Department of the Environment,
Transport and the Regions

Cryptosporidium in Water Supplies

3 Lessons learnt from outbreaks of waterborne cryptosporidiosis

3.1 Review of incidents

3.1.1 This section reviews incidents during the last ten years in which Cryptosporidium oocysts were detected in public drinking water supplies or where there was a suspected public drinking water related outbreak of cryptosporidiosis. The objective of the review is to establish whether there was any correlation between outbreaks of cryptosporidiosis and events which might pose a challenge to the integrity of the treatment or distribution process.

3.1.2 The Group is aware of twenty five outbreaks of cryptosporidiosis in the United Kingdom since 1988 that have been associated with the consumption of public drinking water supplies. These are listed in Table 3.1. The association with mains water varies from possible to probable to strong. Fourteen incidents have involved increased reporting of cases of cryptosporidiosis but in the absence of any reported detection of oocysts in water supplies. These may illustrate the limitations of current sampling and analytical techniques or indicate that the source was not the water supply. In eleven incidents, an increase in reported cases of cryptosporidiosis was associated with detection of oocysts in water supplies. The Group is also aware of eighteen notifications in England and Wales since 1990 involving detection of oocysts in water supplies but without any detectable increase in the level of cryptosporidiosis in the community which supports the practice of not over-reacting to low levels of oocysts in water supplies.

3.1.3 Most outbreaks occurred in situations where oocysts were not detected in water supplies. The tabulated data indicate that the majority of outbreaks of waterborne cryptosporidiosis occurred in situations in which the integrity of treatment had been compromised or where the treatment provided may have been less than adequate. In only two incidents was the correlation less than convincing.

3.1.4 Research funded by the Department of the Environment, Transport and the Regions (DETR 1998) on modelling the risk of infection has indicated that pathogens are non-randomly distributed in drinking water and may be associated with particles arising from turbidity events. It is significant that the majority of outbreaks identified were associated with situations in which turbidity increased although the regulatory standard for turbidity was not necessarily contravened. The groundwater incidents confirm that contamination of groundwater can be a significant risk factor. Time of year and climatic conditions may also be a risk factor, given that most incidents were reported during the late autumn to early spring period.

3.1.5 The Group concluded from its examination of these incidents that outbreaks of water related cryptosporidiosis do not just 'happen'. There appears to be a strong correlation between outbreaks and situations where an inadequacy was identified in the treatment provided or in the operation of the treatment process, or where there was overloading of the treatment process. The unifying factor in all situations was the potential for peaks in turbidity to have been present in the treated water leaving the works. The fact that turbidity events were not recognised in all cases could be a reflection of inadequacy in the continuity of turbidity monitoring or in the calibration and control of the equipment.

Table 3.1 UK outbreaks of cryptosporidiosis associated with public drinking water supplies April 1988 - April 1998


Date

Source/treatment characteristics (all water received normal chlorine based disinfection)

