Volume 10, Number 6—June 2004
Sporadic Cryptosporidiosis, North Cumbria, England, 1996–2000
Risk factors for sporadic cryptosporidiosis were determined in 152 patients and 466 unmatched controls who resided in two local government districts in North Cumbria, North West England, from March 1, 1996, to February 29, 2000. Risk was associated with the usual daily volume of cold unboiled tap water drunk (odds ratio [OR] 1.40, 95% confidence intervals [CI] 1.14 to 1.71 per pint consumed per day [p = 0.001]) and short visits to farms (OR 2.02, 95% CI 1.04 to 3.90, p = 0.04). Fifty-six (84%) of 67 fecal specimens from patients obtained from January 1, 1998, and February 29, 2000, were Cryptosporidium parvum genotype 2 (animal and human strain). Livestock fecal pollution of water sources appears to be the leading cause of human sporadic cryptosporidiosis in this population and shows the need for better protection of water catchments from livestock and improved drinking water treatment in this area of England.
The protozoan parasite Cryptosporidium parvum is a leading cause of infectious diarrhea in humans and livestock with fecal-oral transmission by ingestion of oocysts (1,2). Infection is generally self-limiting, followed by variable protective immunity involving humoral and cell-mediated responses, except in the immune-suppressed, when infection may be prolonged and fatal (2,3). Cryptosporidium oocysts remain viable in water and damp soils for prolonged periods and are resistant to disinfectants at concentrations usually used in water treatment (4,5). Although sound, conventional water treatment is believed to substantially reduce the risk of viable oocysts passing into treated water, the possibility of low-level intermittent contamination has been recognized; whether such contamination affects public health is uncertain (4,6). Outbreak investigations have shown diverse modes of transmission, including contact with livestock (7,8); person-to-person transmission in households and care settings (9); consumption of contaminated foods and drinks, including milk (1,10); water from private supplies (11); and recreational water exposure (12). Infection may also be associated with travel to countries with higher incidence of cryptosporidiosis (13).
In 1992, a community outbreak of cryptosporidiosis occurred in residents of Allerdale and Copeland local government districts in North Cumbria, North West England, which compose part of the Lake District National Park; these areas have a predominantly agricultural and tourism-based economy and a population of approximately 160,000. The lakes have livestock farms and open grazing land abutting them. Approximately one third of the population received public water supplies from Ennerdale Lake, one third from Crummock Lake, and one third from a number of smaller sources. Water from Ennerdale and Crummock Lakes was disinfected with chlorine but unfiltered because the low level of particulate matter in these sources precluded chemically assisted flocculation. The smaller sources used for public supplies received a variety of conventional treatments, including coagulation, filtration, and chlorination, and chlorination alone. A small number of households had private supplies. A case-control study conducted during the outbreak showed a significant association between cryptosporidiosis and consuming cold unboiled mains tap water for persons served by water from Ennerdale Lake but no such association for those served by water from Crummock Lake and other water sources (North Cumbria Health Authority, Carlisle, 1992, unpub. data). After the outbreak, rates of laboratory-confirmed cryptosporidiosis from 1993 to 1995 were 31.2–44.2 per 100,000 in Allerdale and Copeland compared to 19.8 to 23.9 per 100,000 in the neighboring local government districts of Carlisle and Eden. Cases were not obviously clustered in time and could not be linked. A prospective case-control study was therefore undertaken to test the hypothesis that no dose-response relationship existed between consuming unboiled tap water from public water supplies and risk for sporadic Cryptosporidium infection in Allerdale and Copeland.
Medical and diagnostic microbiology services for Allerdale and Copeland residents were provided free at the point of use by the United Kingdom National Health Service and managed by North Cumbria Health Authority (14). The Authority was also responsible for maintaining the computer patient register and updating it for births, deaths, migration, and surveillance and control of infectious disease (14). The register held persons’ name, sex, date of birth, home address, and postal (zip) code but not medical information. The postal codes were geographically referenced but did not share boundaries with public water supply distribution zones, government boundaries, or census enumeration districts.
