ML12221A280

Centers for Disease Control and Prevention - Morbidity and Mortality Weekly Report - Surveillance for Waterborne-Disease Outbreaks ?United States, 1999-2000

ML12221A280
Person / Time
Site:	Watts Bar
Issue date:	11/22/2002
From:	US Dept of Health & Human Services, Centers for Disease Control
To:	Justin Poole Watts Bar Special Projects Branch
Poole J
References
Download: ML12221A280 (52)

Text

Morbidity and Mortality Weekly Report Surveillance Summaries November 22, 2002 / Vol. 51 / No. SS-8 Surveillance for Waterborne-Disease Outbreaks United States, 1999-2000 Centers for Disease Control and Prevention TM SAFER

• HEALTHIER HEALTHIER

• PEOPLE

MMWR CONTENTS The MMWR series of publications is published by the Introduction ......................................................................... 2 Epidemiology Program Office, Centers for Disease Control and Prevention (CDC), U.S. Department of Background ......................................................................... 2 Health and Human Services, Atlanta, GA 30333. EPA Regulations ................................................................ 2 Methods .............................................................................. 6 SUGGESTED CITATION Data Sources .................................................................... 6 General: Centers for Disease Control and Prevention. Definitions ........................................................................ 6 Surveillance Summaries, November 22, 2002.

MMWR 2002:51(No. SS-8). Outbreak Classification ..................................................... 7 Specific: [Author(s)]. [Title of particular article]. In: Results ................................................................................. 7 Surveillance Summaries, November 22, 2002. Outbreaks Associated with Drinking Water ........................ 7 MMWR 2002;51(No. SS-8):[inclusive page numbers]. Outbreaks Associated with Recreational Water ................ 12 Outbreaks Associated with Occupational Exposure to Water ....................................................... 17 Centers for Disease Control and Prevention Previously Unreported Outbreaks .................................... 17 Julie L. Gerberding, M.D., M.P.H. Outbreaks Not Classified as WBDOs ............................... 18 Director Discussion ......................................................................... 18 David W. Fleming, M.D.

Considerations Regarding Reported Results ..................... 18 Deputy Director for Science and Public Health Outbreaks Associated with Drinking Water ...................... 20 Dixie E. Snider, Jr., M.D., M.P.H.

Associate Director for Science Outbreaks Associated with Recreational Water ................ 23 Outbreaks Associated with Occupational Epidemiology Program Office Exposures to Water ...................................................... 26 Stephen B. Thacker, M.D., M.Sc.

Director Conclusion ........................................................................ 26 Acknowledgments ............................................................. 27 Division of Public Health Surveillance and Informatics References ......................................................................... 27 Daniel M. Sosin, M.D., M.P.H. Appendix A ........................................................................ 29 Director Appendix B ........................................................................ 37 Associate Editor, Surveillance Summaries Glossary ............................................................................ 45 Office of Scientific and Health Communications John W. Ward, M.D.

Director Editor, MMWR Series Suzanne M. Hewitt, M.P.A.

Managing Editor C. Kay Smith-Akin, M.Ed.

Project Editor Lynda G. Cupell Malbea A. Heilman Beverly J. Holland Visual Information Specialists Quang M. Doan Erica R. Shaver Information Technology Specialists

Vol. 51 / SS-8 Surveillance Summaries 1 Surveillance for Waterborne-Disease Outbreaks United States, 1999-2000 Sherline H. Lee, M.P.H.1 Deborah A. Levy, Ph.D.1 Gunther F. Craun, M.P.H.2 Michael J. Beach, Ph.D.1 Rebecca L. Calderon, Ph.D.3 1Division of Parasitic Diseases National Center for Infectious Diseases, CDC 2Gunther F. Craun and Associates Staunton, Virginia 3U.S. Environmental Protection Agency Research Triangle Park, North Carolina Abstract Problem/Condition: Since 1971, CDC, the U.S. Environmental Protection Agency (EPA), and the Council of State and Territorial Epidemiologists (CSTE) have maintained a collaborative surveillance system for the occurrences and causes of waterborne-disease outbreaks (WBDOs).This surveillance system is the primary source of data concerning the scope and effects of waterborne diseases on persons in the United States.

Reporting Period Covered: This summary includes data regarding outbreaks occurring during January 1999-December 2000 and previously unreported outbreaks occurring in 1995 and 1997.

Description of the System: The surveillance system includes data for outbreaks associated with drinking water and recreational water. State, territorial, and local public health departments are primarily responsible for detecting and investigating WBDOs and voluntarily reporting them to CDC on a standard form. The unit of analysis for the WBDO surveillance system is an outbreak, not an individual case of a waterborne disease. Two criteria must be met for an event to be defined as a WBDO. First, >2 persons must have experienced a similar illness after either ingestion of drinking water or exposure to water encountered in recreational or occupational settings. This criterion is waived for single cases of laboratory-confirmed primary amebic meningoencephalitis and for single cases of chemical poisoning if water-quality data indicate contamination by the chemical. Second, epidemiologic evidence must implicate water as the prob-able source of the illness.

Results: During 1999-2000, a total of 39 outbreaks associated with drinking water was reported by 25 states. Included among these 39 outbreaks was one outbreak that spanned 10 states. These 39 outbreaks caused illness among an estimated 2,068 persons and were linked to two deaths. The microbe or chemical that caused the outbreak was identi-fied for 22 (56.4%) of the 39 outbreaks; 20 of the 22 identified outbreaks were associated with pathogens, and two were associated with chemical poisoning. Of the 17 outbreaks involving acute gastroenteritis of unknown etiology, one was a suspected chemical poisoning, and the remaining 16 were suspected as having an infectious cause. Twenty-eight (71.8%) of 39 outbreaks were linked to groundwater sources; 18 (64.3%) of these 28 groundwater outbreaks were associated with private or noncommunity wells that were not regulated by EPA. Fifty-nine outbreaks from 23 states were attributed to recreational water exposure and affected an estimated 2,093 persons. Thirty-six (61.0%) of the 59 were outbreaks involving gastroenteritis. The etiologic agent was identified in 30 (83.3%) of 36 outbreaks involving gastroenteritis. Twenty-two (61.1%) of 36 gastroenteritis-related outbreaks were associated with pools or interactive fountains. Four (6.8%) of the 59 recreational water outbreaks were attributed to single cases of primary amebic menin-goencephalitis (PAM) caused by Naegleria fowleri. All four cases were fatal. Fifteen (25.4%) of the 59 outbreaks were associated with dermatitis; 12 (80.0%) of 15 were associated with hot tubs or pools. In addition, recreational water outbreaks of leptospirosis, Pontiac fever, and chemical keratitis, as well as two outbreaks of leptospirosis and Pontiac fever associated with occupational exposure were also reported to CDC.

Interpretation: The proportion of drinking water outbreaks associated with surface water increased from 11.8%

during 1997-1998 to 17.9% in 1999-2000. The proportion of outbreaks (28) associated with groundwater sources increased 87% from the previous reporting period (15 outbreaks), and these outbreaks were primarily associated (60.7%)

2 MMWR November 22, 2002 with consumption of untreated groundwater. Recreational water outbreaks involving gastroenteritis doubled (36 out-breaks) from the number of outbreaks reported in the previous reporting period (18 outbreaks). These outbreaks were most frequently associated with Cryptosporidium parvum (68.2%) in treated water venues (e.g., swimming pools or interactive fountains) and by Escherichia coli O157:H7 (21.4%) in freshwater venues. The increase in the number of outbreaks probably reflects improved surveillance and reporting at the local and state level as well as a true increase in the number of WBDOs.

Public Health Action: CDC and others have used surveillance data to identify the types of water systems, their deficiencies, and the etiologic agents associated with outbreaks and evaluated current technologies for providing safe drinking water and safe recreational water. Surveillance data are used also to establish research priorities, which can lead to improved water-quality regulations. Only the groundwater systems under the influence of surface water are required to disinfect their water supplies, but EPA is developing a groundwater rule that specifies when corrective action (includ-ing disinfection) is required. CDC and EPA are conducting epidemiologic studies to assess the level of waterborne illness attributable to municipal drinking water in nonoutbreak conditions. Rules under development by EPA the Ground Water Rule (GWR), the Long Term 2 Enhanced Surface Water Treatment Rule (LT2ESWTR), and Stage 2 Disinfection Byproduct Rules (DBPR) are expected to further protect the public from contaminants and disinfec-tion byproducts in drinking water. Efforts by EPA under the Beaches Environmental Assessment, Closure, and Health (BEACH) program are aimed at reducing the risks for infection attributed to ambient recreational water by strengthen-ing beach standards and testing; providing faster laboratory test methods; predicting pollution; investing in health and methods research; and improving public access to information regarding both the quality of the water at beaches and information concerning health risks associated with swimming in polluted water. EPAs Beach Watch (available at http://www.epa.gov/waterscience/beaches) provides online information regarding water quality at U.S. beaches, local protection programs, and other beach-related programs. CDC partnered with a consortium of local and national pool associations to develop a series of health communication materials for the general public who attend treated recre-ational water venues and to staff who work at those venues. CDC has also developed a recreational water outbreak investigation toolkit that can be used by public health professionals. All of the CDC materials are accessible at the CDC Healthy Swimming website (http://www.cdc.gov/healthyswimming).

Introduction information can influence research priorities, lead to improved water-quality regulations, and illustrate the burden of illness During 1920-1970, statistical data regarding U.S. waterborne-attributed to water.

disease outbreaks (WBDOs) were collected by multiple researchers and federal agencies (1). Since 1971, CDC, the U.S. Environmental Protection Agency (EPA), and the Coun- Background cil of State and Territorial Epidemiologists (CSTE) have main-tained a collaborative surveillance system that tracks the EPA Regulations occurrences and causes of WBDOs (2-6). This surveillance Drinking Water system includes data regarding outbreaks associated with Public water systems are regulated under the Safe Drinking drinking water, recreational water, and other types of water Water Act (SDWA) of 1974 and its subsequent 1986 and exposures. This report includes data for 1999 and 2000 and 1996 amendments (7-9) (Box). Under SDWA, EPA is for previously unreported drinking water-associated outbreaks authorized to set national standards to protect drinking water that occurred in 1995 and 1997. and its sources against naturally occurring or man-made con-CDCs and EPAs surveillance activities are intended to taminants. The 1996 SDWA amendments require EPA to

1) characterize the epidemiology of WBDOs; 2) identify the publish a list every 5 years of contaminants that are known or etiologic agents that caused WBDOs and determine why the anticipated to occur in public water systems and that might outbreaks occurred; 3) train public health personnel to detect need to be regulated. The first list was called the drinking and investigate WBDOs; and 4) collaborate with local, state, water Contaminant Candidate List (CCL). CCL contained federal, and international agencies on initiatives to prevent 60 contaminants/contaminant groups, included 10 pathogens, waterborne disease. Data obtained through this surveillance and was published in the Federal Register on March 2, 1998 system are useful for identifying major deficiencies in provid- (10). A decision concerning whether to regulate >5 contami-ing safe drinking water and recreational water. Surveillance nants from CCL was required by August 2001. Microbial

Vol. 51 / SS-8 Surveillance Summaries 3 BOX. Environmental Protection Agency (EPA) regulations regarding drinking water, 1974-2003 Regulation/date Description Safe Drinking Water Act/1974 and Authorizes EPA to set national standards to protect drinking water and its sources 1986 and 1996 amendments Total Coliform Rule (TCR)/and Requires routine monitoring for total coliforms of all public water systems plus periodic Maximum Contaminant Level on-site inspections for systems that take <5 samples/month to evaluate and document treat-(MCL)/1989 ment, storage, distribution network, operation and maintenance, and overall management.

Systems that collect >40 samples/month (i.e., typically, systems that serve >33,000 persons) violate MCL if >5.0% of the samples collected during each month are positive for total coliforms; systems that collect <40 samples/month violate MCL if two samples during the month are positive for total coliforms. If a system has a total coliform-positive sample, then 1) that sample must be tested for the presence of fecal coliforms or Escherichia coli, and 2) three repeat samples must be collected (four, if the system collects <1 routine sample/month) within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> and analyzed for total coliforms. If positive, the sample must be analyzed for fecal coliforms or Es. coli. In addition, >5 routine samples must be collected during the next month of sampling, regardless of system size. For any size system, if two consecutive total coliform-positive samples occur at one site during a month, and one of these samples is also fecal coliform-positive or Es. coli-positive, the system has an acute violation of the Maximum Contaminant Level and must notify the state and the public immediately.

Surface Water Treatment Rule Covers all water systems that use surface water or groundwater under the direct influence of (SWTR)/1989 surface water; all systems must disinfect their water, and the majority of systems must filter their water also, unless they meet EPA-specified filter-avoidance criteria that define high-quality source water. Specific requirements include

• a combined filter-effluent-performance standard for turbidity (i.e., for rapid granular filters, 0.5 nephelometric turbidity unit [NTU] maximum for 95% of measurements [taken every 4 hour4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />s] during a month) and no single NTU reading >5.0;

• watershed protection, redundant disinfection capability, and other requirements for unfiltered systems;

• a 0.2-mg/L disinfectant residual entering the distribution system; and

• maintenance of a detectable disinfectant residual in all parts of the distribution system.

This rule requires that all such systems reduce the level of Giardia by 99.9% (3-log reduction) and viruses by 99.99% (4-log reduction) through a combination of removal (filtration) and inactivation (disinfection).

Information Collection Rule/ Requires systems serving >100,000 persons to provide treatment data and monitor disinfec-1996-1998 tion byproducts and source water quality parameters. Surface water systems are also required to monitor Cryptosporidium, Giardia, total culturable viruses, and total and fecal coliforms or Es. coli >1 time/month for 18 months. Results provided information to facilitate development of the Long Term 2 Enhanced SWTR, which is intended to protect against microbial risks by targeting those systems with suboptimal quality source water and to balance the health risks associated with disinfection byproducts and the anticipated Stage 2 Disinfection Byproduct Rule.

Interim Enhanced Surface Water Follow-up to SWTR that covers all public systems using surface water or groundwater under Treatment Rule (IESWTR)/1998 the direct influence of surface water and serving >10,000 persons. Key provisions include

• a 2-log Cryptosporidium-removal requirement for filtered systems;

• strengthened combined filter-effluent-turbidity performance standards for systems using conventional filtration treatment or direct filtration (0.3 NTU maximum for 95% of measurements during a month and no single NTU reading >1.0);

• individual filter turbidity monitoring provisions;

4 MMWR November 22, 2002 BOX (Continued). Environmental Protection Agency (EPA) regulations regarding drinking water, 1974-2003 Regulation/date Description

• disinfection profile and benchmark provisions to ensure continued levels of microbial protec-tion while facilities take necessary steps to comply with new disinfection byproduct standards;

• revision of the definition of groundwater under the influence of surface water and the watershed-control requirements for unfiltered public water systems to include detection of Cryptosporidium;

• requirements for covers on newly finished water reservoirs;

• sanitary surveys for all surface water systems regardless of size; and

• an MCL goal of zero oocysts for Cryptosporidium.

Lead and Copper Rule/2000 Streamlines requirements, promotes consistent national implementation, and reduces the changes burden for water systems.

Long Term 1 Enhanced SWTR Companion regulations for IESWTR; LT1ESWTR applies to public water systems that use (LT1ESWTR)/2002 and the Filter surface water or groundwater under the direct influence of surface water and that serve <10,000 Backwash Recycling Rule (FBRR)/ persons. FBRR regulates how treatment plants recycle water that has been used to backwash a 2001 filter or that has been extracted from treatment plant sludge. FBRR regulates the point in the treatment plant at which the contaminated recycle water may be introduced, assuring that the water is subject to the entire particle and Cryptosporidium parvum removal process.

Long Term 2 Enhanced SWTR Applies to all systems using surface water or groundwater under the influence of surface water; (LT2ESWTR)/expected in 2003 will provide additional protection against Cryptosporidium. Systems will be assigned to a treat-ment category on the basis of their source-water Cryptosporidium levels; the category then determines how much additional treatment is required.

Stage 2 Disinfection Byproduct Will apply to community water systems and nontransient noncommunity water systems that Rule (DBPR)/expected in 2003 use an alternative to ultraviolet disinfection or deliver disinfected water; systems will be required to monitor for total trihalomethanes and the sum of five haloacetic acids and comply with MCLs at each monitoring location as a locational running annual average.

Ground Water Rule (GWR) (1996 Applies to public groundwater systems (i.e., systems that have >15 service connections, or amendment to EPAs Safe Drink- regularly serve >25 persons daily for >60 days/year) or any system that mixes surface and ing Water Act)/expected to be groundwater if the groundwater is added directly to the distribution system and provided to finalized in 2003 consumers without treatment. Establishes multiple barriers to protect against bacteria and viruses in drinking water from groundwater sources; establishes targeted strategy to identify groundwater systems at high risk for fecal contamination. Key areas include

• system sanitary surveys;

• hydrogeologic sensitivity assessments for nondisinfected systems;

• source-water microbial monitoring by systems that do not disinfect and that draw from hydrogeologically sensitive aquifers or have detected fecal indicators within the systems distribution system;

• corrective action by any system with substantial deficiencies or positive microbial samples indicating fecal contamination; and

• compliance monitoring for systems that disinfect to ensure that they reliably achieve 4-log (99.99%) inactivation or removal of viruses.

GWR does not apply to privately owned wells that serve <25 persons (e.g., individual home-owner wells).

Vol. 51 / SS-8 Surveillance Summaries 5 contamination is regulated under the Total Coliform Rule monitoring ended in December 1998, and data were ana-(TCR) of 1989 and the Surface Water Treatment Rule lyzed.