Oocysts detected in treated water

Approximate number of cases of cryptosporidiosis

Association with water1

Conclusion
/Comments5

Apr 1988Surface water with coagulation and rapid gravity filtrationYes27Strong2Agricultural slurry contamination of water in distribution
Mar 1989Impounded reservoir supply, coagulation, rapid gravity filtrationYes500+Strong2Contamination of source water with animal wastes, breakthrough of treatment
Mar 1989Surface water, coagulation, rapid gravity filtrationyesNumber of cases included in 500+ reported abovestrong2River flows abnormally low, severe diarrhoea in cattle upstream of intake
Apr 1989Surface water, no filtrationYes244Probable4Unfiltered water, potential point source discharge from sewage treatment works and farm drains and non-point discharge from grazing animals
Dec 1989Lowland river with bankside storage, roughing filters and slow sand filtrationNo477Strong4Outbreak followed by-passing of filters
Dec 1990Lowland tidally influenced river, direct abstraction, gravity filtrationNo47Probable2Rapid fluctuation in source water quality at the time of the outbreak
Apr 1991Spring, well and stream supply, crude filtrationNo5Possible4Possible agricultural contamination of well supply, inadequate treatment
Apr 1992Surface water, no filtrationYes50Probable4Unfiltered water, potential point source discharge from sewage treatment works and farm drains and non-point discharge from grazing animals
Jul 1992Lowlands river with direct abstraction, separation process and bankside infiltration with no filtrationNo108Probable3Possible link with groundwater turbidity
Nov 1992Upland reservoir and surface water supplying aqueduct, slow sand filtrationYes125Strong2Heavy rainfall in the catchment, high raw water turbidy, increase in treated water turbidity
Nov 1992Groundwater, no filtrationNo47Probable3Outbreak probably caused by faecal contamination from cattle housed adjacent to the well head
Apr 1993Stream source, no filtrationYes3Probable4Inadequate treatment, source open to potential animal contamination
Apr 1993Groundwater from fissured strata, no filtrationNo40Probable3Outbreak possibly caused by rapid recharge of surface water contaminated with oocysts
Apr 1993Upland reservoir, no filtrationNo48Probable4Possible run-off from grazing, very heavy rainfall
Jun 1993Upland reservoir and surface water supplying aqueduct slow sand filtrationNo97Probable3Outbreak caused by poor operating practices and excessive head on the filters
Jun 1994Spring fed natural impoundment upflow filtrationNo8Probable4Possible animal waste contamination following heavy rain
Feb 1995Spring supply, no filtrationNo40Strong4Heavy rain washed in waste animal material
Aug 1995Lowland river with direct abstraction, separation process and bankside infiltration with no filtrationYes575Strong3Plant operating above design output, evidence of turbidity peaks in the bankside infiltration water
Jan 1996Lowland river with coagulation and filtrationYes126Strong4Outbreak occured when works was under strain with excess flow causing solids breakthrough
Mar 1996Surface water with bankside storage, rapid gravity filtration, no coagulationNo20Probable4Outbreak caused by breakthrough of solids as a result of inadequate treatment during an algal bloom
Apr 1996Lowland river with full treatmentNo80Probable4Probable association with water but no evidence of plant operation problems
Feb 1997Groundwater from fissured strata, no filtrationYes345Probable4Outbreak caused by infiltration of surface water containing oocysts
Feb 1997Lowland river, coagulation and filtrationNo22Probable4Outbreak possibly associated with turbidity peak in filtered water
May 1997Spring supply, partial filtrationNo34Possible4Possible run-off from spring grazing
Apr 1998Surface water, no filtrationYes303Possible4Unfiltered water, potential point source discharge from sewage treatment works and farm drains and non-point discharge from grazing animals

1 The definitions of strong, probable and possible association with water are given in Tillet et al (1998).

2 Tillet et al (1998).

3 Furtado et al (1998)

4 Personal communication from Expert Group Secretariat.

5 Conclusions and comments are those of the Drinking Water Inspectorate, health authorities, SCIEH, DWI NI or water utilities

3.1.6 The Group considers that the circumstances of these incidents do not alter the conclusion from the previous Expert Group reports (Badenoch 1990; 1995) that well operated appropriate conventional water treatment plants using processes designed for removal of particulate material minimise the risk of Cryptosporidium contaminating drinking water supplies. However, in the light of its assessment of incidents the Group has a number of recommendations. These are listed in Chapter 5, Advice to Water Utilities.

3.1.7 The Drinking Water Inspectorate (DWI) has published reports on four incidents in England where public drinking water supplies were associated with outbreaks of cryptosporidiosis (DWI 1993; 1997a; 1997b; 1998). One of these incidents is considered in the next section.