Local ethical committees approved the protocol, and a study center was set up at North Cumbria Health Authority. All fecal specimens were examined for Cryptosporidium oocysts, regardless of whether this test was requested by the clinician. Family physicians and hospital clinicians were informed of the study, but not the main hypothesis, and were reminded to ensure best practice in investigating cases of diarrhea with the assistance of the three local microbiology laboratories. Laboratory staff was requested to immediately report Cryptosporidium-positive fecal smear results to the study coordinator by telephone.
Five workers were trained to conduct the study in a standard manner by using written protocols and questionnaires. Both were pilot tested and their techniques refined before enrollment (15).
Case-patients were defined as residents of Allerdale or Copeland with 1) diarrhea (three or more loose stools in a 24-hour period), 2) onset from March 1, 1996, to February 29, 2000, and 3) a fecal smear positive for Cryptosporidium oocysts but negative for other enteric pathogens. Patients were excluded if they had, within 14 days of onset of illness, contact with another household member with cryptosporidiosis or any diarrhea illness, traveled outside the United Kingdom, stayed away from home outside the study area within the United Kingdom for >7 nights, or if they or a household member had already been enrolled as a patient or control at any time during the study.
Controls were defined as residents of Allerdale or Copeland with no history of diarrhea (three or more loose stools in a 24-hour period) in the 14 days before interview. Potential controls were excluded if they had, within 14 days of interview, traveled outside the United Kingdom, stayed away from home outside the study area but within the United Kingdom for >7 nights, or if they or a household member had already been enrolled as a patient or control at any time during the study.
Three controls were randomly selected, by using a computer algorithm, from subtables of the health authority patient register with the same age span (0–5, 6–15, and 16+ years) and the same lead characters of the postal code as the case-patient. The process was repeated when necessary to replace persons who declined, could not be contacted, or met exclusion criteria. The patient register was compared with population estimates from the 2001 census.
Interviews were conducted face-to-face at home. Participants <18 years of age were interviewed with parents or guardians, who acted as proxies for younger children, as appropriate. Informed consent was obtained and recorded. Case-patients and associated controls were enrolled as close to each other in time as possible.
For unboiled cold tap water, bottled water, and soft drinks, we asked about “usual” consumption without a time restriction. For all remaining exposures, including consuming pasteurized and unpasteurized milk, ice, and foods, exposure was sought for the 2 weeks before onset of illness for patients and before date of interview for controls.
The volume of different fluids drunk by study participants was determined by showing a standard picture card of a glass, cup, and mug; the volume was recorded according to the calibration on the card. The usual daily volume of water as cold unboiled tap water drunk at home was determined as water alone and as a diluent in cold fruit-squash type drinks. The usual daily volume of cold unboiled tap water drunk at work, school, or nursery at locations within Copeland and Allerdale District boundaries was determined the same way. The two usual daily volumes consumed at and away from home were added to give the total usual daily volume of cold unboiled tap water consumed within Copeland and Allerdale Districts. The usual daily volume consumed of bottled drinking water and soft drinks not diluted with water was determined separately without distinction between consuming it at home or away from home. Questions about consumption of ice in cold drinks were asked separately for ice made at home and ice consumed at work, school, or nursery.
The following types of contact were recorded: farms, farmed animals, and handling and feeding farm animals; slurry; household pets and feeding household pets; children’s nurseries; and recreational exposure to water in swimming pools, rivers, and streams. Information about consumption and frequency of consumption of uncooked salad items, uncooked meat, uncooked sausage meat and sausages, yogurt, cheese, and cream was also elicited.
The nature of the water supply (public or private) and sewage services (public or septic tank) to the home were recorded and corroborated against water company records. Disruption to tap water or change in tap water color or taste in the week before onset of illness for patients, and in the week before date of interview for controls, was recorded. Information on sources, water treatment works, and blending of waters to each address’s postal code was obtained from water company records and linked to individual patient and control records.
Knowledge about Cryptosporidium Infection
A series of television and newspaper articles on Cryptosporidium occurred in North Cumbria in April and May 1999. Study participants enrolled from April 28, to November 30, 1999, were also asked if they had heard of Cryptosporidium and what they knew about it.