(SWTR) of 1989 (11-13). SWTR covers all water systems EPA also made minor changes in 2000 to the Lead and that use surface water or groundwater under the direct influ- Copper Rule to streamline requirements, promote consistent ence of surface water (Glossary). SWTR is intended to pro- national implementation, and in certain cases, reduce the tect against exposure to Giardia intestinalis, viruses, and burden for water systems. The action levels of 0.015 mg/L for Legionella, as well as selected other pathogens. In 1998, EPA lead and 1.3 mg/L for copper remain the same (21).

promulgated the Interim Enhanced Surface Water Treatment Recreational Water Rule (IESWTR) (14), which provides additional protection against Cryptosporidium and other waterborne pathogens for Regulation of recreational water is determined by state and systems that serve >10,000 persons. In 2002, EPA finalized local governments. Standards for operating, disinfecting, and the Long Term 1 Enhanced SWTR (LT1ESWTR) for public filtering public swimming and wading pools are regulated by water systems that use surface water or groundwater under state and local health departments and, as a result, are varied.

the direct influence of surface water and serve <10,000 per- In 1986, EPA established a guideline for microbiological sons (15). LT1ESWTR was proposed in combination with water quality for recreational freshwater (e.g., lakes and ponds) the Filter Backwash Recycling Rule (FBRR), which was final- and marine water (22). The guideline recommends that the ized in 2001 (16,17). monthly geometric mean concentration of organisms in fresh-The 1996 Amendments require EPA to develop regulations water should be <33/100 mL for enterococci or <126/100 that require disinfection of groundwater systems as necessary mL for Es. coli. States have latitude regarding their guidelines to protect the public health; EPA has proposed the Ground or regulations and can post warning signs to alert potential Water Rule (GWR) to meet this mandate (18). GWR speci- bathers until water quality improves. Unlike treated venues fies the appropriate use of disinfection in groundwater and where disinfection can be used to address problems with addresses other components of groundwater systems to microbiological quality of the water, contaminated freshwa-ensure public health protection. GWR applies to public ter can require weeks or months to improve or return to nor-groundwater systems (systems that have >15 service connec- mal. Prompt identification of potential sources of tions or regularly serve >25 persons/day for >60 days/year). contamination and remedial action is necessary to return bath-This rule also applies to any system that mixes surface and ing water to an appropriate quality for recreational use (23).

groundwater if the groundwater is added directly to the dis- EPAs Action Plan for Beaches and Recreational Waters tribution system and provided to consumers without treat- (Beach Watch) was developed as part of the Clean Water ment. GWR does not apply to privately owned wells. Action Plan. The intent of Beach Watch is to assist state, Additional protection of groundwater from both chemical and tribal, and local authorities in strengthening and extending microbial contamination from shallow wells (including cess- programs that specifically protect users of recreational pools) is expected to be provided as a result of recent revisions waters. As part of the Beaches Act of 2000, the U.S. Congress to the Underground Injection Control Regulations, published directed EPA to also develop a new set of guidelines for recre-December 7, 1999 (19). ational water based on new water-quality indicators. Begin-To fill gaps in existing data regarding occurrence of micro- ning in 2003, EPA will be conducting a series of epidemiologic bial pathogens and other indicators of microbial contamina- studies at recreational fresh and marine beaches in the United tion, occurrence of disinfection byproducts, and States. These studies will be used to develop guidelines for characterization of treatment processes, EPA promulgated the using the new water-quality indicators to be included in new Information Collection Rule in 1996 (20), which required EPA guidelines.

systems serving >100,000 persons to provide treatment data Data collected as part of the national WBDO surveillance and monitor disinfection byproducts and source-water-quality system are used to describe the epidemiology of waterborne parameters. Surface water systems are also required to moni- diseases in the United States. Data regarding water systems tor for the presence of Cryptosporidium, Giardia, total and deficiencies implicated in these outbreaks are used to culturable viruses, and total* (Glossary) and fecal coliforms assess whether regulations for water treatment and monitor-or Escherichia coli >1 time/month for 18 months. The required ing of water quality are adequate to protect the public against disease. Surveillance also enables identifying etiologic agents

• Total coliforms are considered indicator organisms that typically do not and environmental or behavioral risk factors that are cause disease but might be associated with the presence of other disease-causing organisms. Additional information regarding total coliforms is available at http://www.epa.gov/safewater/dwa/electronic/tcr.pdf. Additional information is available at http://www.cleanwater.gov.

6 MMWR November 22, 2002 responsible for these outbreaks. This information is used to illness after either ingestion of drinking water or exposure to inform public health and regulatory agencies, water utilities, water encountered in recreational or occupational settings.

pool operators, and other stakeholders of new or reemerging This criterion is waived for single cases of laboratory-trends that might necessitate different interventions and confirmed primary amebic meningoencephalitis and for single changes in policies and resource allotment. cases of chemical poisoning if water-quality data indicate con-tamination by the chemical. Second, epidemiologic evidence (Table 1) must implicate water as the probable source of the Methods illness. For drinking water, reported outbreaks caused by con-taminated water or ice at the point of use (e.g., a contami-Data Sources nated water faucet or serving container) are not classified as State, territorial, and local public health agencies have pri- WBDOs.

mary responsibility for detecting and investigating WBDOs, If primary cases (i.e., among persons exposed to contami-and they voluntarily report them to CDC on a standard form nated water) and secondary cases (i.e., among persons who (CDC form 52.12, which is available at http://www.cdc.gov/ became ill after contact with primary persons) are distinguished healthyswimming). The form solicits data related to 1) char- on the outbreak report form, only primary cases are included acteristics of the outbreak, including person, place, time, and in the total number of cases. If both actual and estimated case location of the outbreak; 2) results from epidemiological stud- counts are included on the outbreak report form, the esti-ies conducted; 3) specimen and water sample testing; and 4) mated case count can be used if the population was sampled factors contributing to the outbreak, including environmen- randomly or the estimated count was calculated by applying tal factors, water distribution, and disinfection concerns. Each the attack rate to a standardized population.

year, CDC requests reports from state and territorial epide- Public water systems, which are classified as either commu-miologists or from persons designated as WBDO surveillance nity or noncommunity (Glossary), are regulated under SDWA.

coordinators. Additional information regarding water qual- Of the approximately 170,000 public water systems in the ity and treatment is obtained from the states drinking water United States, 113,000 (66.5 %) are noncommunity systems, agency as needed. Numerical and text data are abstracted from of which 93,000 are transient systems (i.e., public water sys-the outbreak form and supporting documents and are tems that regularly serve >25 of the same persons for >6 entered into a database before analysis. months/year [e.g., highway rest stations, restaurants, and parks with their own public water systems]) and 20,000 are Definitions§ nontransient systems (Glossary). A total of 54,000 systems The unit of analysis for the WBDO surveillance system is (31.8%) are community systems. Community water systems an outbreak, not an individual case of a waterborne disease. serve approximately 264 million persons in the United States Two criteria must be met for an event to be defined as a (96.0% of the U.S. population). Approximately 11 million WBDO. First, >2 persons must have experienced a similar persons (4.0%) rely on private or individual water systems (24,25) (Glossary). These statistics exclude outbreaks associ-

§ Additional terms are defined in the glossary. ated with these sources because they are not intended for drink-TABLE 1. Classification of investigations of waterborne-disease outbreaks United States*

Class Epidemiologic data Water-quality data I Adequate§ Provided and adequate Data were provided regarding exposed and unexposed persons, Historical information or laboratory data (e.g., the history that a and the relative risk or odds ratio was >2 or the p-value was <0.05 chlorinator malfunctioned or a water main broke, no detectable free-chlorine residual, or the presence of coliforms in the water)

II Adequate Not provided or inadequate (e.g., stating that a lake was crowded)

III Provided, but limited Provided and adequate Epidemiologic data were provided that did not meet the criteria for Class I, or the claim was made that ill persons had no exposures in common besides water, but no data were provided IV Provided, but limited Not provided or inadequate

• Outbreaks of Pseudomonas and other water-related dermatitis and single cases of primary amebic meningoencephalitis or of illness resulting from chemical poisoning are not classified according to this scheme.

On the basis of epidemiologic and water-quality data that were provided on CDC form 52.12.

§ Adequate data were provided to implicate water as the source of the outbreak.

Vol. 51 / SS-8 Surveillance Summaries 7 ing and are not considered to be public water systems. Also as the vehicle for the outbreak, but also to understand the excluded from these statistics are the millions of persons who circumstances that led to the outbreak.

use noncommunity systems while traveling or working. A classification of I indicates that adequate epidemiologic In this surveillance system, outbreaks associated with water and water-quality data were reported (Table 1); however, the not intended for drinking (e.g., lakes, springs, and creeks used classification does not necessarily imply whether an investi-by campers and boaters; irrigation water and other nonpotable gation was optimally conducted. Likewise, a classification of sources with or without taps) are also classified as individual II, III, or IV should not be interpreted to mean that the systems (Glossary). Sources used for bottled water are also investigations were inadequate or incomplete. Outbreaks and classified as individual systems; bottled water is not regulated the resulting investigations occur under various circumstances, by EPA but is subject to regulation by the Food and Drug and not all outbreaks can or should be rigorously investigated.

Administration (FDA). In addition, outbreaks that affect fewer persons are more likely Each drinking water system associated with a WBDO is to receive a classification of III, rather than I, on the basis of classified as having one of the deficiencies in the following the relatively limited sample size available for analysis.

list. If >1 deficiency is noted on the outbreak report form, the deficiency that most likely caused the outbreak is noted.

Deficiency classifications are as follows: Results 1: untreated surface water; Outbreaks Associated with Drinking Water 2: untreated groundwater; 3: treatment deficiency (e.g., temporary interruption of During 1999-2000, a total of 39 outbreaks associated with disinfection, chronically inadequate disinfection, or drinking water were reported by 25 states (see Appendix A for inadequate or no filtration); selected case descriptions). One of the 39 outbreaks was a 4: distribution system deficiency (e.g., cross-connection, multistate outbreak of Salmonella Bareilly that included cases contamination of water mains during construction or from 10 states. Of the 39 total drinking water outbreaks, 15 repair, or contamination of a storage facility); and outbreaks were reported for 1999 and 24 for 2000. Florida 5: unknown or miscellaneous deficiency (e.g., contaminated reported the most outbreaks (15) during this period. These bottled water) or water source not intended for drinking 39 outbreaks caused illness among an estimated 2,068 (e.g., irrigation water tap). persons; 122 persons were hospitalized, and two died. The Recreational waters include swimming pools, wading pools, median number of persons affected in an outbreak was 13.5 whirlpools, hot tubs, spas, water parks, interactive fountains, and (range: 2-781). Outbreaks peaked during the summer months fresh and marine surface waters. Although the WBDO surveil- (Figure 1), June-August.

lance system includes whirlpool- and hot tub-associated outbreaks of dermatitis caused by Pseudomonas aeruginosa, wound infec- FIGURE 1. Number of waterborne-disease outbreaks associated with drinking water, by etiologic agent and month tions resulting from waterborne organisms are not included. United States, 1999-2000 (n = 38)*

12 Outbreak Classification WBDOs reported to the surveillance system are classified 10 AGI according to the strength of the evidence implicating water as Number of outbreaks Infectious the vehicle of transmission (Table 1). The classification scheme 8 Chemical (i.e., Classes I-IV) is based on the epidemiologic and water-quality data provided with the outbreak report form. Epide- 6 miologic data are weighted more than water-quality data.

4 Although outbreaks without water-quality data might be included in this summary, reports that lack epidemiologic data 2

were excluded. Outbreaks of dermatitis and single cases of either primary amebic meningoencephalitis or illness result- 0 ing from chemical poisoning were not classified according to Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec this scheme. Weighting of epidemiologic data does not 1999-2000 preclude the relative importance of both types of data. The purpose of the outbreak system is not only to implicate water

• One outbreak of Salmonella Bareilly was not included.

Acute gastrointestinal illness of unknown etiology.

8 MMWR November 22, 2002 Nine of the 39 (23.1%) outbreaks were assigned to Class I Parasites. Seven outbreaks affecting 57 persons were on the basis of epidemiologic and water-quality data; three attributed to parasitic infection: six Giardia outbreaks and (7.7%) were Class II; 25 (64.1%) were Class III; and 1 was one Cryptosporidium outbreak. Six outbreaks of Giardia asso-Class IV (Table 1). One of two outbreaks associated with a ciated with drinking water affected 52 persons from five states:

chemical etiology was not assigned a class because that Florida (two outbreaks), New Mexico (one), New Hampshire outbreak was a single case of illness resulting from nitrate (one), Minnesota (one), and Colorado (one). These outbreaks poisoning associated with consumption of water from a pri- occurred in January (one), June (one), July (one), August (one),

vate well. Outbreaks are listed by state (Tables 2 and 3) and and September (two). Four outbreaks were associated with are tabulated by the etiologic agent, the water system type well water systems, and two were associated with surface (Table 4), and by the type of deficiency and type of water water systems. Two outbreaks caused by G. intestinalis involved system type (Table 5). possible contamination of wells by animal feces. G. intestinalis can infect mammalian hosts, which in turn, can serve as res-Etiologic Agents ervoirs for human infection. Water treatment failure was a Twenty (51.3%) of the 39 outbreaks were of known infec- factor in two other outbreaks of Giardia.

tious etiology; 17 (43.6%) were of unknown etiology; and Bacteria. Nine outbreaks affecting an estimated 1,166 per-two (5.1%) were attributed to chemical poisoning. Of the sons were attributed to bacterial infection: four Es. coli O157:

20 outbreaks with known infectious etiology, seven (35.0%) H7 outbreaks, one Campylobacter jejuni, one Salmonella were caused by parasites; nine (45.0%) were caused by Typhimurium, one Sa. Bareilly, and two mixed Ca. jejuni and bacteria; and four (20.0%) were caused by viruses (Figure 2) shiga toxin-producing Es. coli (O157:H7 or O111) outbreaks.

(Appendix A).

TABLE 2. Waterborne-disease outbreaks associated with drinking water United States, 1999 (n = 15)*

Number Type of State Month Class Etiologic agent of cases system§ Deficiency¶ Source Setting California Jul III AGI** 31 Ncom 2 Well Camp Florida Jan III AGI 4 Com 2 Well Community Florida Jan III Giardia intestinalis 2 Ind 2 Well Household Florida Mar III AGI 6 Com 4 River/stream Apartment Florida Mar III AGI 3 Com 4 Well Community Florida May III AGI 3 Ind 2 Well Household Florida Aug III AGI 4 Com 4 River/stream Apartment Missouri Jun II Salmonella Typhimurium 124 Com 3 Well Community New Jersey Nov IV Sodium hydroxide 2 Com 3 Well Community New Mexico Jul I Small round-structured virus§§ 70 Ncom 3 Spring Camp New York Aug I Escherichia coli O157:H7, Campylobacter jejuni¶¶ 781 Ncom 2 Well Fairgrounds Texas Nov I Es. coli O157:H7 22 Com 3 Well Community Washington Jul II AGI 46 Ind 1 River/creek Household Washington Aug I AGI 68 Ncom 2 Well Soccer match Wisconsin Apr NA*** Nitrate 1 Ind 2 Well Household

• An outbreak is defined as 1) >2 persons experiencing a similar illness after ingestion of drinking water and 2) epidemiologic evidence that implicates water as the probable source of the illness.

On the basis of epidemiologic and water-quality data provided on CDC form 52.12.

§ Com = community; Ncom = noncommunity; Ind = individual; community and noncommunity water systems are public water systems that serve >15 connections or an average of >25 residents for >60 days/year. A community water system serves year-round residents of a community, subdivision, or mobile home park with >15 service connections or an average of >25 residents. A noncommunity water system can be nontransient or transient.

Nontransient systems serve >25 of the same persons for >6 months/year (e.g., factories or schools), whereas transient systems do not (e.g., restau-rants, highway rest stations, or parks). Individual water systems are not owned or operated by a water utility and serve <15 connections or <25 persons.

Outbreaks associated with water not intended for drinking (e.g., lakes, springs, and creeks used by campers and boaters; irrigation water; and other nonpotable sources with or without taps) are also classified as individual systems.

¶ 1 = untreated surface water; 2 = untreated groundwater; 3 = treatment deficiency (e.g., temporary interruption of disinfection, chronically inadequate disinfection, and inadequate or no filtration); 4 = distribution system deficiency (e.g., cross-connection, contamination of water mains during construction or repair, and contamination of a storage facility); and 5 = unknown or miscellaneous deficiency (e.g., contaminated bottled water).

• • Acute gastrointestinal illness of unknown etiology.

Unidentified chemical poisoning.

§§ Three persons had stool specimens that tested positive for small round-structured virus, and one persons stool specimen tested positive for Ca. jejuni.

¶¶ A total of 126 persons had stool specimens that tested positive for Es. coli O157:H7; 43 persons had stool specimens that tested positive for Ca. jejuni.

One persons stool specimen tested positive for both organisms.

• • • Not applicable, see Table 1.

