater in the plant enabling concentrated slugs of Cryptosporidium to pass through the filters.
  
  
• Operational factors such as improper filter washing, rapid flow changes, improper coagulation, etc.
  
  

Most waterborne Cryptosporidium outbreaks have been associated with operational problems rather than inherent treatment deficiencies.

Regulatory Stance

The regulatory stance varies considerably. The activities of the United Kingdom Drinking Water Inspectorate, the United States Environmental Protection Agency, the California Department of Health Services, and Canadian Authorities are summarized below.

United Kingdom

Regulatory authorities in the United Kingdom have the most extensive experience with the issue because of the number of outbreaks experienced. Their stance on Cryptosporidium is summarized in two major documents: The Badenoch Report entitled "Cryptosporidium in Water Supplies" (July 1990) and the Drinking Water Inspectorate's report "Cryptosporidium in Water Supplies: Progress With the National Research Program" (July 1992). The basic recommendation of these reports is that water treatment plants need to optimize their practices with respect to coagulation, filtration, and recycling of filter washwater.⁸ The Drinking Water Inspectorate is leading an aggressive research program to address the major research needs.

EPA

Cryptosporidium was not regulated under the 1989 Surface Water Treatment Rule due to the uncertainties surrounding Cryptosporidium (i.e., virulence, infective dose, inactivation, etc.). Source monitoring for Cryptosporidium will be required sometime in 1996 under the Information Collection Rule (ICR). Methodological difficulties with the detection method have slowed down promulgation of the ICR though monitoring finally began in 1997. The proposed interim Enhanced Surface Water Treatment Rule (ESWTR)¹⁹ is in its final stages of revision pending comments to the Notice of Data Availability (November 4, 1997). Some significant items noted in the draft ESWTR Preamble (dated January 26, 1994) are:

• Differing pathogen densities in different waters accounts for the drive to develop site-specific treatment requirements.
  
  
• In light of current and anticipated research EPA senses that they will "soon be in better position to develop a suitable regulation for Cryptosporidium".
  
  
• There appears to be some conferred immunity associated with exposure to Cryptosporidium. A 2-log treatment requirement has been proposed as an option for sources with less than 1 oocyst per 100 liters, though there are other proposals.
  
  
• EPA recognizes that "unfiltered systems would appear to be particularly vulnerable...However, to date, filtered water supplies have been implicated in all identified waterborne Cryptosporidium outbreaks...(filtered) surface water may be more vulnerable to Cryptosporidium than unfiltered supplies with disinfection, depending on the quality of the source water..."
  
  

Great uncertainty still remains as EPA approaches a regulation. Consequently, they are soliciting input into the structuring of the long-term ESWTR. A major emphasis for unfiltered systems will be on watershed controls.

In the 1996 Amendments to the Safe Drinking Water Act, a special section detailed the conditions for a limited alternative to filtration. The text says:

"...a State exercising primary enforcement responsibility for public water systems may, on a case-by-case basis, and after notice and opportunity for public comment, establish treatment requirements as an alternative to filtration in the case of systems having uninhabited, undeveloped watersheds in consolidated ownership, and having control over access to, and activities in, those watersheds, if the State determines (and the Administrator concurs) that the quality of the source water and the alternative treatment requirements established by the State ensure greater removal or inactivation efficiencies of pathogenic organisms for which national primary drinking water regulations have been promulgated or that are of public health concern than would be achieved by the combination of filtration and chlorine disinfection..."

California Department of Health Services (DHS)

DHS is concerned over the status of unfiltered supplies in the state. While anticipating that Cryptosporidium will need to be regulated in the near future, DHS will await for EPA's action. In the interim, DHS has issued a Cryptosporidium Action Plan which places heavy emphasis on treatment optimization.

Centers for Disease Control and Prevention (CDC)

In September 1994, CDC convened a meeting to address concerns over waterborne cryptosporidiosis involving representatives from regulatory and public health agencies, water utilities, and advocacy groups. The focus of the workshop was to avoid unnecessary alarm (i.e., premature or unwarranted boil water advisories) and preventing waterborne outbreaks. Four workgroups addressed the special topics: surveillance systems and epidemiological study designs, public health responses, immunocompromised persons, and water sampling methods and interpretation of results. A document published in 1995¹⁰ noted that there was significant uncertainty regarding waterborne cryptosporidiosis and as such recommended increased surveillance and epidemiological investigations, methods development for Cryptosporidium detection in drinking water, and development of task forces for providing information to the immunocompromised as well as regional population.

Canadian Authorities

There are no current regulations for Cryptosporidium or Giardia, though Giardia has been identified as a public health concern. The only regulatory guidance on microbial water quality is for coliform levels in the distribution system. There are no current plans to regulate Cryptosporidium.

Summary

Of all the regulatory agencies contacted, EPA appears to be the most aggressive in pursuing a Cryptosporidium regulation, but is hampered by poor analytical methods. In the wake of the Las Vegas and Milwaukee outbreaks, EPA is continuing its efforts. Other organizations, due to the meagerness of available information, have decided not to regulate Cryptosporidium at this time.

Current Activities of the Water Industry

A number of water agencies in North America and Europe were contacted to determine their activities with respect to Cryptosporidium. There appears to be much interest in assessing Cryptosporidium levels and optimizing treatment practices to minimize its passage into the water supply.^2,3,43,45 Activities include:

• Monitoring of source and treated waters for Cryptosporidium,
• Optimizing coagulation practices,
• Monitoring turbidity of individual filter cells,
• Backwashing filters prior to restarting,
• Evaluating filter washwater treatment (i.e., clarification and/or disinfection),
• Setting limits for turbidity spikes occurring after filter restarts (i.e., ripening period),
• Controlling filters in a manner to avoid sudden flow changes,
• Optimizing disinfection, and
• Watershed management.

For unfiltered sources there has been a strong emphasis on watershed management,⁴⁵ though some (e.g., New York City and Boston) are receiving pressure to filter.^40,57 Interestingly, few recommendations have been made by the medical community for immunocompromised individuals to seek alternative water sources,⁴¹ even in Milwaukee.^49,50