Data Entry and Analysis
Data were double entered and differences edited and corrected in EpiInfo version 6.0.4.d. (Centers for Disease Control and Prevention, Atlanta, GA). Fluid intakes were analyzed as yes/no responses; within categories (<1/4 pint, 1/4–1 pint, >1–2 pints, and >2 pints) with χ2 tests for trend in single variable analysis; and as actual volume consumed in multivariable analysis (16).
Variables positively associated with infection at the p < 0.2 level in the single variable analysis, age, sex, and water supply zone were included in the initial multivariable model. Backward stepwise logistic regression was undertaken by comparing nested models using Likelihood Ratio Tests and GLIM software (16,17). Variables with significance p > 0.3 in iterations of the multivariable model were removed in stepwise fashion except for age, sex, and water supply zones, which were retained in all models regardless of their significance. Subsidiary analyses modeled usual daily consumption of tap water at home and usual total daily tap water consumption; cases and controls served by mixes of water from more than one source were omitted.
Microscopy of stained fecal smears for Cryptosporidium oocysts (18), Giardia, and culture for pathogenic enteric bacteria were undertaken by using standard methods at each of the three local microbiology laboratories. Smears in which Cryptosporidium oocysts were identified from January 1998 to February 2000 were also analyzed by polymerase chain reaction and restriction fragment length polymorphism typing of a region of the Cryptosporidium oocyst wall protein gene (19,20).
None of the incident cases arising from the study population during the 4 years of the study were linked or clustered in time and space. All were considered sporadic infections eligible for inclusion in the study. No changes occurred in livestock farming, livestock densities, water sources, or water treatment within the study area during the study period.
Potential Study Cases and Exclusions
Two hundred seven case-patients were ascertained during the study period; 152 (73.4%) were enrolled, and 55 (26.6%) were excluded (Table 1). One refused to participate, one was unable to complete the interview, and two did not respond. Thirty-six (17.4%) were secondary to a laboratory-confirmed case in the household, 8 (3.9%) had traveled outside the United Kingdom, and 1 (0.5%) had traveled within the United Kingdom and stayed outside the study area >7 nights. Further single case-patients were excluded for having no history of diarrhea, mixed enteric infection, and being a visitor to, or resident outside, the study area. Two additional case-patients were excluded because a member of their household had already been interviewed as a case-patient or control earlier in the study.
Potential Controls and Exclusions
Seven hundred seventy-eight potential controls were identified; 466 (59.9%) were enrolled, and 312 (40.1%) were excluded (Table 1). One hundred eighty-three (23.5%) refused to participate or were unavailable for interview. The address of three (0.4%) could not be found. Forty-six (5.9%) had a history of diarrhea, 8 (1.0%) had traveled outside the United Kingdom, and 2 (0.3%) had traveled within the United Kingdom away from the study area for >7 nights in the 2 weeks before interview. Twenty-seven (3.5%) had moved from the study area, 2 (0.3%) had an address outside the study area, and 7 (0.9%) shared a household with a patient or a control. Nineteen (2.4%) were not enrolled because 3 controls had already been recruited in association with the case, 9 (1.2%) were in the wrong age band, and 3 (0.4%) had no reason recorded.
Of the 152 study case-patients, 86 (56.6%) were <6 years of age; 47 (30.9%) were 6–15; and 19 (12.5%) were >16 years. Eighty-two (53.9%) were male. More cases were detected in the first half of each year of the study (Figure). The average annual incidence rate was similar in populations served by water from Crummock Lake, Ennerdale Lake, and the other water sources combined (Table 2).
In addition to diarrhea, the 152 patients reported abdominal pain (110 [72.4%]), vomiting (94 [61.8%]), fever (69 [45.4%]), anorexia (68 [44.7%]), and weight loss (56 [36.8%]). Thirty-seven patients were ill at interview. In 115 patients who had recovered when interviewed, the median duration of illness was 9 days (range 2–21). Nineteen (22.1%) of the 86 patients <5 years of age and 4 (8.5%) of 47 case-patients ages 6-15 years were admitted to hospital.