Vol. 51 / SS-8 Surveillance Summaries 9 TABLE 3. Waterborne-disease outbreaks associated with drinking water United States, 2000 (n = 24)*

Number Type of State Month Class Etiologic agent of cases system§ Deficiency¶ Source Setting California Jul I Norwalk-like virus 147 Ncom 2 Well Camp California Jul I Escherichia coli O157:H7 5 Ind 5 River/creek Camp California Sep III AGI** 63 Ind 5 Irrigation system Football game Colorado Aug III Giardia intestinalis 27 Ncom 3 River Resort Florida Mar III AGI 19 Com 3 Well Trailer park Florida Apr III AGI 21 Com 3 Well Trailer park Florida Apr I AGI 71 Ind 2 Well Community Florida Jun III AGI§§ 2 Ind 2 Well Household Florida Jul III AGI 3 Ind 2 Well Household Florida Jul III AGI 3 Ind 2 Well Household Florida Aug III AGI 4 Ind 2 Well Household Florida Sep III G. intestinalis 2 Ind 4 Well Household Florida Dec III Cryptosporidium parvum 5 Com 4 Well Community Idaho Apr III Es. coli O157:H7 4 Ind 5 Irrigation canal Household Idaho Jun III Campylobacter jejuni 15 Ncom 2 Spring Camp Idaho Jul III AGI 65 Ncom 2 Well Restaurant Kansas Jun III Norwalk-like virus 86 Ncom 2 Well Reception hall Minnesota Jun III G. intestinalis¶¶ 12 Ncom 2 Well Camp New Hampshire Sep III G. intestinalis 5 Ind 3 Well Household New Mexico Jul II G. intestinalis 4 Ind 5 River Rafting trip Ohio Aug I Es. coli O157:H7 29 Com 4 Surface water*** Fairgrounds Utah Aug III Ca. jejuni 102 Ind 5 Irrigation water Football camp West Virginia Jun III Norwalk-like virus 123 Ncom 3 Wells Camp Multistate Apr-Aug I Salmonella Bareilly 84 Ind 5§§§ Municipal/spring§§§ Wells/bottled water***

• An outbreak is defined as 1) >2 persons experiencing a similar illness after ingestion of drinking water and 2) epidemiologic evidence that implicates water as the probable source of the illness.

On the basis of epidemiologic and water-quality data provided on CDC form 52.12.

§ Com = community; Ncom = noncommunity; Ind = individual; community and noncommunity water systems are public water systems that serve

>15 service connections or an average of >25 residents for >60 days/year. A community water system serves year-round residents of a community, subdivision, or mobile home park with >15 service connections or an average of >25 residents. A noncommunity water system can be nontransient or transient. Nontransient systems serve >25 of the same persons for >6 months/year (e.g., factories or schools), whereas transient systems do not (e.g., restaurants, highway rest stations, or parks). Individual water systems are not owned or operated by a water utility and serve <15 connections or

<25 persons. Outbreaks associated with water not intended for drinking (e.g., lakes, springs, and creeks used by campers and boaters; irrigation water; and other nonpotable sources with or without taps) are also classified as individual systems.

¶ 1 = untreated surface water; 2 = untreated groundwater; 3 = treatment deficiency (e.g., temporary interruption of disinfection, chronically inadequate disinfection, and inadequate or no filtration); 4 = distribution system deficiency (e.g., cross-connection, contamination of water mains during construction or repair, and contamination of a storage facility); and 5 = unknown or miscellaneous deficiency (e.g., contaminated bottled water).

• • Acute gastrointestinal illness of unknown etiology.

Persons also reported rashes in addition to acute gastrointestinal illness.

§§ One person had a stool specimen that tested positive for Blastocystis hominis.

¶¶ Eight persons had stool specimens that tested positive for G. intestinalis; one stool specimen tested positive for Dientamoeba fragilis.

• • • Type of water was not specified on report form.

Thirty-seven persons had stool specimens that tested positive for Ca. jejuni; four persons stool specimens tested positive for Es. coli O157:H7, and three persons had stool that tested positive for Es. coli O111.

§§§ The outbreak implicated both drinking water from private wells and springs and water bottled by one facility. The bottling facility used two sources of water.

The two outbreaks with multiple pathogens caused the two affected by contamination of drinking water from nitrate and largest bacterial drinking water outbreaks reported during this sodium hydroxide.

study period. Unidentified Etiologic Agents. Seventeen outbreaks Viruses. During this period, four outbreaks involving viral involving gastroenteritis of unknown etiology were reported gastroenteritis were reported. A total of 426 persons reported from four states, affecting an estimated 416 persons and illness; no hospitalizations or deaths were reported in associa- resulting in five hospitalizations. Testing for certain tion with these four viral outbreaks. Three of the four out- enteric pathogens (including ova and parasite testing) was breaks occurred in camp facilities in California, New Mexico, attempted in five of the 17 outbreaks. In a June 2000 and West Virginia. All three water sources were noncommunity outbreak affecting 2 persons, stool specimens collected from groundwater sources. one person tested negative for G. intestinalis but positive for Chemicals. During 1999, two outbreaks involving chemi- Blastocystis hominis. However, whether B. hominis was the cause cal contamination were reported. A total of three persons were of the reported illness is unclear, and the pathogenicity of

10 MMWR November 22, 2002 TABLE 4. Waterborne-disease outbreaks associated with drinking water, by etiologic agent and type of water system United States, 1999-2000 (n = 39)

Type of water system*

Community Noncommunity Individual Total Etiologic agent Outbreaks Cases Outbreaks Cases Outbreaks Cases Outbreaks Cases AGI 6§ 57 3 164 8 195 17 416 Giardia intestinalis 0 0 2 39 4 13 6 52 Escherichia coli O157:H7 2 51 0 0 2 9 4 60 Norwalk-like viruses (NLV) 0 0 3 356 0 0 3 356 Salmonella species¶ 1 124 0 0 1 84 2 208 Campylobacter jejuni 0 0 1 15 1 102 2 117 Es. coli O157:H7/Ca. jejuni 0 0 1 781 0 0 1 781 Small round-structured virus 0 0 1 70 0 0 1 70 Cryptosporidium parvum 1 5 0 0 0 0 1 5 Sodium hydroxide 1 2 0 0 0 0 1 2 Nitrate 0 0 0 0 1 1 1 1 Total 11 239 11 1,425 17 404 39 2,068 Percentage 28.2 11.6 28.2 68.9 43.6 19.5 100.0 100.0

• Community and noncommunity water systems are public water systems that serve >15 service connections or an average of >25 residents for >60 days/

year. A community water system serves year-round residents of a community, subdivision, or mobile home park with >15 service connections or an average of >25 residents. A noncommunity water system can be nontransient or transient. Nontransient systems serve >25 of the same persons for >6 months/year (e.g., factories or schools), whereas transient systems do not (e.g., restaurants, highway rest stations, or parks). Individual water systems are not owned or operated by a water utility and serve <15 connections or <25 persons. Outbreaks associated with water not intended for drinking (e.g.,

lakes, springs, and creeks used by campers and boaters; irrigation water; and other nonpotable sources with or without taps) are also classified as individual systems.

Acute gastrointestinal illness of unknown etiology.

§ One outbreak of four cases was caused by an unidentified chemical.

¶ One outbreak was serotype Typhimurium, and one outbreak was serotype Bareilly.

TABLE 5. Waterborne-disease outbreaks associated with drinking water, by type of deficiency and type of water system United States, 1999-2000 (n = 39)

Type of water system*

Community Noncommunity Individual Total Type of deficiency Outbreaks  % Outbreaks  % Outbreaks  % Outbreaks  %

Untreated surface water 0 0 0 0 1 5.9 1 2.6 Untreated groundwater 1 9.0 8 72.7 8 47.0 17 43.6 Inadequate treatment 5 45.5 3 27.3 1 5.9 9 23.1 Distribution system 5 45.5 0 0 1 5.9 6 15.4 Miscellaneous or unknown 0 0 0 0 6 35.3 6 15.4 Total 11 100.0 11 100.0 17 100.0 39 100.0

• Community and noncommunity water systems are public water systems that serve >15 service connections or an average of >25 residents for >60 days/

year. A community water system serves year-round residents of a community, subdivision, or mobile home park with >15 service connections or an average of >25 residents. A noncommunity water system can be nontransient or transient. Nontransient systems serve >25 of the same persons for >6 months/year (e.g., factories or schools), whereas transient systems do not (e.g., restaurants, highway rest stations, or parks). Individual water systems are not owned or operated by a water utility and serve <15 connections or <25 persons. Outbreaks associated with water not intended for drinking (e.g.,

lakes, springs, and creeks used by campers and boaters; irrigation water; and other nonpotable sources with or without taps) are also classified as individual systems.

Examples of treatment deficiencies include temporary interruption of disinfection, chronically inadequate disinfection, or inadequate or no filtration; examples of distribution system deficiencies include cross-connection, contamination of water mains during construction or repair, or contamination of a storage facility; and examples of unknown or miscellaneous deficiencies include contaminated bottled water.

B. hominis has been debated in the scientific community (26). confirmed. G. intestinalis was suspected in an April 2000 out-Stool specimens were negative for parasitic and bacterial break in a Florida trailer park affecting 21 persons, on the enteric pathogens in two outbreaks in Washington (July 1999 basis of the incubation period and symptoms reported. In and August 1999) and in two Florida outbreaks (March 1999 another outbreak in a Florida trailer park in March 2000 and April 2000) (Appendix A). among 19 persons, a bacterial pathogen was suspected as the In addition, suspected pathogens were noted in four other cause of the outbreak on the basis of the symptoms, which outbreak reports submitted. On the basis of symptoms of ill- included conjunctivitis and dermatitis in addition to gastro-ness, Norwalk-like virus (NLV) was suspected in an Idaho enteritis. A chemical agent was suspected as the cause of ill-outbreak among firefighters that caused 65 illnesses and four ness among four residents in a Florida apartment building hospitalizations, but the outbreak was not laboratory- who had a cross-connection between their drinking water and

Vol. 51 / SS-8 Surveillance Summaries 11 FIGURE 2. Waterborne-disease outbreaks associated with drinking water, by etiologic agent, toxins were not detected. Es. coli water system, water source, and deficiency United States, 1999-2000 (n = 39) O157:H7 was found in the water in a July 2000 California outbreak. In Etiologic agent (n = 39) Water system (n = 39) a 2000 Colorado outbreak, the pres-ence of G. intestinalis was demon-Unidentified Individual strated in a sample from the water 43.6% 43.6%

holding tank, despite the lack of Chemical Bacterial coliform data.

5.1% 23.1% Community 28.2% Of the three outbreaks with either Viral Noncommunity 10.3% 28.2% a confirmed or suspected chemical Parasitic etiology, only one demonstrated that 17.9%

the chemical had been directly in the water. Tap water was tested after the Water source (n = 38)* Deficiency (n = 39)

Untreated health department was notified that groundwater an infant had methemoglobinemia.

Well 43.6%

68.4% Both fecal coliforms and 28 mg/L Treatment Spring 5.3%

Untreated surface deficiency of nitrate were detected in the water 2.6% 23.1%

Miscellaneous water. For an outbreak where burns Irrigation system 7.9% 15.4% and gastroenteritis were reported Surface water Distribution system 15.4%

and linked to a sodium hydroxide 18.4%

spill, a pH test of the water that

• One outbreak of Salmonella Bareilly was not included.

could indicate whether NaOH or another basic substance had spilled into the water was not documented.

a toilet flush-valve. The residents of the apartment had noted However, the environmental assessment indicated the tank blue tap water before onset of illness on multiple occasions contents had emptied into the water. A third suspected chemi-before an improper flush valve in the toilet tank was discovered. cal outbreak involving a cross-connection between a toilet Four outbreaks of gastroenteritis were associated with con- flush-valve and the drinking water system did not have water-sumption of untreated water from private wells. These four quality data available.

outbreaks occurred in Florida and affected 3-4 persons each. In 11 of the 35 outbreaks, water was not sampled for In July 2000, flooding was a possible contributor to two out- coliforms until >1 month after the first case associated with breaks. Water in each of the homes tested positive for coliforms the outbreak was reported (range: 5-16 weeks). In four of and did not have adequate disinfection. these 11 outbreaks, the water samples did not test positive for coliforms (fecal or total), chemicals, or pathogens. Instead, Water-Quality Data these were confirmed as outbreaks by epidemiologic data or Water-quality data (i.e., information regarding the presence by reports that treatment deficiencies had occurred.

of coliform bacteria, pathogens, or chemical contaminants) were available for 35 (89.7%) of the 39 drinking water out- Water Systems and Water Sources breaks. Two reports of outbreaks of confirmed or suspected Eleven (28.2%) of the 39 drinking water outbreaks were infectious etiology and two reports of outbreaks of confirmed associated with community systems, 11 (28.2%) with non-or suspected chemical etiology did not provide water-quality community systems, and 17 (43.6%) with individual water data. systems (Tables 4 and 5). Ten (25.6%) of the 39 drinking Of the 36 reports of outbreaks with a suspected or con- water outbreaks were associated with surface water, including firmed infectious etiology, 33 outbreaks provided water-quality three outbreaks that implicated irrigation water not intended data. Twenty-six (78.8%) of the 33 outbreaks with a suspected for consumption. Twenty-nine (74.4%) of the 39 drinking or confirmed infectious etiology reported a positive coliform, water outbreaks, including the outbreak associated with total coliform, or fecal coliform result. Organisms also were bottled water, were associated with groundwater sources (wells detected in the water in two of these outbreaks. In August and springs).

2000, Ca. jejuni was detected in the water in a mixed Five (45.5%) of the 11 outbreaks associated with commu-Ca. jejuni/Es. coli O157:H7 outbreak in Utah, although shiga nity water systems were caused by treatment deficiencies; one

12 MMWR November 22, 2002 (9.0%) outbreak was related to contaminated, untreated of the 16 outbreaks occurred when persons drank water not groundwater, and five (45.5%) outbreaks were related to intended for direct consumption from irrigation systems or problems in the water distribution system. Two of the five when they consumed surface water that had been ineffectively distribution system problems were related to cross-connections or improperly treated. One (6.3%) of the 16 outbreaks in a between the distribution system and an irrigation well. The system occurred in a home where creek water on the property third outbreak related to a community water source had a was directly consumed without treatment.

household cross-connection between the toilet water and main Of the nine bacterial outbreaks, four occurred in ground-kitchen tap. One outbreak of Cr. parvum (Florida, December water systems (one was associated with a deficiency in the 2000) was related to a repeated history of water main breaks. distribution system, one with a treatment deficiency, and two In another outbreak in Ohio in August 2000, deficiencies in occurred in untreated systems). Six of seven parasitic outbreaks the distribution system of a fairgrounds might have allowed occurred in groundwater systems: three occurred in untreated back-siphonage of animal manure into the water used by food systems; two involved problems in the distribution system; and beverage vendors. and one was related to a treatment deficiency. All four viral Ten (90.1%) of 11 outbreaks associated with noncommu- outbreaks occurred in noncommunity groundwater systems.

nity water systems occurred in groundwater systems. Seven of Two occurred in untreated wells, and two were related to treat-the 10 groundwater outbreaks were linked to untreated wells, ment deficiencies in a spring and well. Two chemical out-and one of the 10 involved consumption of untreated spring breaks were related to treatment deficiencies in well water.

water. Two of the 10 outbreaks were related to treatment Fourteen of the 17 outbreaks of unknown etiology were linked deficiencies in water taken from wells or a spring and to groundwater systems. Ten of these 14 outbreaks occurred were associated with outbreaks of NLV and a small round- in untreated systems; two were related to distribution system structured virus. An outbreak associated with G. intestinalis problems, and two were related to treatment deficiencies.

related to consumption of surface water occurred when a pump failure and a defective filter cartridge resulted in river water Outbreaks Associated entering the drinking water holding tank without filtration. with Recreational Water No information concerning chlorine levels from water samples During 1999-2000, a total of 23 states reported 59 out-was provided.

breaks associated with recreational water (Tables 6-9)

Nine (52.9%) of 17 outbreaks associated with individual (see Appendix B for selected case descriptions). Twenty-three water systems occurred in groundwater systems. Eight of these outbreaks were reported for 1999, and 36 for 2000. The states groundwater systems were wells that were not treated rou-that reported the largest number of outbreaks were Florida tinely; one outbreak of giardiasis occurred when the filtration (14 outbreaks) and Minnesota (eight outbreaks). These system for a well was inadvertently turned off. Five (31.3%)

TABLE 6. Waterborne-disease outbreaks of gastroenteritis associated with recreational water United States, 1999 (n = 15)

Number State Month Class* Etiologic agent Illness of cases Source Setting California Jun III AGI Gastroenteritis 23 Pool Apartment complex Connecticut Jul II Escherichia coli O121:H19 Gastroenteritis 11 Lake Lake Florida Mar III Campylobacter jejuni Gastroenteritis 6 Pool Private home Florida Aug I Shigella sonnei, Interactive Cryptosporidium parvum§ Gastroenteritis 38 fountain Beach park Florida Aug IV Cr. parvum Gastroenteritis 6 Pool Private home Florida Sep III Es. coli O157:H7 Gastroenteritis 2 Ditch water Community Idaho Jun IV Norwalk-like virus Gastroenteritis 25 Hot springs Resort Illinois Jun III AGI Gastroenteritis 25 Lake Lake Massachusetts Jul III Giardia intestinalis Gastroenteritis 18 Pond Pond Minnesota Jul III Cr. parvum Gastroenteritis 10 Pool Trailer park Nebraska Jun IV Es. coli O157:H7 Gastroenteritis 7 Wading pool Child care center New York Jun II Norwalk-like virus Gastroenteritis 168 Lake County park Washington Aug I Es. coli O157:H7 Gastroenteritis 36 Lake State park Wisconsin Jul IV Cr. parvum Gastroenteritis 10 Pool Municipal pool Wisconsin Aug II Es. coli O157:H7 Gastroenteritis 5 Lake/pond Swimming beach

• On the basis of epidemiologic and water-quality data provided on CDC form 52.12.

Acute gastrointestinal illness of unknown etiology.

§ Five persons had stool specimens that tested positive for Shigella sonnei; two stools tested positive for Cryptosporidium parvum.