3.2 Cryptosporidiosis in north west London and Hertfordshire, spring 1997

3.2.1 The outbreak of cryptosporidiosis in north west London and Hertfordshire, in early spring 1997, with 345 confirmed cases, which was associated with drinking water derived from underground strata was one of the main reasons for re-convening the Expert Group. The DWI assessment of the incident was published in June 1998 (DWI 1998). It made recommendations specifically to the water utility involved on matters concerned with water quality monitoring, emergency procedures, risk assessment of groundwater sources, and implementation of recommendations contained in the two Reports of the Expert Group on Cryptosporidium in Water Supplies. It also made recommendations to all water utilities on matters concerned with risk assessment of groundwater sources and water quality monitoring; and to the Incident Management Team on matters associated with communications. The Group was concerned to note that it was necessary for the DWI report to refer to implementation of the recommendations from the two previous Expert Group reports. As most of these recommendations are still sound and relevant

the Group recommends that all water utilities review their implementation of the Badenoch recommendations.

The main Badenoch recommendations are listed in Appendix A1.

3.2.2 The DWI assessment of the north west London and Hertfordshire incident concluded that there were lessons to be learnt for all water utilities.It recommended that all water utilities:

(i) carry out risk assessments of groundwater sources to identify vulnerable sources requiring more rigorous surveillance;

(ii) consider the routine use of tracer tests as a measure of groundwater source vulnerability to pollution, especially those sources identified by risk assessment as most susceptible;

(iii) increase surveillance of shaft and adit chalk water sources vulnerable to fast infiltration of recharge; and 

(iv) review their operational sampling programmes and establish where none exists or where the programme is shown to be deficient, a regular sampling and analytical programme for all separate sources of raw water.

3.2.3 The DWI report also recommended that a national survey of shaft and adit systems be undertaken to provide a better understanding of such systems, especially in chalk aquifers.

3.2.4 The Expert Group endorses these recommendations to water utilities as part of their risk assessment programme.

3.3 Identifying and reporting the potential presence of oocysts in drinking water

3.3.1 The Group considers that there are four broad ways that a water utility could become aware of oocysts in the water supply. It notes that there are differences in the way this information would be handled in England and Wales, Scotland and Northern Ireland.

3.3.2 Water utility detects oocysts in a sample of treated water. In England and Wales and Northern Ireland the laboratory would normally notify the local operations manager or scientist. Further investigations would be made and the local authority environmental health department and local health authority Consultant in Communicable Disease Control (CCDC) or equivalent should be informed in accordance with locally agreed arrangements for notification. In Scotland, notification within the utility is the same but the local authority, Consultant in Public Health Medicine (CPHM) and the Scottish Office are informed.

Recommendations

3.3.3 The Group recommends that the water utility should investigate immediately when oocysts are detected in raw water to establish if any circumstances exist to allow Cryptosporidium to enter water supplies. Investigations should include review of recent treatment plant operational data.

3.3.4 The water utility should develop local liaison arrangements with the local authority and health authority for rapid appraisal of the potential health risk, particularly when oocysts are detected in final water or in distribution. 

3.3.5 Water utility detects a change of operational circumstance which potentially risks oocysts contaminating drinking water supplies.Water utility response is essentially the same in England and Wales, Scotland and Northern Ireland with investigation by a treatment specialist or local operations manager. The most crucial step however, is early recognition and interpretation of the change of operational circumstance by a plant operative or local manager.

Recommendation

3.3.6 The Group recommends that water utilities should ensure that employees operating assets producing drinking water are aware of the types of circumstance which can potentially put water supplies at risk of Cryptosporidium contamination. Procedures should be in place to ensure rapid recognition and appraisal of risks associated with any relevant change in operational circumstance. 

3.3.7 Local health authority gets notification of an increase in stools containing Cryptosporidium oocysts. There are important differences in reporting practice between England and Wales and Northern Ireland and Scotland. In Scotland details on every positive stool are sent by most CPHMs to the water utility for it to provide information on water supply source and treatment plant. The CPHM then collates these reports and comes to a view if there is evidence of a water-related clustering of cases. In England and Wales and Northern Ireland, the CCDC receives stool notification but rarely contacts the water utility immediately unless there is evidence accumulating to indicate clustering or patterns above expected background rates. Only when the CCDC is concerned about clusters or higher frequencies of isolation will the water utility be routinely asked to provide details of the water supply to each case. Ready availability to the CCDC/CPHM of local water supply zone maps will aid cluster identification.