Three or four controls were recruited in association with 131 (86.2%) patients and one or two in association with the remainder. Patient and control groups were comparable by sex, local government district of residence, water sources and water supply zones, disruption and discoloration of tap water, nights spent away from home within the United Kingdom in the 2 weeks before onset or interview, and sewage services to the home (Table 3).
Time until Study Recruitment
One hundred twenty-eight (84.2%) patients were interviewed within 1 week and 151 (99%) within 2 weeks of the date of the Cryptosporidium-positive fecal smear test report. The delay between reporting a case and enrolling the patient and associated controls was a median of 2.3 weeks (range 1–8).
Knowledge about Cryptosporidium
Thirty six (75%) of 48 patients and 113 (67.7%) of 167 controls recruited from April 29 to November 30, 1999, stated that they had not heard of Cryptosporidium before being contacted for the study. The proportion without knowledge was similar for patients and controls recruited before and after July 8, 1999. Of the 66 persons who had previously heard of Cryptosporidium, 16 had knowledge of modes of transmission: 4 reported transmission could occur through water, drinks, or contact with farms and animals; 6 reported that transmission was by water only; and 6 reported transmission was by farm contact only.
Single Variable Analysis
Significant associations were seen with consuming cold unboiled tap water (odds ratio [OR] 2.12, 95% confidence interval [CI] 1.16 to 3.91, p = 0.012) with a significant dose-response relationship (χ2 test for trend p = 0.017). A significant dose-response relationship was also seen for the usual volume of cold unboiled tap water consumed at home (χ2 test for trend p = 0.005), but not for that consumed at the workplace, nursery, or school (χ2 test for trend p = 0.495) (Appendix). No association was found between consuming bottled water, ice, soft drinks, and pasteurized or unpasteurized milk. Consuming lettuce, tomatoes, mixed salad, and cream was associated with lower risk (p < 0.05).
Any contact with a farm was associated with a twofold increase in risk (OR 2.11, CI 1.4 to 3.2, p < 0.001). Risk was higher for short farm visits (OR 2.56, CI 1.57 to 4.17) and increased with the frequency of farm visits (χ2 test for trend p = 0.003) (Appendix). Risk was also increased by contact with farm animals (OR 2.23, CI 1.45 to 3.43), eating food within 2 hours of contact with farm animals (OR 3.11, CI 1.79 to 5.38), and stroking farm animals (OR 2.01, CI 1.17 to 3.42). Walking near slurry applied to fields was not associated with increased risk, but contact with slurry showed some evidence of increased risk (OR 2.0, CI 0.99 to 4.02).
Contact with pets at home or contact with pets with diarrhea did not increase risk. Risk was increased for feeding pets leftovers (OR 3.79, CI 1.0 to 14.69), with marginal evidence of risk for feeding pets raw vegetables (OR 2.09, CI 0.95 to 4.56) and biscuits (OR 1.76, CI 0.95 to 3.21). Contact with animals other than farm animals and home pets was not associated with infection (OR 0.84, CI 0.54 to 1.30). Risk was increased by having accidentally touched feces from any animal (OR 3.04, CI 1.33 to 6.94). Attendance at a playgroup or nursery and recreational exposure to water was not associated with infection (Appendix).
The usual volume of cold unboiled tap water consumed was independently associated with cryptosporidiosis (OR 1.40, CI 1.14 to 1.71 per pint consumed per day) in the final multivariable model (Table 4). A short visit to farms (OR 2.02, CI 1.04 to 3.9) was also significant. No difference in risk was found between the different water supply zones, irrespective of whether the zones received unfiltered water from Crummock and Ennerdale Lakes, other public supplies with a variety of conventionally filtered and unfiltered water, or private water supplies (Tables 3 and 4). Slight evidence was found for increased risk for feeding pets raw vegetables (OR 2.11, CI 0.98 to 4.56) and biscuits (OR 1.77, CI 0.94 to 3.35) but not for age, sex, consuming nonlocally produced cheese, contact with farms without cattle or sheep, or with cattle (Table 4). Subsidiary analysis showed that usual volume of cold unboiled tap water consumed at home was also a significant risk factor. A further analysis that excluded persons whose house was served by a mixed public water supply found similar results (not shown).