Vol. 51 / SS-8 Surveillance Summaries 13 TABLE 7. Waterborne-disease outbreaks of gastroenteritis associated with recreational water United States, 2000 (n = 21)

Number State Month Class* Etiologic agent Illness of cases Source Setting Colorado Aug I Cryptosporidium parvum Gastroenteritis 112 Pool Municipal pool Florida May IV AGI Gastroenteritis 2 Lake Lake Florida Jul III AGI Gastroenteritis 4 Outdoor spring County park Florida Jul III Cr. parvum Gastroenteritis 3 Pool Apartment complex Florida Aug III Cr. parvum Gastroenteritis 5 Pool Country club Florida Aug I Cr. parvum Gastroenteritis 19 Pool Resort Florida Aug III AGI Gastroenteritis 9 Pool Motel Florida Aug III Cr. parvum Gastroenteritis 5 Pool Condominium Georgia Jun II Cr. parvum Gastroenteritis 36 Pools§ Community Maine Jul II AGI Gastroenteritis 32 Lake/pond Swimming beach Minnesota Jul II Cr. parvum¶ Gastroenteritis 220 Lake Swimming beach Minnesota Jul IV Shigella sonnei** Gastroenteritis 15 Lake/pond Swimming beach Minnesota Jul III Cr. parvum Gastroenteritis 7 Pool Day camp Minnesota Jul II Cr. parvum Gastroenteritis 6 Pool Hotel Minnesota Aug II Sh. sonnei Gastroenteritis 25 Lake Public beach Minnesota Aug IV Cr. parvum Gastroenteritis 4 Pool Municipal pool Missouri Sep III Shigella flexneri Gastroenteritis 6 Wading pool Community Nebraska Jun I Cr. parvum Gastroenteritis 225 Pools Community Ohio Jun I Cr. parvum Gastroenteritis 700 Pool Private swim club South Carolina Jul IV Cr. parvum Gastroenteritis 26 Pool Community Wisconsin Jan IV Norwalk-like virus Gastroenteritis 9 Pool Motel

• On the basis of epidemiologic and water-quality data provided on CDC form 52.12.

Acute gastrointestinal illness of unknown etiology.

§ Persons swam in a community pool and an inflatable pool.

¶ Seventeen persons had stool specimens that tested positive for Cr. parvum. One person had a stool specimen that tested positive for Giardia intestinalis.

One person had a stool specimen that tested positive for both organisms.

• • Fourteen of 15 stool specimens tested positive for Shigella; one person tested positive for Cr. parvum; and one tested positive for both.

TABLE 8. Waterborne-disease outbreaks of meningoencephalitis, keratitis, leptospirosis, and Pontiac fever associated with recreational water United States, 1999-2000 (n = 8)

Number State Year Month Class* Etiologic agent Illness of cases Source Setting California 2000 Apr NA Naegleria fowleri Meningoencephalitis 1 Mudhole Mudhole Florida 1999 Oct NA N. fowleri Meningoencephalitis 1 Pond Pond Florida 2000 § NA N. fowleri Meningoencephalitis 1 Guam 2000 Jul II Leptospira interrogans Leptospirosis 21 Lake Adventure race Vermont 2000 Feb NA Bromine Chemical keratitis 3 Pool Pool Texas 1999 Sep II Unknown¶ Acute respiratory infection 12 Hot tub Ranch Texas 2000 Jul NA N. fowleri Meningoencephalitis 1 Lake Lake Wisconsin 2000 May I Legionella pneumophila Pontiac fever 20 Whirlpool Motel

• On the basis of epidemiologic and water-quality data provided on CDC form 52.12.

Not applicable.

§ The month the outbreak occurred was not reported; the source and setting were not reported.

¶ Clinical specimens tested negative for Legionella pneumophila serotypes 1 and 6, adenovirus, influenza virus, and parainfluenza virus.

59 outbreaks affected 2,093 persons and resulted in 25 hospi- recreational water contact were reported more frequently in talizations and four deaths. The median size of the outbreak February, March, June, and July. The four cases of primary was 10 persons (range: 1-700). amebic meningoencephalitis occurred in the warmer months Of the 59 outbreaks, 36 were outbreaks of gastroenteritis (April-October).

(Tables 6 and 7); 15 were outbreaks of dermatitis (Table 9);

Etiologic Agents four were cases of meningoencephalitis; and the remaining four outbreaks were of leptospirosis, chemical keratitis, acute Of the 59 recreational water outbreaks, 44 (74.6%) were of respiratory infection of unknown etiology, and Pontiac fever known infectious etiology (Tables 6-9). Of the 36 outbreaks (Table 8). Thirty-one (86.1%) of the 36 outbreaks involving involving gastroenteritis, 17 (47.2%) were caused by para-gastroenteritis occurred during the summer months (i.e., June- sites; nine (25.0%) by bacteria; three (8.3%) by viruses; one August) (Figure 3). Outbreaks of dermatitis associated with (2.8%) by a combination of parasites and bacteria; and the remaining six (16.7%) were of unknown etiology (Figure 4).

14 MMWR November 22, 2002 TABLE 9. Waterborne-disease outbreaks of dermatitis associated with recreational water United States, 1999-2000 (n = 15)

Number State Year Month Class* Etiologic agent of cases Source Setting Alaska 2000 Oct NA Pseudomonas aeruginosa§ 29 Pool/hot tub Hotel Arkansas 1999 Jun NA P. aeruginosa§ 10 Pool Community Arkansas 2000 Feb NA P. aeruginosa§ 26 Pool/ hot tub Motel California 2000 Jun IV Schistosomes** 6 Pond Pond California 2000 Jul IV Schistosomes** 4 Pond Pond Colorado 1999 Feb NA P. aeruginosa¶§ 19 Hot tub Hotel Colorado 1999 Dec NA P. aeruginosa 5 Hot tub Ski lodge Florida 2000 Aug NA P. aeruginosa 6 Hot tub Apartment complex Maine 2000 Feb NA P. aeruginosa¶ 9 Hot tub/pool Hotel Maine 2000 Mar NA P. aeruginosa§ 11 Hot tub Hotel Minnesota 2000 Dec NA P. aeruginosa¶ 16 Hot tub Private Oregon 1999 Jul IV Schistosomes** 2 Lake Lake Vermont 1999 Jun NA P. aeruginosa 9 Hot tub Hotel Vermont 1999 Feb NA P. aeruginosa 11 Hot tub Vacation home Washington 2000 Mar NA P. aeruginosa 10 Pool/hot tub Motel

• On the basis of epidemiologic and water-quality data provided on CDC form 52.12.

Not applicable.

§ Organism isolated from water.

¶ Laboratory-confirmed case.

• • Suspected etiology on the basis of clinical syndrome and setting.

Suspected etiology on the basis of clinical syndrome.

FIGURE 3. Number of waterborne-disease outbreaks fresh or surface water, and 37 (62.7%) with treated (e.g., chlo-associated with recreational water, by illness and month United States, 1999-2000 (n = 58)* rinated) water. Information regarding the water venue for an outbreak of meningoencephalitis was not provided.

18 16 Parasites. Sixteen of the 17 parasitic recreational water out-14 Other breaks involving gastroenteritis were caused by Cr. parvum.

Number of outbreaks Meningoencephalitis 12 Dermatitis The seventeenth outbreak was caused by G. intestinalis.

10 Gastroenteritis Fifteen of the 17 parasitic outbreaks occurred in chlorinated 8

venues; in these outbreaks, inadequate treatment, disrupted 6

chlorine disinfection, or suboptimal pool maintenance were 4

contributing factors to the outbreaks. Cr. parvum is highly 2

resistant to chlorine disinfection and can survive for days in 0

adequately chlorinated pools; therefore, suboptimal chlori-Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec nation of the pool might not be the sole factor contributing 1999-2000 to the occurrence of an outbreak.

Three outbreaks of laboratory-confirmed cryptosporidiosis

• Information regarding the month was not provided for one outbreak of meningoencephalitis. occurred during the 1999 summer swim season. During the 2000 summer swim season, three substantial outbreaks of Cr. parvum occurred that were related to swimming in Of the 23 nongastroenteritis-related recreational outbreaks, municipal pools. In August 2000, an outbreak occurred in seven were attributed to P. aeruginosa, four to free-living amoe-Colorado that affected 112 persons attending a private pool bae, one to Leptospira species, one to Legionella species, and party. In June 2000, the two other cryptosporidiosis outbreaks, one to bromine (Tables 8 and 9). Nine nongastroenteritis-one in Ohio affecting 700 persons and the other in Nebraska related recreational outbreaks were of unknown etiology, eight affecting 225 persons (27), occurred among members of pri-of which were suspected but not confirmed to be caused by vate swim clubs. In both outbreaks, the protracted nature of P. aeruginosa or schistosomes. The ninth outbreak of unknown the outbreaks during >2 months was the result of repeated etiology was suspected to be caused by a virus or by Legionella recontamination of the pools by infected persons continuing pneumophila on the basis of observed symptoms and the epi-to swim; 37 (18%) of 205 persons interviewed in the demiologically implicated vehicle of transmission. Of the 59 Nebraska outbreak admitted to swimming while symptom-recreational water outbreaks, 21 (35.6%) were associated with atic, and 32% swam while ill or during the 2 weeks after their

Vol. 51 / SS-8 Surveillance Summaries 15 FIGURE 4. Waterborne-disease outbreaks of gastroenteritis associated with recreational who had visited the same swimming water, by etiologic agent and type of exposure United States, 1999-2000 beach. After a review of potential risk factors, the only common link found Etiologic agent (n = 36) Type of exposure (n = 36) was swimming at the implicated Cryptosporidium parvum 44.4%

beach. The popular beach Treated water featured a shallow, dammed area that 61.1%

Shigella species AGI* was used for wading. Total and fecal 16.7%

8.3% coliforms were detected in water NLV Escherichia coli Fresh water samples collected before and during 8.3% O157:H7 Other§ 11.1%

38.9% the outbreak, although the levels 11.1%

detected did not exceed levels of Etiologic agent in fresh water (n = 14) Etiologic agent in treated water (n = 22) EPA-recommended guidelines for AGI* Escherichia coli Cryptosporidium microbiologic quality of water (22).

28.6% O157:H7 21.4%

parvum 68.2% One sample that was tested for Shigella Cryptosporidium

/

Cryptosporidium 4.5% Es. coli O157:H7 was negative.

parvum 7.1% NLV Escherichia coli Es. coli O157:H7 was implicated in 14.3% O157:H7 4.5%

Giardia 7.1% Campylobacter jejuni another outbreak among 36 persons Escherichia coli 4.5%

. NLV AGI*

(August 1999) who visited a state park Shigella. species Shigella species O121:H19 14.3% 4.5% 4.5% 9.1% in Washington. Es. coli O121:H19 was 7.1%

implicated in an outbreak in a Con-necticut community (July 2000).

• Acute gastrointestinal illness of unknown etiology.

Norwalk-like virus. Shigella sonnei was implicated in two

§ These included outbreaks of Campylobacter jejuni, Giardia, Escherichia coli O121:H19 and one mixed outbreaks that occurred at swimming Shigella/Cryptosporidium outbreak.

beaches in Minnesota (July and August 2000).

illness (27). Another outbreak (Florida, August 2000) was Nonfreshwater sources were implicated in four bacterial associated with the outbreak that occurred in Ohio. A family recreational water outbreaks involving gastroenteritis. In who were members of the implicated swim club in the Ohio March 1999, an outbreak of Ca. jejuni was associated with a outbreak were vacationing with a sick child. While in a pool private pool in Florida that did not have continuous chlorine in Florida, the infant had two fecal accidents. The resulting disinfection and reportedly had ducks swimming in the pool.

outbreak caused five cases of diarrheal illness and two hospi- Outbreaks of Shigella flexneri and Es. coli O157:H7 (Mis-talizations. souri, September 2000 and Nebraska, June 1999) occurred Eight other outbreaks of cryptosporidiosis occurred in among children using unchlorinated wading pools. Fecal treated venues during the 2000 swim season. Two outbreaks accidents were factors contributing to the contamination of of gastroenteritis occurred in untreated venues: one in a fresh- the water in both outbreaks. Es. coli O157:H7 also was impli-water lake in Minnesota in July 2000 and one in a Massachu- cated as a cause of illness in an outbreak (Florida, September setts pond in July 1999 (Appendix B). 1999) among two young children who had been playing in Four cases of laboratory-confirmed primary amebic men- ditch water. Both clinical specimens and water samples tested ingoencephalitis attributed to Naegleria fowleri occurred dur- positive for Es. coli O157:H7.

ing this 2-year reporting period. All four persons were aged Two nongastroenteritis-related recreational water outbreaks

<19 years. Three of the persons died from infection after hav- were also reported. One outbreak of leptospirosis was reported ing contact with a pond, lake, or mud hole. The fourth persons among 21 persons who participated in an adventure race in freshwater exposure could not be determined; that person had Guam in July 2000 (Table 8). These persons reported mul-fallen from a jet ski into an unspecified body of water, sus- tiple outdoor exposures, including running through jungles tained injuries, and died from an infection shortly after it was and savannahs, swimming in a river and a reservoir, and bicy-detected. cling and kayaking in the ocean. Leptospira was confirmed by Bacteria. Nine recreational outbreaks involving gastroen- serology, and an epidemiologic investigation demonstrated teritis were attributed to bacterial pathogens, and five of the that swimming in the reservoir, submerging ones head in the nine were linked to freshwater sources. Five cases (Wisconsin, water, and swallowing water while swimming were risk fac-August 1999) of Es. coli O157:H7 occurred among persons tors for illness. Water samples were not tested, and an

16 MMWR November 22, 2002 environmental assessment of the reservoir was not conducted. illness. Although clinical specimens (urine, blood, sputum, The second nongastroenteritis-related recreational water out- and throat swabs) were tested for organisms, including break was an outbreak of Pontiac fever epidemiologically Leg. pneumophila serogroups 1 and 6, influenza virus, parain-linked to use of a whirlpool at a hotel. fluenza virus, and adenovirus, no infectious agent was identi-Viruses. During 1999-2000, three outbreaks of NLV fied. No testing for biologic or chemical agents was performed (Calicivirus) that affected a total of 202 persons were reported. on water samples because the hot tub had already been drained, Two NLV outbreaks occurred in untreated systems; one out- refilled, and hyperchlorinated before the environmental break of NLV occurred (Idaho, June 1999) at a resort and investigation.

water park and affected 25 persons. The pool implicated in During the 1999-2000 reporting period, 15 outbreaks of the investigation was untreated because the source of the pools dermatitis were identified (Table 9). Three of these outbreaks water was a natural hot springs that was high in mineral con- were associated with swimming in freshwater and were tent. The investigators noted that geothermal pools used for assumed to be cercarial dermatitis caused by contact with the swimming are not required to be regulated by public health larval form (cercariae) of schistosomes, which are present in officials in that locale. The pool implicated by the freshwater environments. Two of these dermatitis outbreaks investigation also had been implicated in a previous outbreak occurred in lakes in California that were associated with past of NLV in June 1996. cases of cercarial dermatitis. The onset of dermatitis occurred Other. During 1999-2000, six recreational water outbreaks within hours after swimming in the lake and resolved after a involving gastroenteritis of unknown etiology were reported. limited number of days (median days of illness were 2 and 3 One outbreak (Florida, August) involved a motel pool that days [range: 2-3 and 3-5 days], respectively). The 12 remain-was cloudy and dirty at the time of exposure. Nine persons ing outbreaks were associated with pool and hot tub use and who swam in this pool and did not share any other common affected 5-29 persons each. P. aeruginosa was confirmed in exposure became ill with gastroenteritis. Disinfectant residu- clinical isolates in 3 of the 12 outbreaks and was confirmed als and operation of the filtration system at the time of the in water/filter samples in five outbreaks, two of which also investigation were deficient. Problems were also noted with had a clinical isolate. In eight of these outbreaks of dermati-the equipment used for adjusting pH. tis, specific treatment deficiencies or problems were identi-Another outbreak (Florida, August 1999) among 38 per- fied. Outbreaks in Arkansas (June 1999), Florida (August sons who visited a beach park was attributed to both Sh. sonnei 2000), Colorado (December 1999), and Washington (March and Cr. parvum (28). Illness was epidemiologically linked to 2000) were attributed to deficiencies in treatment.

playing in an interactive fountain at the park, ingesting water, In one outbreak of dermatitis (Maine, February 2000), nine and consuming food and beverages at the fountain. The persons reported rash in addition to headache, fever, fatigue, fountains recirculation, filtration, and disinfection systems and sore throat (29). Swimming in the hot tub or swimming were not approved by the health department and were inad- in the pool was a risk factor. The pool and hot tub were on equate or not completely operational at the time of its use. separate filtration systems, and both were used by the major-Samples of the fountain water tested positive for coliforms ity of persons in the outbreak. Low levels of free chlorine were but did not test positive for fecal coliforms. Nevertheless, the found in the pool and hot tub, but the presence of chlorinate cause of the outbreak was determined to be the fountain, which isocyanurates (chlorine stabilizers) might have influenced was closed until the health departments concerns could be measured levels of free chlorine. A clinical isolate of P. eruginosa remedied. was obtained from an ill person; P. aeruginosa also was isolated Three cases of chemical keratitis (Vermont, February 2000) from the pool filter even after the pool had been cleaned twice.

resulted from exposure to bromine in a hotel swimming pool. P. aeruginosa was isolated from clinical specimens and Bromine levels were >5 ppm (acceptable bromine levels are water samples in an outbreak at a Colorado hotel that 1-3 ppm), and the pH level was >8.5. Patrons who spent affected 19 persons in February 1999 (28), 13 of whom were time with their heads underwater with their eyes open were children aged <15 years. Symptoms were not limited to rash; affected. they included diarrhea, vomiting, nausea, fever, fatigue, muscle Twelve persons affected in an outbreak (Texas, September aches, joint pain, swollen lymph nodes, and subcutaneous 1999) reported symptoms that included exhaustion, sore nodules on hands and feet. Because of the severity and range muscles, headache, chills, and fever after attending a confer- of symptoms, clinical specimens were examined for enteric ence at a guest ranch. One woman reported a miscarriage bacterial and parasitic pathogens as well as Legionella species, during her illness. Exposure to a hot tub, defined as either Leptospira species, and Entamoeba histolytica but did not test immersion or being near the hot tub, was associated with positive for any of these etiologic agents. Swabs taken from

Vol. 51 / SS-8 Surveillance Summaries 17 the hot tub floor and rail were positive for P. aeruginosa and An outbreak of acute respiratory illness occurred among other Pseudomonas species. Pool and hot tub records indicated sugar beet processing plant workers (Minnesota, August 2000).

that chlorine and pH had declined below the state-mandated Of the 15 cases identified, 13 were hospitalized. Serology for levels at the time of exposure. Epidemiologic evidence impli- 4 (26.7%) of the 15 persons tested positive for Leg.

cated the hot tub as the likely vehicle of exposure for the out- pneumophila; three (20.0%) persons were confirmed positive break. In both the Colorado outbreak and the Maine outbreak for Leg. pneumophila by sputum polymerase chain reaction that occurred in February 2000, an offsite contractor had been (PCR). Fourteen (93.3%) of the 15 persons worked on a crew engaged to monitor disinfectant and pH levels. Insufficient that had performed high-pressure cleaning in one area of the communication between pool staff and the remote monitor- plant; the fifteenth patient had conducted high-pressure clean-ing company might have contributed to extended periods of ing elsewhere in the plant. The sources of water for the high-usage with inadequate disinfection (28). pressure cleaning contained 105 colony-forming unit (CFU)/

mL of Leg. pneumophila and endotoxin levels of 22,200 Outbreaks Associated with Occupational endotoxin units/mL. Although the attack rate, symptoms, and Exposure to Water laboratory findings were consistent with an outbreak of Pontiac fever, endotoxin exposure might have contributed to this Two outbreaks not associated with drinking or recreational outbreak.

water exposure were reported during this period (Table 10).