3.3.8 The Group is anxious to avoid over or under reaction by health authorities and to not recommend a burdensome administrative process.

Recommendation

3.3.9 The Group recommends that water utilities should provide copies of water supply zone maps to CCDCs/CPHMs or their equivalent and health authorities should make early contact with the local water utility if an outbreak of cryptosporidiosis is suspected.

3.3.10 The national disease surveillance centre is the first to detect a cluster of cases. It is possible for the Centre for Disease Surveillance and Control (CDSC) for England and Wales and the Scottish Centre for Infection and Environmental Health (SCIEH), to establish case clustering when epidemiological data are pooled from each health authority area. It is usual for both centres to report findings to the local medical officer rather than directly to a water utility. The Group is aware that case reporting rates to CDSC and to SCIEH differ because cryptosporidiosis is a laboratory reportable disease in Scotland, but these centres routinely monitor outbreaks. For that reason, these centres do have a role to play in making drinking water safer. The national centres are able to show, over time, areas where there are higher than average rates of cryptosporidiosis incidence. However, in England and Wales, any attempt to link with water supplies is presently constrained, as cases are not coded by postcode. This makes integration of case locations into utility Geographic Information Systems (GIS) time consuming and inaccurate.

Recommendations

3.3.11 Human cryptosporidiosis should be made a laboratory reportable disease in England and Wales and consideration should be given to making the disease notifiable.This should provide earlier identification of increased incidence of cryptosporidiosis and greater consistency in case reporting and recording.

3.3.12 The Group recommends that, wherever possible, health authorities should make postcodes of cases of human cryptosporidiosis available to water utilities to help both organisations identify as early as possible if particular water sources are involved and to allow regional trends to be assessed.

3.4 Risk assessment

3.4.1 Research has highlighted the ability of oocysts of Cryptosporidium to survive in the aquatic environment and their tolerance of the disinfection process used in conventional water treatment. A consequence of this resistance to disinfection is that oocysts of Cryptosporidium can still be present when coliform bacteria, the traditional indicators of drinking water quality, have been inactivated.

3.4.2 To try and ensure freedom from risk of Cryptosporidium, other control strategies have had to be considered. Waterborne outbreaks occur even though oocysts cannot be detected in the water. This supports the general view that the 'contamination' occurs for only a few hours during which time it would be complete chance that routine samples coincided with the event. Random spot sampling is, therefore, unlikely to be effective for operational monitoring and a recommendation on continuous sampling or sampling triggered by turbidity events is made at 5.4.9. However, risk assessment, as recommended in the previous Expert Group reports (Badenoch 1990; 1995) has become an accepted strategy for the water industry.

3.4.3 Water utilities are now expected to make an assessment of the risk of contamination by oocysts for all water abstractions, including springs and groundwaters. If they judge there is a significant or unacceptable risk of contamination, utilities are expected to make an assessment of the effectiveness of the installed treatment in removing oocysts and the risk of an oocyst being present in the water supply. Utilities should review the assessments from time to time, particularly when there is a change in circumstances potentially affecting the risk.

3.4.4 Such assessments are not intended to predict Cryptosporidium outbreaks but they can highlight where they are most likely to occur. Changes in catchments and at treatment works can have significant affects on risk and it is important that these are reported and incorporated into assessments at regular reviews. Treatment works staff should be trained to be aware of the potential effect on the final water quality of even very small changes in the catchment or the treatment stream.

3.4.5 Risk assessment for water catchments and treatment works should cover anything in the catchment area that has the potential to allow Cryptosporidium into raw water. At the treatment works it should cover any factors which do not present a barrier to Cryptosporidium, or could contribute to Cryptosporidium breaking through filters, or which do not alert staff to filter breakthrough. Areas for consideration include the following.