Genotyping was undertaken in 67 of 101 cases from 1998 to 2000. All the smears were confirmed positive, and 56 (83.6%) of tested smears were C. parvum genotype 2 (animal and human) strain, 1 (1.5%) was genotype 1 (human strain), and 10 (14.9%) could not be typed.
The primary care patient register was reviewed after patients were recruited for the study. The register contained 166,376 names of Allerdale and Copeland residents compared to a population of 162,809 enumerated at the 2001 census (available from www.statistics.gov.uk/census2001). The computer algorithm used to randomly select potential controls generated 125 tables of registered patients’ names, where registered patients were within the same age category as study case-patients and had the same lead characters of the postal code of residence as study case-patients. The tables contained a median of 496 (range 9–7,800) names.
Drinking cold unboiled tap water from public drinking water supplies was a highly significant risk factor for sporadic human cryptosporidiosis, regardless of the water source. To our knowledge, this study is the first to show that drinking from public water supplies is an important risk factor for sporadic human Cryptosporidium infection. Most cases were in children, consistent with previous reports from England and Canada (21,22). Many patients required admission to hospital, showing the seriousness of illness. Infection in study patients was also associated with short visits to farms and predominantly with the C. parvum genotype 2 (animal and human) strain, consistent with farmed livestock’s being a major source of infection. Risk was not increased by contact with pets.
Excluding study participants by recent travel ensured that environmental exposures most likely occurred within the study area. Excluding household contact with an earlier onset case ensured that person-to-person transmission within the household was unlikely to have occurred in study case-patients.
Ascertaining patients who sought medical attention was likely to have been high because the National Health Service provided free medical care within the study area at the point of use and because all fecal specimens were tested for Cryptosporidium by National Health Service laboratories without charge. Although only half of patients with cryptosporidiosis may seek medical attention in the United Kingdom (23), we have no reason to suppose that risk factors for patients who do not visit healthcare facilities would differ substantially from those that did.
Matching refers to pairing cases with one or more controls on the basis of their similarity in selected variables, with the objective of eliminating bias (24,25). We undertook stratified random sampling from a population list to select potential controls and adjust by using multiple regression analysis, which is one of a number of alternative designs to matching (24,25).
Refusal to participate was low at 4 (1.9%) of 205 cases and 183 (23.5%) of 778 potential controls. Lower response rates in controls compared to cases is expected (24). Care was taken during the design and conduct of the study to mask interviewers to the tap water hypothesis. Interviewer training emphasized that all risk factors were plausible and required equal care in measurement. A survey after media coverage indicated little knowledge about risk factors for cryptosporidiosis by patients or controls. Patients may have increased fluid consumption after the onset of illness. Study participants were therefore asked to report their “usual consumption” of unboiled tap water, bottled water, and soft drinks, without a time restriction. Although a bias towards a patient’s recalling consumption of fluids after onset of illness could explain some of the association with tap water that we observed, we do not think it can explain it entirely. In particular, fluid volumes were measured in the same way for unboiled tap water, bottled water, and soft drinks, but the highly significant association and dose-response relationship were observed only for unboiled tap water. An interview date bias was avoided because patients and associated controls were enrolled within a short time of each other.
Cryptosporidiosis in HIV-infected persons is associated with consuming unboiled tap water. We do not believe that undetected HIV infection or other causes of immune suppression could have been a major confounding factor. Only two new cases of HIV infection would be expected in Allerdale and Copeland each year even if rates for the whole North West of England were applied to the study population. However, our findings reinforce the need for immune-compromised persons, including those with HIV infection, to avoid drinking unboiled tap water (26,27). Allerdale and Copeland had very similar levels of social deprivation (28). Moreover, controls resided in approximately the same locality as patients, as defined by shared lead characters of the postal code. Therefore, a systematic difference in social deprivation between patients and controls was unlikely and would not have explained the associations we observed.