One outbreak of leptospirosis (Hawaii, August 1999) occurred among persons landscaping a pond. Leptospirosis was con- Previously Unreported Outbreaks firmed serologically for the two persons who had contact with Three previously unreported drinking water outbreaks that the pond. Both persons reported multiple skin abrasions and occurred in 1995 and 1997 were submitted during this were exposed to the pond water for a period of 5-10 days. reporting period (Table 11). An illegal cross-connection One of the two persons was hospitalized. (Washington, July 1995) between a domestic water supply TABLE 10. Waterborne-disease outbreaks associated with occupational exposures United States, 1999-2000 (n = 2)

Number State Year Month Class* Etiologic agent Exposure of cases Source Setting Hawaii 1999 Aug IV Leptospira Contact with pond water 2 Pond Outdoor landscaping Minnesota 2000 Aug III Pontiac fever High-pressure cleaning 15 Plant lagoon Sugar beet plant using lagoon water

• On the basis of epidemiologic and water-quality data provided on CDC form 52.12.

Endotoxin was also isolated from environmental samples; the role of endotoxin is unclear.

TABLE 11. Waterborne-disease outbreaks associated with drinking water that were not included in the previous surveillance summaries United States, 1995-1997 (n = 3)*

Number Type of State Year Month Class* Etiologic agent of cases system§ Deficiency¶ Source Setting Washington 1995 Jul III Giardia intestinalis** 87 Com 4 Well Community California 1997 Nov III Nitrite (sodium metaborite) 7 Com 4 Mixed river/groundwater Hospital cafeteria New York 1997 Dec I Norwalk-like virus 1,450 Ncom 3 Well Ski resort

• An outbreak is defined as 1) >2 persons experiencing a similar illness after either ingestion of drinking water or exposure to water used for recreational purposes and 2) epidemiologic evidence that implicates water as the probable source of illness.

On the basis of epidemiologic and water-quality data provided on CDC form 52.12.

§ Com = community; Ncom = noncommunity; Ind = individual; Community and noncommunity water systems are public water systems that serve >15 connections or an average of >25 residents for >60 days/year. A community water system serves year-round residents of a community, subdivision, or mobile home park with >15 service connections or an average of >25 residents. A noncommunity water system can be nontransient or transient. Nontransient systems serve >25 of the same persons for >6 months/year (e.g., factories or schools), whereas transient systems do not (e.g., restaurants, highway rest stations, or parks). Individual water systems are not owned or operated by a water utility and serve <15 connections or <25 persons. Outbreaks associated with water not intended for drinking (e.g., lakes, springs, and creeks used by campers and boaters; irrigation water; and other nonpotable sources with or without taps) are also classified as individual systems.

¶ 1 = untreated surface water; 2 = untreated groundwater; 3 = treatment deficiency (e.g., temporary interruption of disinfection, chronically inadequate disinfection, and inadequate or no filtration); 4 = distribution system deficiency (e.g., cross-connection, contamination of water mains during construction or repair, and contamination of a storage facility); and 5 = unknown or miscellaneous deficiency (e.g., contaminated bottled water).

• • Thirty-three persons had stool specimens that tested positive for G. intestinalis. One specimen tested positive for Entamoeba coli. One other specimen tested positive for Blastocystis hominis. One cultured specimen tested positive for Campylobacter jejuni.

18 MMWR November 22, 2002 and an irrigation system at a plant nursery resulted in con- bathroom facility; a third outbreak involved water taken from tamination of multiple wells in a community. Eighty-seven a garden hose (the water had been stored in an ice chest cases of gastroenteritis were reported, and one hospitalization before consumption at a private residence); and a fourth out-was recorded. G. intestinalis was determined in 33 (52.4%) of break associated with bottled water that might have been con-63 stool specimens; Entamoeba coli and B. hominis were each taminated at the point of use. Two of the six reported found in one stool specimen. One (7.1%) of 14 stool speci- point-of-use outbreaks involving a suspected chemical expo-mens that were cultured for Ca. jejuni tested positive. sure occurred in food service facilities, but water testing was NLV was implicated as the cause of an outbreak (New York, not performed to verify the presence of the chemical; and, December 1997) of 1,450 cases at a restaurant at a ski resort. because of the relatively limited number of cases associated Epidemiologic data implicated water or consumption of ice with these incidents, the epidemiologic information was not made from water as the cause of the outbreak. The environ- adequate to include these incidents as outbreaks.

mental assessment revealed possible problems with the well Data from six other possible or confirmed outbreaks were operation and location. The chlorinator for the well had been also not included in this analysis. One confirmed outbreak of malfunctioning and had already been disconnected before the leptospirosis was related to travel outside the United States or assessment. Testing of the water by the local health depart- its territories and therefore was excluded. This outbreak ment determined that neither a free nor total chlorine residual occurred among student travelers who became ill after their was detectable in the potable water supply and indicated the return from Ecuador. Three cases of leptospirosis were con-presence of fecal coliforms. In addition, the well was located firmed by laboratory testing among the cohort, and four

<24 inches away from a stream. During the period the chlori- additional cases were suspected. Three other outbreaks of nator was not functioning, the pump for the well had been G. intestinalis, Cr. parvum, and NLV could not be included continuously pumping water. Surface water that might have in the analysis. Although these outbreaks were probably caused been introduced into the water supply, plus a deficiency in by a recreational water exposure, the data provided did not treatment of the water, played a key role in the outbreak. meet the criteria for inclusion (i.e., the outbreaks did not meet Seven persons who were either employees or visitors at a the criteria for Classes I-IV).

hospital (California, November 1997) were symptomatic for Two additional outbreaks were excluded because of inad-methemoglobinemia in one outbreak. An epidemiologic equate information: one outbreak of dermatitis caused by investigation indicated that the only shared exposure among in-home bathing and one potential drinking water outbreak these persons was a visit to the hospital cafeteria and the con- of Cr. parvum in a New England community. This outbreak sumption of a carbonated beverage with ice from the self- of Cr. parvum occurred in a community near another reported service soda dispenser. The onset of symptoms occurred 1-5 community outbreak of G. intestinalis and Cr. parvum in 1999.

minutes after or while drinking a carbonated beverage. One The pond implicated in the recreational water outbreak of person was hospitalized, and no deaths occurred. The envi- G. intestinalis also served as a surface water source, which was ronmental investigation discovered a cross-connection in the intermittently mixed into the municipal drinking water that plumbing system that might have allowed water from the cool- supplied the community. However, not all the persons received ing tower, which had been recently shock-treated with their drinking water from the municipal water source.

sodium metaborate, to be drawn into the drinking water sys- Although raw surface water samples later tested positive for tem. Sodium metaborate has been associated with nitrate poi- Cr. parvum by immunomagnetic spectroscopy (IMS), house-soning and methemoglobinemia in past incidents (30). hold water samples either tested negative, were not tested, or the results were not provided. The epidemiologic informa-Outbreaks Not Classified as WBDOs tion and water-quality information provided were not con-Outbreaks attributed to drinking water that was contami- clusive.

nated or potentially contaminated at the point of use rather than at the source or in the distribution system are not classi-Discussion fied as WBDOs. Six outbreaks, causing illness among a total of 102 persons, are in this category. None of the six outbreaks Considerations Regarding Reported reflected a common vehicle of contamination: one outbreak Results of Cr. parvum was epidemiologically associated with ice con-The WBDO surveillance system provides information con-sumption; a school-based outbreak of Sh. sonnei was related cerning epidemiologic and etiologic trends in outbreaks. In to consumption of water from a dispenser stored in a

Vol. 51 / SS-8 Surveillance Summaries 19 previous years, a decrease in the number of drinking water- where the majority of outbreaks actually occurred. An increase associated outbreaks had been observed. However, the in the number of outbreaks reported could either reflect an cumulative number of drinking water outbreaks reported for actual increase in outbreaks or an improved sensitivity in the 1999-2000 period demonstrates a reversal of this trend surveillance practices.

(Figures 5 and 6). The number of recreational water outbreaks Recognition of WBDOs also is dependent on certain out-has been gradually increasing for the past 15 years and is at break characteristics; outbreaks associated with serious illness the highest level since CDC began receiving such reports in or affecting a substantial number of persons are more likely 1978. to receive attention from health authorities. Outbreaks Although the number of outbreaks reported through involving acute diseases, including those characterized by a the surveillance system has increased, the significance of this FIGURE 5. Number of waterborne-disease outbreaks associated with drinking water, by increase is unclear. Whether this year and etiologic agent United States, 1971-2000 (n = 730)*

indeed reflects a true increase in the number of outbreaks that occurred 60 in the United States is unknown. AGI Not all outbreaks are recognized, 50 Chemical investigated, and then reported to Viral Number of outbreaks Parasitic CDC or EPA, and studies have not 40 Bacterial been performed that assess the sen-sitivity of this system and indicate 30 what percentage of actual out-breaks this system is able to detect. 20 Multiple factors exist that can influence whether WBDOs are 10 recognized and investigated by local, territorial, and state public 0 health agencies: the size of the out- 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998 break; severity of disease caused by Year the outbreak; public awareness of the outbreak; routine laboratory

• The total from previous reports has been corrected from n = 691 to n = 688.

testing for organisms; require- Acute gastrointestinal illness of unknown etiology.

ments for reporting cases of dis-eases; and resources available to the FIGURE 6. Number of waterborne-disease outbreaks associated with drinking water, by year and type of water system United States, 1971-2000 (n = 730)*

local health departments for sur-veillance and investigation of prob- 60 able outbreaks. This surveillance Individual system probably underreports the 50 Noncommunity Number of outbreaks true number of outbreaks because Community 40 of the multiple steps required before an outbreak is identified 30 and investigated. In addition, changes in the capacity of local, 20 county, and state public health agencies and laboratories to detect 10 an outbreak might influence the 0

numbers of outbreaks reported in each state relative to other states. 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998 The states with the majority of Year outbreaks reported during this period might not be the states

• The total from previous reports has been corrected from n = 691 to n = 688.

20 MMWR November 22, 2002 short incubation period, are more readily identified than Outbreaks Associated with Drinking outbreaks associated with chronic, low-level exposure to an Water agent (e.g., certain chemicals) or are associated with organ- The number of outbreaks reported during 1999 (15) and isms that have a longer incubation period (e.g., certain para-2000 (24) is higher than the number reported during 1997 sitic organisms). Outbreaks involving larger drinking water (7) and 1998 (10). As described previously, the number of systems (e.g., community systems) are more likely to be drinking water outbreaks had declined (2,3). The increase in detected than outbreaks that involve noncommunity systems reported outbreaks should be carefully interpreted. Although because these systems serve mostly nonresidential areas and the number of drinking water outbreaks has changed, the transient populations. Outbreaks associated with individual total number of persons affected by a drinking water out-systems are the most likely to be underreported because they break during 1999-2000 (n = 2,027) is comparable to what typically involve a limited number of persons. Recreational was initially reported in 1997-1998 (n = 2,038) and 1995-outbreaks where persons congregate in one venue and then 1996 (n = 2,567) (2,3). Changes in surveillance and report-are geographically dispersed can be difficult to document.

ing of outbreaks might have improved detection of outbreaks The identification of the etiologic agent of a WBDO affecting limited, private systems that in turn, affect a rela-depends on the timely recognition of the outbreak so that tively limited number of persons. However, the increase in appropriate clinical and environmental samples can be col-outbreaks that affect persons in limited, private systems mer-lected. The laboratory involved in the testing of specimens its further investigation by public health and water-quality must have the capacity and capability to test for a particular agencies.

organism. In certain cases, specific tests must be requested.

Certain states reported drinking water outbreaks for the Routine testing of stool specimens at laboratories will include first time in >10 years (e.g., Connecticut since 1976 or Utah tests for the presence of enteric bacterial pathogens and might since 1986). California reported multiple drinking water out-also include an ova and parasite examination. However, breaks after reporting no outbreaks in 1997 and 1998, and Cr. parvum, one of the most commonly reported waterborne compared with other years during 1990-2000, the number parasites, is not often included in standard ova and parasite of reported outbreaks in California increased slightly. The examinations and in certain instances, must be specifically number of outbreaks reported by Florida also increased.

requested (31). During 1999-2000, tests for NLV and other Although the numbers of reported outbreaks increased over-possible agents of viral origin were rarely performed or docu- all, the seasonality of the drinking water outbreaks is consis-mented in the outbreaks that were reported to CDC. Collec- tent with previous years, with the number peaking during the tion of water-quality data depends primarily on local and state summer months. The observed increase in the number of out-statutory requirements, the availability of investigative per- breaks is associated with an increase in outbreaks associated sonnel, and the technical capacity of the laboratories that test with consumption of untreated water from both surface and the water. Furthermore, certain outbreaks can substantially groundwater sources, but specifically private wells.

alter the relative proportion of cases of waterborne disease The percentage of drinking water outbreaks associated with attributed to a particular agent. The number of reported cases surface water during 1999-2000 was 17.9% (i.e., seven out-is typically an approximate figure, and the method and accu-breaks) (Figure 2). This percentage is higher than the 11.8%

racy of the approximation vary among outbreaks. reported during 1997-1998 period (i.e., two outbreaks).

One key limitation of the data collected as part of the However, three of the seven surface water outbreaks reported WBDO surveillance system is that the information collected during 1999-2000 were associated with the direct ingestion pertains only to outbreaks of waterborne illness. The epide- of surface water without any treatment or with inadequate miologic trends and water-quality concerns observed in out- individual treatment. Two of these outbreaks were associated breaks might not necessarily reflect or correspond with trends with consumption of water during outdoor excursions where associated with endemic waterborne illness. CDC and EPA point-of-use treatment (e.g., filtration or disinfection) might are collaborating on a series of epidemiologic studies to assess have been attempted and was either inadequate to protect the magnitude of nonoutbreak waterborne illness associated health or was inconsistently or incorrectly applied. The third with consumption of municipal drinking water and with outbreak occurred after a household had run out of potable exposure to recreational marine and freshwaters. water and instead served untreated creek water to their guests.

These three outbreaks illustrate that the public might be unaware that surface water, despite its clarity, is prone to contamination by organisms. Surface waters should not be

Vol. 51 / SS-8 Surveillance Summaries 21 directly consumed without being treated at the point of use from groundwater sources and should establish a targeted or boiled. Manufacturers of point-of-use devices and the strategy to identify groundwater systems at high risk for fecal National Sanitation Foundation (NSF) provide information contamination.

regarding different devices, instructions for use, and their Twenty-six of these 28 groundwater outbreaks had a well as ability to make water safe for human consumption.¶ The the implicated water source, and two were linked to a spring.

remaining four outbreaks comprise approximately 11% of all The percentages of outbreaks associated with wells and springs drinking water outbreaks, an equivalent percentage to that were similar during this reporting period to the 1997-1998 reported in 1997-1998. These four outbreaks were associ- period. Although GWR is expected to have public health ben-ated with systems that routinely received treatment. One out- efits, these protections extend primarily to community ground-break of giardiasis occurred at a resort (Colorado, August 2000) water systems. Of the 26 well-related outbreaks that occurred served by a noncommunity system. The increased demand during the 1999-2000 period, only eight of 26 were associ-for water during the summer, coupled with multiple treat- ated with community wells. Ten were associated with indi-ment failures, resulted in the delivery of unfiltered and vidual private wells, and eight were associated with nondisinfected water to the resort. These multiple failures noncommunity wells. These systems would not necessarily illustrate the importance of routine maintenance, specifically benefit from the promulgation of GWR, and therefore, the among noncommunity systems, which do not have consis- quality of water in wells remains a public health concern.

tent demand for water year-round. Two outbreaks (Florida, Approximately 14-15 million households in the United States March 1999 and August 1999) were associated with cross- rely on a private, household well for drinking water each year, connections: one to an irrigation well and another to a toilet. and >90,000 new wells are drilled throughout the United States Another surface water outbreak (Ohio, August 2000) at a fair- each year (32). In addition, contamination of a private well is grounds was suspected to have resulted from back-siphonage not only a health concern for the household served by the into the drinking water from an animal manure site. These well, but can impact households using other nearby water outbreaks indicate that even when treatment of water at the supplies and could potentially contaminate the aquifer.

source is adequate, deficiencies in the distribution system or Additional education efforts should be targeted towards well at the home can result in illness. Such deficiencies are pre- owners, users, well drillers, and local and state drinking water ventable, and the public should be informed of how to detect personnel to encourage practices that best ensure safe drink-and avoid creating cross-connections. ing water for private well users.