3.4.6 Microbiological risk assessment (MRA) is an emerging methodology to predict the risks to drinking water consumers from small numbers of pathogens breaking through into drinkingwater supplies. MRA models work by assessing the daily pathogen exposures to drinking waterconsumers and then translating those exposures into a risk by way of a dose-response curve.Limited dose-response data from human volunteer studies are available for c.parvum (DuPont et al 1995).

3.4.7 Current MRA models assume that pathogens are randomly distributed within the drinking water supply and therefore exposures to pathogens through drinking water is modelled by the Poisson distribution. There is considerable evidence however that bacteria are heterogeneously distributed in the drinking water supply and that treatment will cause clustering to occur. Intensive sampling studies using aerobic bacterial spores have shown that drinking water treatment removed 94-98% of spores but increased the clustering of those remaining. It is possible that oocysts are similarly affected. Through clustering a small proportion of consumers could be exposed to high doses through drinking water and models which do not consider rare high count of oocysts in samples could cause risk to human health to be underestimated (DETR 1998).

3.4.8 For these reasons the Group concludes that risk assessment should be based on a combination of factors including the degree of exposure of the catchment to oocysts, the treatment processes currently in place and the history of cryptosporidiosis in the community. Monitoring systems and water treatment requirements should be reviewed against the level of risk.

References

Badenoch, J. (1990) Cryptosporidium in water supplies. Report of the Group of Experts; Department of the Environment, Department of Health. London, UK. HMSO. 230pp.

Badenoch, J. (1995) Cryptosporidium in water supplies. Second Report of the Group of Experts; Department of the Environment, Department of Health. London, UK. HMSO. 108pp.

DETR (1998) Cryptosporidium: Development of a risk assessment methodology for drinking water. Report No: DWI 4519. WRc plc for Department of the Environment, Transport and the Regions, London, UK.

DuPont, H.L., Chappell, C.L., Sterling, C.R., Okhuysen, P.C., Rose, J.B., Jakubowski, W. (1995) The infectivity of Cryptosporidium parvum in healthy volunteers. N Eng J Med 332 855-859.

DWI. (1993) An outbreak of cryptosporidiosis in north Humberside, December 1989/January 1990: lessons for water supply. Drinking Water Inspectorate. London, UK.

DWI. (1997a) Assessment of water supply and associated matters in relation to the incidence of cryptosporidiosis in Torbay in 1992. Drinking Water Inspectorate. London, UK.

DWI. (1997b) Assessment of water supply and associated matters in relation to the incidence of cryptosporidiosis in Torbay in August and September 1995. Drinking Water Inspectorate. London, UK.

DWI. (1998) Assessment of water supply and associated matters in relation to the incidence of cryptosporidiosis in west Herts and north London in February and March 1997. Drinking Water Inspectorate. London, UK.

Furtado, C., Adak, G.K., Stuart, J.M., Wall, P.G., Evans, H.S. and Casemore, D.P. (1998) Outbreaks of waterborne infectious intestinal disease in England and Wales, 1992-5. Epidemiol Infect 121, 109-119

Tillet, H.E., de Louvois, J. and Wall, P.G. (1998) Surveillance of outbreaks of waterborne infectious disease: categorizing levels of evidence. Epidemiol Infect 120, 37-42.

[ Previous ] [ Contents ] [ Next ]
| A - Z Index | Business and Technical | Consumer Information | Contacting us |
| DWI Home Page | News | Product Approval | Search our site | Wales |

Drinking Water Inspectorate,
Floor 2/A1, Ashdown House, 123 Victoria Street, London, SW1E 6DE
Telephone : 020 7944 5956 - Facsimile : 020 7944 5969
E-mail: dwi_enquiries@detr.gov.uk

Updated 11 July 2001
Return to Publications and Reports Index
Return to Drinking Water Inspectorate Home Page
Department for Environment, Food & Rural Affairs (DEFRA)Home Page
The National Assembly for Wales / Cynulliad Cenedlaethol Cymru Home Page
Web site terms
© Crown copyright 2001