The spring peak and smaller autumn peak in our cases is consistent with previous reports from England and Wales. These peaks are attributed to lambing, calving, and runoff from spring rains and to summer travel to countries with higher incidence of cryptosporidiosis (4,13).
Our findings contrast with a case-control study in Adelaide and Melbourne, which did not detect increased risk for sporadic cryptosporidiosis associated with the public water supplies (29). This difference may reflect the quality of the source waters, of water treatment, or both in these cities. Swimming pool exposure was the most significant risk factor in the Australian study. Regularly consuming raw vegetables was protective in that study. We also observed a protective association with lettuce, tomatoes, and mixed salad, and additionally for cream, in single variable analysis. These foods might have conferred a direct protective nutrient effect, been markers for more favorable general nutrition, or contained small numbers of oocysts derived from water used for irrigation and preparation sufficient to induce immune boosting (3,30,31). A recent study of sporadic cryptosporidiosis in the San Francisco Bay Area also failed to show an association with tap water, but the study was small (32).
Most of our patients were children, which suggests that older members of our study population were mainly immune, probably because of long-term immune boosting from low-level intermittent contamination of water supplies and contact with livestock (31,33). This observation is consistent with recent seroprevalence studies in blood donors resident in midwestern American cities and indicates lower seroprevalence of Cryptosporidium antibodies in populations served by deep borehole water compared to lake and river water supplies (34,35).
No association was seen for contact with pets at home, pets with diarrhea, or feeding tinned pet meat, raw meat, or pellets. However, feeding pets biscuits and raw vegetables was associated with slightly increased risk in single variable analysis and in the final multivariable model. These food types may be markers for more intimate contact with animal secretions; feeding raw vegetables may indicate contact with contaminated water in preparation. Although contamination of pet biscuits is possible, these products are dry and manufactured at high temperature; thus, survival of oocysts within these foods seems unlikely. Accidental hand contact with the feces of any type of animal was significant in single variable analysis but not in the final multivariable model. These results suggest that sound hygiene in cleaning animal feeding utensils, avoiding cross-contamination between pet and human food preparation areas, and good hand hygiene are desirable but that pets and pet feeding were not major risk factors for cryptosporidiosis in this population.
The findings of this study are consistent with the decline in human Cryptosporidium infection throughout England and Wales, coincident with the foot and mouth disease epidemic in livestock during 2001 (13). These two facts strongly suggest that livestock reservoirs of Cryptosporidium contribute substantially to sporadic human cryptosporidiosis in North Cumbria and in England and Wales as a whole, through low-level intermittent Cryptosporidium oocyst contamination of public drinking water supplies. Our results support the need for rigorous risk assessment of water sources and, where indicated, improved catchment control. Our results are also in accord with recent U.K. legislation that requires continuous monitoring of Cryptosporidium oocyst concentrations in treated water from at-risk supplies (36). Advanced methods of filtration, disinfection, and UV light treatment may be required to further decrease the risk for cryptosporidiosis from public water supplies (5,37,38).
The water company installed membrane filtration during 2000 at works served by Crummock and Ennerdale Lakes, which previously provided chlorination alone. The impact of this intervention will be presented in a separate article.
Dr. Goh is an honorary consultant to the Carlisle and District Primary Care Trust and was formerly the consultant in Communicable Disease Control to North Cumbria Health Authority, in North West England. She specialized in public health medicine and developed an interest in childhood immunization, enteric infection, Cryptosporidium, and water supplies.
We thank colleagues in the Steering Group and in the many local and national agencies that contributed to this research, in particular, Brian White, Vic Emmerson, Mary Cosgrove, Peter Daley, John Cain, Rachel Horton, Susan Partridge, Judith Hilton, Patrick Wall, Ros Stanwell-Smith, James Stuart, Claire Gilham, Alan Godfree, Matthew Wilkinson, Charmian Kerr, David Counter, Andrew Holliman, John Gray, Tony Lloyd, Mark Smith, David Holt, Joy Graham, Jennifer Clay, Paul Blaylock, Richard Lamb, Dennis Massey, Emma Wigginton, Anthony Pennington, Kristin Elwin, and Anne Thomas.