Twenty-eight (71.8%) of the 39 outbreaks related to drink- Three outbreaks were associated with direct consumption ing water were associated with groundwater sources. This of water from irrigation systems, comprising approximately number is an 87% increase from the number reported in the 8% of drinking water outbreaks (Figure 2). Cross-connections previous period (i.e., 15). Seventeen of the 28 outbreaks to irrigation systems were implicated as contaminating fac-(60.7%) were linked to consumption of untreated ground- tors in three other irrigation-related outbreaks. Irrigation water; eight of 28 (28.6%) outbreaks were associated with waters are not regulated under the Safe Drinking Water Act, treatment deficiencies; and three (10.7%) were linked to because they are typically intended for agricultural purposes, deficiencies in the distribution system. The observed pattern not for human consumption. Therefore, irrigation water would of deficiencies is contrary to what was observed in the previ- not be expected to be treated to reduce the level of microor-ous reporting period, where the majority of groundwater out- ganisms or other contaminants potentially in the water to the breaks were associated with treatment or distribution system same standards as water intended for consumption. In one problems. This pattern indicates that untreated groundwater outbreak, children drank directly from an irrigation systems are increasingly associated with outbreaks of illness. canal while playing outside a home. In two other outbreaks, Groundwater systems, with the exception of systems influ- water was directly consumed from an irrigation tap by sports enced by surface water, are not routinely required to use fil- team members. In the first instance, the sports team consumed tration or treatment that would be expected to reduce the water from a labeled irrigation tap despite being informed number of pathogens in the water. EPAs pending GWR** is that the water was not intended for consumption. In the expected to establish multiple barriers in groundwater sys- second instance, two teams drank from taps on the field tems to protect against bacteria and viruses in drinking water Although EPA does not regulate private wells and will not regulate them

¶ Additional information is available at http://www.nsf.org. as part of the proposed GWR, EPA lists recommendations for protecting

• • Additional information is available at http://www.epa.gov/safewater/ private water supplies at http://www.epa.gov/safewater/pwells1.html and gwr.html. provides links to other sources of information.

22 MMWR November 22, 2002 because no other source of potable water was available on number of outbreaks reported that were associated with indi-field. vidual systems during this period is the highest reported level The multistate outbreak of Sa. Bareilly, which was detected since 1984. The drinking water quality of community sys-through CDCs Salmonella Outbreak Surveillance Algorithm tems, which typically have been the focus of increased EPA (SODA), epidemiologically implicated the consumption of regulation, has continually improved. But noncommunity bottled water as a risk factor for illness. This is the first wide- systems and individual systems, which are not regulated to spread outbreak implicating bottled water in the United States. the same extent, are continuing problems. The majority of Previous bottled water outbreaks occurred in New Jersey in these individual system outbreaks are linked to currently 1973 (33), Pennsylvania in 1980 (34), and in the Northern unregulated groundwater supplies, specifically private wells.

Mariana Islands in 1993 (4). Bottled water standards and regu- The populations served by these systems merit increased lations, unlike the majority of drinking water standards, are attention by public health officials.

not set and enforced by EPA but by FDA. FDA regulates The etiologic agent was not identified in 17 (43.6%) of 39 bottled water as a packaged food product and bases their outbreaks (Figure 2). These outbreaks of unknown etiology bottled water standards on EPAs tap water standards. In comprised the largest group of outbreaks, followed by out-addition, bottled water might be subject to state and volun- breaks caused by bacteria (nine), parasites (seven), viruses tary industry regulation. Bottled water, before this outbreak, (four), and chemicals (two). During 1997-1998, parasites had not been identified as a vehicle for transmission of infec- accounted for the largest percentage of the 17 outbreaks (six tious organisms in the United States, although a bottled [35.3%]), followed by unidentified pathogens (five [29.4%]),

water outbreak of Ca. jejuni associated with consumption of bacteria (four [23.5%]), chemicals (two [11.6%]), and viruses water bottled in Greece was documented during the 1997- (zero [0%]). The number of outbreaks per type of agent were 1998 surveillance period (2). Because of the wide geographic increased for all categories during 1999-2000, with the distribution of bottled water products, an outbreak associ- exception of chemicals. Although the number of reported ated with the consumption of bottled water would be diffi- viral outbreaks increased, indicating an improvement in the cult to recognize. FDA, EPA, CDC, and the bottled water availability and usage of laboratory detection methods dur-industry together should address concerns regarding consump- ing previous years, viral outbreaks are probably substantially tion of bottled water and public health. underreported. Although viruses were suspected in other out-Overall, the number of outbreaks associated with the five breaks, specifically in those of unknown etiology, testing for drinking water deficiencies (untreated surface water, untreated viruses was not performed. The technology for detection of groundwater, treatment deficiency, distribution system defi- viruses in stool and water samples has improved, but testing ciency, and unknown/miscellaneous deficiency) increased in for viruses is not widely practiced. Investigators are encour-each category from the 1997-1998 levels. The percentage of aged to submit clinical specimens to CDC or state laborato-outbreaks caused by a treatment deficiency and distribution ries that conduct these tests. Guidelines for collecting stool system problem decreased relative to reported increases in the specimens for identification of viral organisms are available other three categories. Although problems with treatment and from CDC (35). Investigators are also encouraged to contact with distribution systems remain critical concerns for safe CDC and EPA regarding testing of water samples.

drinking water, the publics lack of understanding of the risk Only two outbreaks of chemical origin were identified dur-associated with consumption of untreated water and the ing this surveillance period, the same number as was reported assumption that all water is suitable for consumption is a during the 1997-1998 period. One outbreak related to a spill concern also. of sodium hydroxide at a community water treatment plant The relative proportion and number of outbreaks associ- demonstrated the need for safe water treatment practices. The ated with different water systems also differs from the figures other outbreak was a single case of methemoglobinemia in an from the 1997-1998 period (Figure 2). Outbreaks in com- infant who required hospitalization after having been fed munity systems increased from 8 to 11 outbreaks (37.5% boiled water taken from a private well. Coordination of pub-increase); noncommunity outbreaks doubled from 5 to 11; lic health messages is critical; an intervention that was and individual system outbreaks quadrupled, increasing from intended to reduce the transmission of infections agents con-4 to 17 outbreaks. However, the proportion of outbreaks in centrated the chemicals present in the water. These figures, as community systems decreased from 47.1% during 1997-1998 in the past, probably underrepresent the actual waterborne to 28.2% during 1999-2000, whereas the relative proportion chemical poisonings that occur. Multiple factors can explain of outbreaks in individual systems increased from 23.5% dur- the low reporting rate, including the likelihood that 1) the ing 1997-1998 to 43.6 during 1999-2000. In addition, the majority of waterborne chemical poisonings typically occur

Vol. 51 / SS-8 Surveillance Summaries 23 in private residences and affect a relatively limited number of Because swimming is essentially a shared water activity or persons; 2) exposures to chemicals through drinking water communal bathing, rinsing of soiled bodies and overt fecal might cause illness that is difficult to link to a chemical expo- accidents cause contamination of the water. Unintentional sure; 3) the mechanisms for reporting waterborne chemical ingestion of recreational water contaminated with pathogens poisonings to the WBDO surveillance system are not as can then lead to gastrointestinal illness, even in nonoutbreak established for chemicals as they are for WBDOs attributed settings (36,37). Fresh and marine waters are also subject to to infectious agents; and 4) health-care providers and those other modes of contamination from point sources (i.e., sew-affected might not as easily recognize chemical poisonings. age releases), watersheds (i.e., runoff from agriculture and resi-As a result of these factors, WBDOs of chemical poisonings dential areas), and floods.

are less likely to be reported to public health officials. Outbreaks involving gastroenteritis are more frequently Strengthening the capacity of local and state public health observed during the swimming season, which usually starts epidemiologists and environmental health specialists to on Memorial Day weekend (the last weekend in May) and detect and investigate outbreaks remains a priority at CDC ends Labor Day weekend (the first weekend in September).

and EPA. As part of that effort, CDC and EPA should part- However, swimming also occurs year-round in indoor venues ner with the states, CSTE, and the Association of Public and in states with more temperate climates. Outbreaks of ill-Health Laboratories to develop training materials and online ness by month (Figure 3) include two outbreaks that occurred resources that would be useful and easily accessible to local noticeably outside the summer months: one outbreak in a and state public health personnel. Although no federal regu- Florida pool in March and another outbreak in an indoor lation exists for monitoring private wells, developing educa- pool in Wisconsin in January.

tional materials targeted towards the general public, informing As during the previous reporting period, Cr. parvum them of ways to maintain the safety and water quality of their accounted for the largest percentage of outbreaks involving wells would be valuable. In addition, health messages regard- gastroenteritis (44.4%), followed by Es. coli O157:H7 ing the consumption of nonpotable water and appropriate (11.1%), NLV (8.3%), and Shigella (8.3%). An outbreak of point-of-use treatment should be developed and distributed G. intestinalis was also reported in 1999. The last reported to the public. recreational water outbreak of Giardia occurred in 1996. Out-breaks of Ca. jejuni, Es. coli O121:H19, and a mixed Sh. sonnei/

Outbreaks Associated with Recreational Cr. parvum outbreak were also reported for the first time to Water the surveillance system. Outbreaks of unknown etiology com-prised 16.7% of the recreational water outbreaks involving Of the 59 recreational WBDOs, those involving gastroen-gastroenteritis.

teritis were most frequently reported (n = 36). The 15 outbreaks reported in 1999 and 21 outbreaks reported in 2000 equal FIGURE 7. Number of waterborne-disease outbreaks associated with recreational water, by year and illness United States, 1989-2000 (n = 229)*

or surpass the number reported in 1998, which previously was the 40 highest number of outbreaks Other 35 involving recreational water-related Dermatitis Number of outbreaks gastroenteritis reported in one year 30 Meningoencephalitis since the inception of the surveil- Gastroenteritis 25 lance system. Together, the out-breaks involving gastroenteritis 20 reported during the 1999-2000 15 period are higher than the 18 out-breaks documented in the previous 10 reporting period (Figures 7 and 8).

5 Since 1989, the number of gastro-enteritis-related outbreaks has been 0 gradually increasing, and this 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 increase is statistically significant Year (p = 0.01).

• The total from previous reports has been corrected from n = 171 to n = 170.

24 MMWR November 22, 2002 FIGURE 8. Number of outbreaks involving gastroenteritis associated with recreational water, recirculation of water in the pool.

by year and illness United States, 1978-2000 (n = 146)

Because a low number of oocysts 25 might cause illness in a person, even ingestion of a limited amount of water can cause infection.

20 Although the number of Cr.

Number of outbreaks parvum outbreaks has been steadily increasing during 1990-2000, 15 multiple explanations could exist for the increase. The properties of the organism, coupled with the 10 popularity of swimming and the tendency of persons to aggregate in 5 larger water venues, increases the likelihood that swimming water can become contaminated and that 0 swimmers will ingest the water and 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 become infected. However, the Year increases in outbreaks could be explained by a higher awareness of Cr. parvum as a potential cause of Twenty-two (61.1%) of the 36 outbreaks of gastroenteritis illness among swimmers by the public health community and occurred in treated systems (i.e., pools) that would usually be the recreational water industry and, as a result, are more likely expected to be chlorinated or disinfected to prevent transmis- to be detected.

sion of infectious agents after unintentional ingestion. The majority of these Cr. parvum outbreak investigations However, the term treated system might pertain to systems not noted inadequate pool maintenance. Although low chlorine routinely treated, including wading pools, interactive levels are unlikely to have been the cause of the outbreaks, the fountains, and in one case, an untreated pool that was served frequent reporting of low chlorine levels in these outbreaks by a natural hot springs source. Multiple interrelated factors indicates a disturbing lack of awareness concerning the role can impede disinfection in treated venues, including an of chlorine and pH control as the major protective barrier increased bather load in a pool, high levels of organic mate- against infectious disease transmission in pools. Inadequate rial (e.g., fecal material or environmental or skin debris) and disinfectant levels in any pool increases the risk for transmis-ultraviolet light, all of which deplete chlorine residuals that sion of chlorine-sensitive pathogens (e.g., Es. coli O157:H7 usually maintain protection in the system. In certain outbreaks, or Shigella species) if an infected swimmer contaminates the fecal material was indicated on the report as a contributing pool. Pool operators and staff should be appropriately trained factor to the outbreak; the majority of fecal accidents were regarding the spread of recreational water illnesses and the attributed to young children who were in or near the water at critical role of pool maintenance (i.e., disinfection, pH con-the time the accident occurred. trol, and filtration) in preventing WBDOs.§§ Unlike previous years, a substantial number of different Fourteen outbreaks involving gastroenteritis after freshwa-bacterial and viral organisms were reported as causing gas- ter exposure were reported during 1999-2000, compared with trointestinal illness in these treated recreational water venues eight during 1997-1998. Es. coli O157:H7 accounted for the (Figure 9). Nevertheless, >66% of these outbreaks were most outbreaks of known etiology (three), followed by NLV attributed to Cr. parvum (Figure 4). Unlike other organisms, (two), Shigella species (two), Es. coli O121:H19 (one),

which are more susceptible to the levels of chlorine typically G. intestinalis (one), and Cr. parvum (one). Four outbreaks found in a pool, Cr. parvum is highly chlorine-resistant and were of unknown etiology. Certain outbreaks occurred in requires increased levels of chlorine and longer contact times beach areas that had substantial numbers of families bathing with chlorine for inactivation. Cr. parvum can survive for days and swimming in the water. Again, a common element noted in public health-mandated chlorine concentrations required for pools. In addition, its relatively limited size (4-6 µm) can §§ Guidelines for pool operators and other information related to recreational allow it to pass through particulate filtration systems during water illnesses is available at http://www.cdc.gov/healthyswimming.

Vol. 51 / SS-8 Surveillance Summaries 25 FIGURE 9. Number of outbreaks involving gastroenteritis associated with recreational water, inadequate disinfection of the by water type United States, 1985-2000 (n = 135) water. The higher temperatures commonly found in hot tubs deplete disinfectant levels at a more 25 Fresh rapid rate; hot tub operators should be encouraged to actively check Treated 20 and maintain adequate disinfectant Number of outbreaks levels. In addition to rashes, reports 15 have been received of other symp-toms. In Alaska, three of 29 per-sons reported nausea. In the two 10 Maine outbreaks, persons also reported headache, fatigue, and 5 other symptoms. The Colorado outbreak was notable for its severe 0 symptomatology and an extended 1985-1986 1987-1988 1989-1990 1991-1992 1993-1994 1995-1996 1997-1998 1999-2000 duration of illness. Extended and painful rashes associated with Year P. aeruginosa outbreaks are unusual but have been documented in these reports was the presence of diaper-aged children in (38,39). One report (39) also indicates that a substantial num-the water, diaper-changing on the beach, and even washing ber of children are being affected by these outbreaks. In the off young children in the water. One incident involved per- Colorado outbreak of P. aeruginosa, the persons affected were sons who swam in a lake that was had posted signs indicating primarily children, but no indication was provided that age that the lake was unsafe for swimming. was a risk factor for infection. More remarkable is the Reports of infants and children swimming when they have observed duration of illness. Certain persons reported chronic diarrhea is a problem common to both freshwater systems illness (i.e., rash, joint pain, abdominal pain, and chest pain) and treated venues. Although health communication messages that lasted >6 weeks. Using remote pool monitoring services have been targeted in the past for treated venues, similar in two of these outbreaks underscores the need for training messages should be provided to those who swimming in fresh- pool staff regarding the role of monitoring service and prompt water venues. EPA, as part of the Beaches Action Plan, is communication between service and pool operators when developing guidelines and information for users of freshwa- problems are detected.

¶¶ ters. Geothermal pools and hot springs should be examined Three outbreaks of dermatitis that occurred after persons closely. In one outbreak, pools in a complex were exempt from swam in fresh or marine water were presumed to be caused by public health regulation because they were naturally occur- an allergic reaction to the cercariae, the larval form of certain ring hot springs and mineral waters. Hot springs, which fea- nonhuman species of schistosomes. Cercarial dermatitis was ture high levels of minerals and elevated temperatures, are an identified problem in two of these lakes, and signs posted potentially ideal venues for microbial growth or contamina- by the health department regarding this problem were ignored tion. These springs and geothermal pools pose an increased by swimmers. The extent of the problem of cercarial dermati-risk to swimmers, compared with treated pools because of tis caused by freshwater exposure is unknown, although it their lack of disinfection and filtration. Improved consumer probably occurs more frequently than what is reported to the and staff education and supplementary treatment might be surveillance system. As schistosomes occur naturally in eco-necessary to prevent future outbreaks in these enclosed fresh- systems that bring snails and birds or aquatic mammals close water pools. together, a substantial number of freshwater lakes in the United Twelve of the 15 outbreaks of dermatitis were associated States might cause illness among swimmers. Swimmers should with hot tub or pool use. The majority of these reports of pay careful attention to where they swim, avoid shallow swim-dermatitis are associated with deficient maintenance and ming areas known to be appropriate snail habitats in lakes associated with cercarial dermatitis, and report any incidents

¶¶ Additional information is available at http://www.epa.gov/waterscience/ to their local health department to prevent further illnesses.

beaches.

26 MMWR November 22, 2002 The four deaths associated with primary amebic menin- Conclusion goencephalitis (PAM) reported during the 1999-2000 Data collected as part of the national WBDO surveillance period were all linked to freshwater exposure. Typically, these system are used to describe the epidemiology of waterborne cases are associated with swimming in freshwater bodies in diseases in the United States. Data regarding water systems the late summer months because N. fowleri, which has been and deficiencies implicated in these outbreaks are used to implicated in >90% of the cases reported to CDC, prolifer-assess whether regulations for water treatment and monitor-ate in warm, stagnant waters. Previous cases of PAM have ing of water quality are adequate to protect the publics health.

been reported from states with more temperate climates (e.g.,

Identification of the etiologic agents responsible for these California, Florida, and Texas) or from areas with hot springs.

outbreaks is also critical because new trends might necessitate The amoebas associated with PAM are believed to enter different interventions and changes in policies and resource through the nasal passage. Preventing forceful entry of water allotment.

up the nasal passages during jumping or diving by holding Surveillance for waterborne agents and outbreaks occurs ones nose or wearing nose plugs could reduce the risk for primarily at the local and state level. Local and state public infection.

health agencies need to detect and recognize WBDOs and Swimming in waters contaminated by animal urine was implement appropriate prevention and control measures.

the likely explanation for an outbreak of leptospirosis among Improved communication among local and state public health persons participating in an adventure race in Guam. Leptospira departments, regulatory agencies, water utilities, and recre-species can be found frequently in wild animal urine, and can ational water facilities would aide the detection and control be contracted through inhalation of aerosolized water or of outbreaks. Routine reporting or sharing of water-quality ingestion of water while swimming. Leptospirosis can also be data with the health department is recommended. Other acquired through abrasions. In this instance, the exposure was means of improving surveillance at the local, state, and fed-associated with immersion of persons heads in a body of water eral level could include the additional review and follow-up while they swam and swallowed water. Although outdoor of information gathered through other mechanisms (e.g.,

swimming is not necessarily dangerous, swimmers should be issuances of boil-water advisories or reports of illness associ-educated regarding the potential risks resulting from swim-ated with agents thought to be waterborne).

ming in areas that are not secured from wild animal use.

One repeated observation regarding outbreak data collected An increased level of bromine, which is used to disinfect as part of the WBDO system was that the timely collection of pools and hot tubs, caused certain cases of chemical keratitis.

clinical specimens and water samples for testing and com-Inadequate disinfection of a whirlpool resulted in an outbreak mencement of an environmental investigation would have of legionellosis among 20 persons who stayed at a motel. Safe resulted in an improved ability to detect the outbreaks etio-disinfection practices and appropriate pool maintenance pro-logic agent and the source of water contamination. However, tocols should be communicated to operators and managers the course of an investigation is influenced by the ability and of facilities that treat recreational water.

capacity of public health departments and laboratories to rec-ognize and investigate potential outbreaks of illness. Even when Outbreaks Associated with Occupational personnel are available to investigate a potential outbreak in a Exposures to Water timely manner, a common observation is that investigations Two outbreaks that do not fit into the previous categories cannot always be completed thoroughly. WBDO outbreak were reported to CDC by Minnesota and Hawaii. Outbreaks investigations typically require input from different disciplines, associated with exposure to aerosolized water have previously including infectious disease epidemiology, environmental epi-occurred but have not been reported to the WBDO surveil- demiology, clinical medicine, sanitation, water engineering, lance system (40,41). These outbreaks are discussed in this and microbiology. Either further cross-training of existing report to demonstrate that water exposures are not limited to personnel needs to be implemented or additional personnel ingestion and contact (e.g., through swimming), and these and resources need to be made available or linked to those outbreaks are preventable. Using barrier masks to prevent who typically investigate reports of WBDOs.

inhalation of aerosolized water or disinfection of water that is State health departments can request epidemiologic assis-not being used for drinking or swimming purposes could have tance and laboratory testing from CDC to investigate prevented the respiratory illnesses associated with these two WBDOs. CDC and EPA can be consulted regarding engi-outbreaks. neering and environmental aspects of drinking water and rec-reational water treatment and regarding collection of

Vol. 51 / SS-8 Surveillance Summaries 27 large-volume water samples to identify pathogenic viruses and 3. Levy DA, Bens MS, Craun GF, Calderon RL, Herwaldt BL. Surveil-parasites, which require special protocols for their recovery. lance for waterborne-disease outbreaksUnited States, 1995-1996.

In: CDC Surveillance Summaries, December 11, 1998. MMWR Requests for tests for viral organisms should be made to CDCs 1998;47(No. SS-5):1-34.

Viral Gastroenteritis Section, Respiratory and Enterovirus 4. Kramer MH, Herwaldt BL, Craun GF, Calderon RL, Juranek DD.

Branch, Division of Viral and Rickettsial Diseases, National Surveillance for waterborne-disease outbreaksUnited States, 1993-Center for Infectious Diseases (NCID), at 404-639-3577. 1994. In: CDC Surveillance Summaries, April 12, 1996. MMWR Requests for tests for parasites should be made to CDCs 1996;45(No. SS-1):1-33.

Division of Parasitic Diseases, NCID, at 770-488-7760. 5. Moore AC, Herwaldt BL, Craun GF, Calderon RL, Highsmith AK, Juranek DD. Surveillance for waterborne disease outbreaksUnited Additional information is available from States, 1991-1992. In: CDC Surveillance Summaries, November 19,

• EPAs Safe Drinking Water Hotline at 800-426-4791, on 1993. MMWR 1993:42(No. SS-5):1-22.

the Internet at http://www.epa.gov/safewater, or by e-mail 6. Herwaldt BL, Craun GF, Stokes SL, Juranek DD. Waterborne-disease at hotline-sdwa@epa.gov; outbreaks, 1989-1990. In: CDC Surveillance Summaries, December

• CDCs NCID website at http://www.cdc.gov/ncidod; 1991. MMWR 1991;40(No. SS-3):1-21.

7. Environmental Protection Agency. 40 CFR Part 141. Water programs:

• CDCs Healthy Swimming website at http://ww.cdc.gov/

national interim primary drinking water regulations. Federal Register healthyswimming; includes recreational water health com- 1975;40:59566-74.

munication materials for the general public and pool 8. Pontius FW, Roberson JA. Current regulatory agenda: an update. Jour-maintenance staff (e.g., information regarding disinfec- nal of the American Water Works Association 1994;86:54-63.

tion, guidelines on response to fecal accidents [42], fact 9. Pontius FW. Implementing the 1996 SDWA amendments. Journal of sheets concerning recreational water illnesses), and an the American Water Works Association 1997;89:18-36.

10. Environmental Protection Agency. Announcement of the drinking outbreak investigation toolkit that can be used by public water contaminant candidate list; notice. Federal Register health professionals; 1998;63:10274-87.

• CDCs Voice and Fax Information System, 888-232-3228 11. Environmental Protection Agency. 40 CFR Parts 141 and 142. Drink-(voice) or 888-232-3299 (fax). Choose cryptosporidiosis ing water; national primary drinking water regulations; filtration, in the disease category; and disinfection; turbidity, Giardia lamblia, viruses, Legionella, and het-erotrophic bacteria; final rule. Federal Register 1989;54:27486-541.

• for reporting WBDOs, CDCs Division of Parasitic Dis-

12. Environmental Protection Agency. 40 CFR Parts 141 and 142. Drink-eases, NCID, at 770-488-7760 or by fax at 770-488-7761. ing water; national primary drinking water regulations; total coliforms (including fecal coliforms and E. coli); final rule. Federal Register Acknowledgments 1989;54:27544-68.

The authors thank the following persons for their contributions 13. Environmental Protection Agency. 40 CFR Parts 141 and 142. Drink-to this report: state waterborne-disease surveillance coordinators; ing water; national primary drinking water regulations; total coliforms; state epidemiologists; state drinking water administrators; Susan corrections and technical amendments; final rule. Federal Register Shaw, M.S., M.P.H., Office of Ground Water and Drinking Water, 1990;55:25064-5.

14. Environmental Protection Agency. 40 CFR Parts 9, 141, and 142.

EPA; Robert Tauxe, M.D., Division of Bacterial and Mycotic National primary drinking water regulations: interim enhanced sur-Diseases, NCID, CDC; Roger Glass, M.D., Ph.D., Steve Monroe, face water treatment; final rule. Federal Register 1998;63:69477-521.

Ph.D., and Marc-Alain Widdowson, M.A., Vet.M.B., M.Sc., 15. Environmental Protection Agency. 40 CFR Parts 9, 141, and 142.

Division of Viral and Rickettsial Diseases, NCID, CDC; Matthew National primary drinking water regulations: long term 1 enhanced Arduino, Dr.P.H. and Joe Carpenter, Division of Healthcare Quality surface water treatment rule; final rule. Federal Register 2002;67:1812-44.

and Promotion, NCID, CDC; Mark Eberhard, Ph.D., Mary 16. Environmental Protection Agency. 40 CFR Parts 9, 141, and 142.

Bartlett, Ali Khan, M.D., James Maguire, M.D., and Dennis National primary drinking water regulations: long term 1 enhanced Juranek, D.V.M., Division of Parasitic Diseases, NCID, CDC; and surface water treatment and filter backwash rule; proposed rule. Fed-Lorraine Backer, Ph.D., Division of Environmental Hazards and eral Register 2000;67:19046-150.

Health Effects, National Center for Environmental Health, CDC. 17. Environmental Protection Agency. 40 CFR Parts 9, 141, and 142, National primary drinking water regulations: filter backwash recycling References rule; final rule. Federal Register 2001;66:31085-105.

1. Craun GF, ed. Waterborne diseases in the United States. Boca Raton, 18. Environmental Protection Agency. 40 CFR Parts 141 and 141.

FL: CRC Press, Inc., 1986. National primary drinking water regulations: ground water rule; pro-

2. Barwick RS, Levy DA, Craun GF, Beach MJ, Calderon RL. Surveil- posed rules. Federal Register 2000;65:30193-274.

lance for waterborne disease outbreaksUnited States, 1997-1998. 19. Environmental Protection Agency. 40 CFR Parts 9, 144, 145, and In: CDC Surveillance Summaries, May 26, 2000. MMWR 146. Underground injection control regulations for class V injection 2000;49(No. SS-4):1-35. wells, revision; final rule. Federal Register 1999;64:68545-73.

28 MMWR November 22, 2002

20. Environmental Protection Agency. 40 CFR Part 141. National pri- 31. Jones JL, Lopez A, Wahlquist SP, Nadle J, Wilson M. Survey of clini-mary drinking water regulations: monitoring requirements for public cal laboratory practices, parasitic diseases. Clinical Infectious Diseases drinking water supplies; final rule. Federal Register 1996;61:24353-88. (in press).

21. Environmental Protection Agency, 40 CFR Parts 9, 141, and 142. 32. US General Accounting Office. Drinking water: information on the National primary drinking water regulations for lead and copper. quality of water found at community water systems and private wells.

Final rule. Federal Register 2000;65:1949-2015. Washington, DC: US General Accounting Office, 1997. GAO publi-

22. US Environmental Protection Agency, Office of Water. Ambient cation no. GAO/RCED-97-123.

water quality criteria for bacteria1986. Cincinnati, OH: National 33. CDC. Foodborne & waterborne disease outbreaks [Annual summary Service Center for Environmental Publications, 1986. EPA publica- 1973]. Atlanta, GA: US Department of Health, Education, and Wel-tion no. 440584002. fare, CDC, 1974. Publication no. 76-8185.

23. Dufour AP. Health effects criteria for fresh recreational waters. 34. CDC. Water-related outbreaks [Annual summary 1980]. Atlanta, GA:

Research Triangle Park, NC: US Environmental Protection Agency, US Department of Health and Human Services, CDC, 1981. Publi-Office of Research and Development, Health Effects Research Labo- cation no. 82-8385.

ratory, 1984; EPA publication no. 600184004. 35. CDC. Norwalk-like viruses: public health consequences and out-

24. US Environmental Protection Agency, Office of Water. Factoids: drink- break management. MMWR 2001;50(No. RR-9):1-18.

ing water and ground water statistics for 2000. Washington, DC: US 36. Calderon RL, Mood EW, Dufour AP. Health effects of swimmers and Environmental Protection Agency, Office of Water, 2001. EPA publi- nonpoint sources of contaminated water. International Journal of cation no. 816K01004. Available at http://www.epa.gov/cgi-bin/ Environmental Health Research 1991;1:21-31.

claritgw. 37. Seyfried PL, Tobin RS, Brown NE, Ness PF. Prospective study of

25. US Environmental Protection Agency. EPA safe drinking water infor- swimming-related illness. I. Swimming-associated health risk. Am J mation system factoids: FY 1999 inventory data. Washington, DC: Public Health 1985;75:1068-70.

US Environmental Protection Agency, 2002. Available at http:// 38. Berrouane YF, McNutt LA, Buschelman BJ, et al. Outbreak of severe www.epa.gov/safewater/data/99factoids.pdf. Pseudomonas aeruginosa infections caused by a contaminated drain in a

26. Anonymous. Blastocystis hominis: a new pathogen in day-care centers? whirlpool bathtub. Clin Infect Dis 2000;31:1331-7.

Can Commun Dis Rep 2001;27:76-84. 39. Fiorillo LM, Zucker M, Sawyer D, Lin AN. Pseudomonas hot-foot

27. CDC. Protracted outbreaks of cryptosporidiosis associated with swim- syndrome. N Engl J Med 2001;345:335-8.

ming pool useOhio and Nebraska, 2000. MMWR 2000;50:406-10. 40. CDC. Sustained transmission of nosocomial Legionnaires disease

28. CDC. Outbreak of gastroenteritis associated with an interactive water Arizona and Ohio. MMWR 1997;46:416-21.

fountain at a beachside parkFlorida, 1999. MMWR 2000;49:565-8. 41. CDC. Legionnaires disease associated with a whirlpool spa display

29. CDC. Pseudomonas dermatitis/folliculitis associated with pools and hot Virginia, September-October, 1996. MMWR 1997;46:83-6.

tubsColorado and Maine, 1999-2000. MMWR 2000;49:1087-91. 42. CDC. Responding to fecal accidents in disinfected swimming venues

30. CDC. Methemoglobinemia attributable to nitrite contamination of [Notice to readers]. MMWR 2001;50:416-7.

potable water through boiler fluid additivesNew Jersey, 1992 and 1996. MMWR 1997;46):202-4.

Vol. 51 / SS-8 Surveillance Summaries 29 Appendix A Selected Case Descriptions of Outbreaks Associated with Drinking Water State where Suspected Number Outbreak outbreak or confirmed of persons date(s) occurred etiologic agent affected Case description Parasites January Florida G. intestinalis 2 Persons became ill after drinking well water. Environmental 1999 inspection revealed that pigs were maintained 15 feet from the well. Fecal coliforms were not detected in the water samples, but data indicated inadequate levels of chlorine.

June Minnesota G. intestinalis 12 Initial water testing confirmed the presence of fecal coliforms 2000 in the drinking water where persons were working in a wild-life refuge. Further investigation of the well revealed that a drain line from the well pit had been connected to the sewer line, allowing contamination of the water by raw sewage.

July New G. intestinalis 4 Persons drinking water provided by others during a rafting 2000 Mexico trip became ill. The water was either incorrectly purified river water or river water that was untreated before consumption.

August Colorado G. intestinalis 27 A resort experienced multiple failures in the pumping mecha-2000 nism and filtration system during drinking water treatment that resulted in untreated river water entering the drinking water supply. The water supply included a mix of spring water and conventionally treated river water that usually was pumped into and held in a water storage tank. A failure in the primary system was reported; the backup pump was used, and multiple filtration cartridges were later determined to be defective. Although water in the holding tank was usu-ally chlorinated, because of the demands placed on the sys-tem during the summer season, chlorine might not have had adequate time to inactivate Giardia cysts. Giardia was iden-tified in samples from 5 of the 27 affected persons and in the finished water in the holding tank.

September Florida G. intestinalis 2 Persons became ill after drinking water contaminated from 2000 a cross-connection between watering troughs for exotic ani-mals and the drinking water system, which created an opportunity for back-siphonage of the trough water.

Coliforms were present in the water, but fecal coliforms were not detected in the water samples. Data indicated inadequate levels of chlorine.

30 MMWR November 22, 2002 State where Suspected Number Outbreak outbreak or confirmed of persons date(s) occurred etiologic agent affected Case description September New G. intestinalis 5 Household members used a private well where the water-2000 Hampshire filtration system had inadvertently been shut off during a 1-month period by construction workers. The water was not otherwise treated (e.g., chlorination).

December Florida Cryptosporidium 5 An outbreak occurred among persons who had consumed 2000 parvum water from a community system well. The well had a series of breaks and repairs, including a break reported days before initial reports of illness. Investigation revealed low levels of chlorine (0.03 mg/L) in the tap water.

B acteria June Missouri Salmonella 124 Persons consumed water from a community well that was 1999 Typhimurium routinely disinfected by chlorine, but investigators reported that chlorine had been inadequate at the time of the out-break. Seventeen persons were hospitalized.

August New York Ca. jejuni and 781 A fairground was served primarily by chlorinated water, but 1999 Es. coli certain vendors used water from an unchlorinated shallow O157:H7 well to make beverages and ice. Epidemiologic data associ-ated illness (including 71 hospitalizations and two deaths) with consumption of water from that well. Drought condi-tions had lowered the water table and heavy rains occurring during the fair might have contributed to contamination of the well by surface water. Dye tests revealed a cross-connection between a dormitory septic system and a well on the fair-grounds, but did not confirm the presence of a hydraulic connection between a nearby manure storage site and the implicated well. However, testing had occurred 2 months after the outbreak and after Hurricane Floyd had passed through the area, raising the water table. Therefore, the pos-sibility of a previous connection during the time of the fair could not be ruled out. Although stool sampling of ill per-sons indicated the presence of both Es. coli O157:H7 and Ca. jejuni, only Es. coli O157:H7 was identified in water samples.

November Texas Es. coli 22 A city well water supply was implicated because all 22 per-1999 O157:H7 sons received municipal water and had no other common exposures. Water quality data indicated that the routinely chlorinated water was inadequately chlorinated for weeks during the beginning of the outbreak. Subsequent to the outbreak, a new chlorination system was installed for the water supply.