This study was funded by the Department of Environment, Food and Rural Affairs, Department of Health, United Kingdom Water Industry Research Limited, and was supervised by the Drinking Water Inspectorate.
- Fayer R, Morgan U, Upton SJ. Epidemiology of Cryptosporidium: transmission, detection, and identification. Int J Parasitol. 2000;30:1305–22.
- Guerrant RL. Cryptosporidiosis: an emerging highly infectious threat. Emerg Infect Dis. 1997;3:51–7.
- Kosek M, Alcantara A, Lima AAM, Guerrant RL. Cryptosporidiosis: an update. Lancet Infect Dis. 2001;1:262–9.
- Meinhardt PL, Casemore DP, Miller KB. Epidemiological aspects of human cryptosporidiosis and the role of waterborne transmission. Epidemiol Rev. 1996;18:118–36.
- Korich DG, Mead JR, Madore MS, Sinclair NA, Sterling CR. Effects of ozone, chlorine dioxide, chlorine, and monochloramine on Cryptosporidium parvum oocyst viability. Appl Environ Microbiol. 1990;56:1423–8.
- Le Chevalier MW, Norton WD, Lee RG. Giardia and Cryptosporidium spp in filtered drinking water supplies. Appl Environ Microbiol. 1991;57:2617–21.
- Miron D, Kenes JDK. Calves as a source of an outbreak among young children in an agricultural closed community. Pediatr Infect Dis J. 1991;10:438–41.
- Sayers GM, Dillon MC, Connolly E. Cryptosporidiosis in children who visited an open farm. Commun Dis Rep CDR Rev. 1996;6:R140–4.
- Cordell RL, Addiss DG. Cryptosporidiosis in child care settings: a review of the literature and recommendations for prevention and control. Pediatr Infect Dis J. 1994;13:310–7.
- Gelletli R, Stuart J, Soltano N, Armstrong R, Nicols G. Cryptosporidiosis associated with school milk. Lancet. 1997;350:1005–6.
- Furtado C, Adak GK, Stuart JM, Wall PG, Evans HS, Casemore DP. Outbreaks of waterborne infectious intestinal disease in England and Wales, 1992–5. Epidemiol Infect. 1998;121:109–19.
- Kramer MH, Sorhage FE, Goldstein ST, Dalley E, Wahlquist SP, Herwaldt BL. First reported outbreak in the United States of cryptosporidiosis associated with a recreational lake. Clin Infect Dis. 1998;26:27–33.
- Smerdon WJ, Nichols T, Chalmers RM, Heine H, Reacher M. Foot and mouth disease in livestock and reduced cryptosporidiosis in humans, England and Wales. Emerg Infect Dis. 2003;9:22–8.
- Ham C. Health policy in Britain. 4th ed. New York: Palgrave; 1999. p. 4–26, 27–50.
- Armstrong BK, White E, Saracci R. Principles of exposure measurement in epidemiology. Oxford: Oxford University Press; 1992. p. 1–45, 115–36, 294–317.
- Armitage P, Berry G. Statistical methods in medical research. 3rd ed. Oxford: Blackwell; 1998. p. 93–153.
- Francis B, Green M, Payne C, eds. The GLIM system release 4 manual. Oxford: Clarendon Press; 1993.
- Casemore DP. Laboratory methods for diagnosing cryptosporidiosis. J Clin Pathol. 1991;44:445–51.
- Spano F, Putigagni L, McLauchlin J, Casemore DP, Crisanti A. PCR-RFLP analysis of the Cryptosporidium oocyst wall protein (COWP) gene discriminates between C. wrairi and C. parvum, and between C. parvum isolates of human and animal origin. FEMS Microbiol Lett. 1997;150:209–17.
- Peng MM, Xiao L, Freeman AR, Arrowood MJ, Escalante AA, Weltman AC, Genetic polymorphism among Cryptosporidium parvum isolates: evidence of two distinct human transmission cycles. Emerg Infect Dis. 1997;3:567–73.
- Public Health Laboratory Service Study Group. Cryptosporidiosis in England and Wales: prevalence and clinical and epidemiological features. BMJ. 1990;300:774–7.
- Majowiecz SE, Michel P, Aramini JJ, McEwen SA, Wilson JB. Descriptive analysis of endemic cryptosporidiosis cases reported in Ontario, 1996–1997. Can J Public Health. 2001;92:62–6.
- Wheeler JG, Sethi D, Cowden JM, Wall PG, Rodrigues LC, Tompkins DS, Study of infectious intestinal disease in England: rates in the community, presenting to general practice, and reported to national surveillance. The Infectious Intestinal Disease Study Executive. BMJ. 1999;318:1046–50.
- Schlesselman JJ, Stolley PD, Schlesselman JJ, eds. Case-control studies: design, conduct and analysis. New York: Oxford University Press; 1982. p. 124–43.
- Hennekens CH, Buring JE. Epidemiology in medicine. 1st ed. Boston: Little Brown; 1987. p. 287–324.
- Chief Medical Officer for England. Cryptosporidium in water: advice to the immunocompromised. London: Department of Health; 1999; CMO’s Update 21 February 1999. Available from: http://www.doh.gov.uk/pub/docs/doh/cmo21.pdf
- Centers for Disease Control and Prevention. Preventing cryptosporidiosis: a guide for people with compromised immune systems. [accessed 24 Mar 2004]. Available from: http://www.cdc.gov/ncidod/dpd/parasites/cryptosporidiosis/factsht_crypto_prevent_ci.htm
- 1998 index of local deprivation. U.K. Government; 2003 [accessed 24 Mar 2004]. Available from: http://www.odpm.gov.uk/stellent/groups/odpm_urbanpolicy/
- Robertson B, Sinclair MI, Forbes AB, Veitch M, Kirk M, Cunliffe D, Case-control studies of sporadic cryptosporidiosis in Melbourne and Adelaide, Australia. Epidemiol Infect. 2002;128:419–31.
- Calder PC, Kew S. The immune system: a target for functional foods? Br J Nutr. 2002;88(Suppl):S165–77.
- Casemore DP. Epidemiological aspects of human cryptosporidiosis. Epidemiol Infect. 1990;104:1–28.
- Khalakdina A, Vugia DJ, Nadle J, Rothrock GA, Colford JM. Is drinking water a risk factor for endemic cryptosporidiosis? A case-control study in the immunocompetent general population of the San Francisco Bay area. BMC Public Health. 2003;3:11 Available from http://www.biomedcentral.com/1471-2458/3/11 accessed 25 Mar 2004.
- Okhuysen PC, Chappell CL, Sterling CR, Jakubowski W, DoPont HL. Susceptibility and serologic response to healthy adults to reinfection with Cryptosporidium parvum. Infect Immun. 1998;66:441–3.
- Frost FJ, Muller T, Craum GF, Calderon RL, Roefer PA. Paired city Cryptosporidium serosurvey in the southwest USA. Epidemiol Infect. 2001;126:301–7.
- Frost FJ, Muller T, Craun G, Lockwood WB, Calderon RL. Serological evidence of endemic waterborne Cryptosporidium infections. Ann Epidemiol. 2002;12:222–7.
- The Stationery Office. The Water Supply (water quality) (Amendment) Regulations 1999. Statutory Instrument 1999 No. 1524. [accessed 24 Mar 2004]. Available from: http://www.dwi.gov.uk/regs/si1524/index.htm
- Hsu BM, Yeh HH. Removal of Giardia and Cryptosporidium in drinking water treatment: a pilot study. Water Res. 2003;37:1111–7.
- Morita S, Namikoshi A, Hirata T, Oguma K, Katayama H, Ohgaki S, Efficacy of UV irradiation in inactivating Cryptosporidium parvum oocysts. Appl Environ Microbiol. 2002;68:5387–93.