Vol. 51 / SS-8 Surveillance Summaries 31 State where Suspected Number Outbreak outbreak or confirmed of persons date(s) occurred etiologic agent affected Case description April Idaho Es. coli 4 Children reportedly drank water from the canal in their back-2000 O157:H7 yard while playing. Water from the affected household and from the canal located near the home was tested. Coliforms were not present in the homes tap water, but 100 fecal coliforms/100 mL were present in the canal water. Agricul-tural runoff might have contributed to contamination of the canal water.

June Idaho Ca. jejuni 15 Contaminated groundwater was implicated in this outbreak 2000 among persons attending a geology summer camp. The source of water was a spring box, a device used to collect water, that was incorrectly constructed, potentially allowing contamination by surface water and agricultural runoff.

Water samples from the spring box were positive for total coliforms but negative for fecal coliforms. Breaks in the line running from the spring to the camp were noted. Water in the campsite hose tested positive for both total and fecal coliforms.

July California Es. coli 5 Campers ingested filtered creek water both directly and in 2000 O157:H7 food reconstituted with creek water. They reported defecat-ing near the campsite, and they had defecated in the water while swimming. In addition, deer droppings, a possible source of Es. coli O157:H7, were found near the creek.

Although campers also reported swimming in the creek, swimming could not be implicated as an exposure. Illness was confirmed through laboratory tests.

August Utah Ca. jejuni and 102 Participants at a football camp drank from an irrigation sys-2000 Es. coli tem tap that was not intended for human consumption. No O157:H7 hospitalizations or deaths were reported. Ca. jejuni was the most commonly isolated organism among submitted stool specimens and was also isolated from water samples; Es. coli O157:H7 and O111 were also isolated from stool speci-mens, but shiga toxins for these organisms were not found in water samples.

August Ohio Es. coli 29 Persons who had attended a county fair were confirmed for 2000 O157:H7 Es. coli O157:H7 by stool culture. Additional cases of gastrointestinal illness were identified in the community; however, either they could not be laboratory-confirmed as Es. coli O157:H7 or they were probably secondary cases.

Nine of the 29 culture-confirmed persons were hospitalized.

Other potential risk factors (e.g., a particular food item, beverage, or direct exposure to animals) could not be impli-cated. However, consumption of food and beverages sold within a particular area of the fairgrounds was a substantial

32 MMWR November 22, 2002 State where Suspected Number Outbreak outbreak or confirmed of persons date(s) occurred etiologic agent affected Case description exposure. Investigation revealed that the branch of the dis-tribution system supplying the vendors in the implicated area also supplied an animal show barn. Spigots connecting hoses to both the animal areas and the concession stands were not equipped with backflow-prevention devices. There-fore, possible back-siphonage might have resulted in con-tamination of the water supply used for human consumption.

Although, water samples taken from the municipal system supplying the fair and water taken after the closure of the fair were negative for coliforms, one spigot serving the implicated vendor area tested positive for coliforms, but nega-tive for Es. coli.

April- 10 South- Salmonella 84 The outbreak was detected by CDCs Surveillance Outbreak August eastern Bareilly Detection Algorithm. Forty-three of the 84 persons were 2000 states: enrolled in a case-control study with 76 matched controls.

Alabama, The outbreak was associated with drinking bottled water Arkansas, and water from private wells or springs. Twenty-five persons Georgia, drank water from private wells; Sa. Bareilly was recovered Kentucky, from one patients well. The majority of source wells were Louisiana, located in a limestone area in the southeastern United States; Missouri, wells drilled in limestone are subject to contamination from North different sources, including surface water. Among eight per-Carolina, sons who did not drink well or spring water in the week Tennessee, before illness, the consumption of either of two brands of Virginia, water bottled by one company was epidemiologically impli-and West cated as a risk factor for illness. A Food and Drug Adminis-Virginia tration (FDA) traceback investigation of the two implicated brands (spring water and infant water, water marketed spe-cifically for consumption by infants) linked the bottled water to one facility in the southeastern United States. An investigation of the facilities determined that the two impli-cated products had different sources, but water flowed through the same lines before bottling. Infant water was derived from a municipal source and passed through mul-tiple treatment steps at the facility (separate from the spring water), and then stored before the bottling process. The bot-tling process included two filtration steps, ultraviolet irra-diation and an nonautomated ozonization step. The bottled spring water came from water transported from a spring in north Georgia. After the spring water was fluoridated at the plant, the spring water shared the same treatment and bot-tling path as the infant water. Logs from daily in-house test-ing did not indicate coliforms; however, raw spring water samples tested by an independent laboratory in June were positive for total coliforms. Environmental samples collected

Vol. 51 / SS-8 Surveillance Summaries 33 State where Suspected Number Outbreak outbreak or confirmed of persons date(s) occurred etiologic agent affected Case description by CDC and FDA during the investigation did not yield Sa.

Bareilly, but one of the 105 samples taken from a lot of bottled water produced in September was positive for Es. coli, and two of 105 samples were positive for other coliforms. Treat-ment provided should have been adequate to prevent con-tamination of the bottled water, but finding coliforms implies that interruption of treatment or suboptimal operation might have allowed fecal contamination of the bottled water.

Inspection of the treatment process did not reveal any obvi-ous deficiencies; therefore, the source of the contamination remains unknown.

Viruses July New Small round- 70 At a scout camp, a spring box influenced by surface water 1999 Mexico structured virus was the key factor in the outbreak. Small round-structured virus was observed in stool samples from three of 22 speci-mens submitted. An association with consumption of water or any drinks prepared with water was observed, although the numbers were not statistically significant. No other food items were associated with the outbreak. Assessment of the drinking water system indicated that the spring box was situ-ated at a lower elevation than the latrines and other build-ings with individual septic systems. Water samples taken from the well and distribution system were positive for fecal coliforms.

June West Norwalk-like 123 Multiple cohorts of camp attendees reported gastrointesti-2000 Virginia virus nal illness occurring during a multiweek period. Consump-tion of food items and a history of swimming were ruled out as vehicles of transmission. The epidemiologic informa-tion could not statistically implicate water as the vehicle of exposure; however, the environmental investigation deter-mined that two wells that provided the drinking water were located near a lagoon and one well was visibly contaminated with sewage. Fecal coliforms were isolated from both wells.

June Kansas Norwalk-like 86 Persons who attended two different parties at a reception 2000 virus hall on 2 separate days became ill. The facilitys well water was the only common source between the two groups, and the water was not filtered or treated. Norwalk-like virus was isolated from individual stool specimens from both groups.

Inspection of the well serving the facility indicated improper well construction, and sampled water tested positive for coliforms.

34 MMWR November 22, 2002 State where Suspected Number Outbreak outbreak or confirmed of persons date(s) occurred etiologic agent affected Case description July California Norwalk-like 147 Untreated well water at a camp facility was implicated. Test-2000 virus ing identified fecal and total coliforms in one of the wells and in the areas of the distribution system serving the recre-ational vehicles and tent areas in the campground.

Chemicals April Wisconsin Nitrate 1 An infant was fed formula prepared with boiled tap water 1999 from a private well for >2 days after the household ran out of bottled water normally used to prepare infant formula.

The infant became ill, was hospitalized, and was treated to reverse increased methemoglobin levels. Subsequent testing of the tap water revealed the presence of fecal coliforms and nitrate. Inorganic tests for metals were negative.

November New Jersey Sodium 2 Lack of a check valve at a community well resulted in spill-1999 hydroxide age of approximately 200 gallons of sodium hydroxide into a 110,000-gallon well during a multihour period. Although only two persons reported illness, a cohort estimated at 100-1,000 persons might have been affected. The first person to report an adverse effect suffered a first-degree burn from showering in the contaminated water. The second person suffered cramps, presumably from ingestion of the water.

Unidentified etiologic agents January Florida Unknown 4 A well was improperly constructed, and water was not rou-1999 tinely disinfected or filtered. Persons interviewed indicated water had been turbid before the outbreak. Two water samples indicated coliforms were present.

March Florida Unknown 3 An outbreak occurred at an apartment building that usually 1999 received municipal water from a well disinfected with chlo-rine. The outbreak involved a cross-connection within the apartment complex to an irrigation well. Total coliforms were found in samples taken after the outbreak had occurred, but fecal coliforms were not found. A stool specimen from one of the affected persons tested negative for bacterial and parasitic enteric pathogens.

March Florida Unknown 6 Outbreak involved a cross-connection to an irrigation well 1999 from a municipal surface water system. The municipal water system was routinely treated and filtered. The irriga-tion well in question was in unsatisfactory sanitary condi-tion and was located 25 feet from a commercial septic system and 10 feet from a garbage container. No clinical testing was performed.

Vol. 51 / SS-8 Surveillance Summaries 35 State where Suspected Number Outbreak outbreak or confirmed of persons date(s) occurred etiologic agent affected Case description May Florida Unknown 3 An outbreak occurred after persons in the household began 1999 drinking well water. Well was located 10 feet from a chicken coop. Duration of illness was approximately 60 days. No clinical results were provided. Chlorine residuals were not detected in water samples.

July California Unknown 31 An outbreak occurred among persons who visited a camp 1999 cabin. Diarrhea, vomiting, and cramps were reported.

Median incubation period was 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />, and median dura-tion of illness was 24 days. The well used for drinking water was not routinely disinfected or filtered. Water samples tested positive for fecal coliforms.

July Washington Unknown 46 Persons had consumed untreated creek water, but tests for 1999 Cryptosporidium parvum, G. intestinalis, Shigella species, Salmonella, Ca. jejuni, Es. coli O157:H7, and Yersinia enterocolitica performed on stool specimens were all nega-tive. One stool specimen was positive for Blastocystis hominis and another for Endolimax nana. The median incubation period before onset of illness among the 46 persons was 34 hours3.935185e-4 days <br />0.00944 hours <br />5.621693e-5 weeks <br />1.2937e-5 months <br /> (range: 10-95.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />), and their median duration of illness was 38 days (range: 2-192 days).

August Washington Unknown 68 Participants of a soccer match reported gastrointestinal ill-1999 ness. The most commonly reported symptoms were diar-rhea (72%), vomiting (78%), and nausea (84%). No one food or beverage item was implicated, but drinking any bev-erage prepared on site was a substantial risk for illness.

Although contaminants were not identified, the soccer match took place at a polo field that reportedly had horse manure at the site. Coliforms were found in the water tested, but samples taken from the wellhead and the clubhouse were negative for Es. coli. Stool specimens tested negative for enteric bacteria and parasites.

April Florida Unknown 71 Clinical testing was attempted when the outbreak occurred 2000 among persons who drank water from untreated wells. A sinkhole in a local lake had developed, allowing water to be directed into the aquifer. Before the sinkhole was plugged, water from 78 of the surrounding wells was tested. Nine-teen of the wells were negative for coliforms; 33 were posi-tive for total coliforms; and 26 were positive for fecal coliforms. Cryptosporidium was found in one of the wells.

After the sinkhole was plugged, 58 wells were negative for coliforms; 21 wells were positive for total coliforms; and none

36 MMWR November 22, 2002 State where Suspected Number Outbreak outbreak or confirmed of persons date(s) occurred etiologic agent affected Case description was positive for fecal coliforms. No enteric pathogens were identified in one clinical specimen that was submitted. No viral testing of either clinical specimens or water samples was reported.

June Florida Unknown 2 An outbreak was associated with consumption of untreated 2000 well water; two persons reported diarrhea that lasted approxi-mately 60 days. The stool specimen collected from one per-son tested negative for Giardia but positive for B. hominis.

The well was located in an area that had experienced flood-ing and heavy rainfall before the outbreak. The affected per-sons indicated that dark residue and air bubbles had appeared in their water around the time of illness. Environmental assessment confirmed that the well was in unsatisfactory sani-tary condition and noted the presence of grit and air bubbles in the water. Total coliforms were observed in water samples taken after disinfection of the well, although fecal coliforms were absent. The conclusion of the investigators was that the well might have been contaminated by surface water.

September California Unknown 63 An outbreak occurred after football players consumed water 2000 after a game from an irrigation system coupler located on the playing field. The water was not intended for human consumption and was subject to contamination from back-siphonage of surface water that collected around the sprin-kler heads and control valves. Both total and fecal coliforms

>16,000 most probable number/100 mL were isolated from the irrigation source water. No hospitalizations or deaths resulted. Because the median incubation period was 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> (range: 5-31 hours) and the median duration of illness was also 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> (range: 1 hour-3 days), the outbreak was probably of viral or bacterial origin.

Vol. 51 / SS-8 Surveillance Summaries 37 Appendix B Selected Case Descriptions of Outbreaks Associated with Recreational Water State where Suspected Number Outbreak outbreak or confirmed of persons date(s) occurred etiologic agent affected Case description Parasites July and Wisconsin Cryptosporidium 10 and 6, Both outbreaks were linked to private swimming pools; chlo-August and Florida parvum respectively rine levels had not been tested in either of these outbreaks.

1999 July Minnesota Cr. parvum 10 Chlorine levels were inadequate at a trailer park swimming 1999 pool because the chlorinator needed repair.

July Massachusetts Giardia 18 The common exposure for persons affected was swimming 1999 intestinalis in a local pond. Giardiasis was laboratory-confirmed.

Although the source water was not tested for G. intestinalis, water tests for the pond indicated high levels of total coliforms.

June Ohio Cr. parvum 700 A swimming pool at a private club was implicated. An 2000 investigation determined that additional risk factors for cryptosporidiosis (e.g., drinking municipal water, eating unpasteurized food, or visiting a local zoo) were not statisti-cally significant. However, the investigation did determine that a substantial risk factor for illness was oral contact with water. The pool consisted of a zero-entry-level pool (i.e., a simulated beach entry) that was connected to both a baby wading pool and adult pool with a water slide. Review of pool records indicated that on >2 days, the chlorine residual was inadequate, including on days when the bather load was high and air temperature hot. In addition, multiple fecal accidents had been reported. Testing of water samples at CDC, using Environmental Protection Agency (EPA)

Method 1622, demonstrated the presence of Cr. parvum oocysts in the water.*

June Nebraska Cr. parvum 225 Surveillance conducted by a county health department 2000 revealed a cluster of laboratory-confirmed cases of cryptosporidiosis, primarily among persons who were mem-bers of two private clubs with swimming facilities (clubs A and B). Case-control studies were conducted among mem-bers of the two clubs. Swimming and being splashed with pool water at club A was statistically associated with illness.

Club A had four pools: outdoor adult and baby pools and

• CDC. Protracted outbreaks of cryptosporidiosis associated with swimming pool useOhio and Nebraska, 2000. MMWR 2000;50:406-10.

38 MMWR November 22, 2002 State where Suspected Number Outbreak outbreak or confirmed of persons date(s) occurred etiologic agent affected Case description indoor adult and baby pools. The two outdoor pools shared a filtration system, whereas the two indoor pools had sepa-rate filtration systems but were connected by a pipe used to equilibrate the water level of the pools. Fecal accidents had occurred before the outbreak. Review of the pool mainte-nance logs revealed a 2-day period when the baby pool had not been chlorinated.

June Georgia Cr. parvum 36 Persons reported illness after attending a private pool party 2000 where, in addition to swimming in a neighborhood pool, party attendees also swam in an inflatable pool. Inadequate chlorine levels were detected in the neighborhood pool. No treatment was documented for the inflatable pool.

July Minnesota Cr. parvum 7 Inadequate chlorination and filtration were documented in 2000 this outbreak among attendees at a day camp.

July Minnesota Cr. parvum 6 Information from interviews implicated a hotel pool in an 2000 outbreak of cryptosporidiosis among members of a baseball team. Both whirlpool and pool samples were negative for Cr. parvum.

July Florida Cr. parvum 3 Inadequate water quality and low chlorine levels were docu-2000 mented in an outbreak at an apartment complex. Fecal material had been visible in the pool at the apartment complex.

July South Cr. parvum 26 Water-quality data from a neighborhood pool implicated in 2000 Carolina an outbreak of cryptosporidiosis were inconclusive. Coliform tests were negative; however, these tests were conducted after the outbreak had occurred and the pool had been treated with high levels of chlorine. Chlorine levels at the time of and after the outbreak were unreported. Certain factors might have contributed to the outbreak, including a history of symptomatic children swimming in the pool before the out-break and increased rain activity that might have decreased the water quality by diluting chlorine levels or facilitated runoff into the pool.

July Minnesota Cr. parvum 220 An outbreak of cryptosporidiosis occurred at a public swim-2000 ming beach. A total of 220 persons reporting illness were asked questions regarding swimming exposure and food con-sumed while at the beach. Exposure to the water (i.e., get-ting the head wet while swimming) was associated with CDC. Protracted outbreaks of cryptosporidiosis associated with swimming pool useOhio and Nebraska, 2000. MMWR 2000;50:406-10.

Vol. 51 / SS-8 Surveillance Summaries 39 State where Suspected Number Outbreak outbreak or confirmed of persons date(s) occurred etiologic agent affected Case description illness. The inspection conducted by the county health department did not identify any plumbing deficiencies at the beach clubhouse. Reports of construction on a sewer line were investigated, but the investigators determined that the source of contamination was persons washing babies in the lake while changing diapers on the beach.

August Colorado Cr. parvum 112 The epidemiologic investigation implicated swimming in 2000 the main pool or having contact with the water as a risk factor for illness among the persons who attended a private party at a municipal pool. Although the main pool shared a silica-sand filtration system with a smaller, adjacent wading pool, contact with the wading pool was a substantial risk factor.

August Florida Cr. parvum 19 Exposure to the pool at a resort was a substantial risk factor 2000 for infection. In addition, infection was associated with longer times of exposure to the resort pool water. Complaints of cloudy water and diapered children swimming in the pool were reported.

August Florida Cr. parvum 5 This outbreak was linked to the outbreak that occurred in 2000 Ohio in June 2000. A family from Ohio who were members of the implicated swim club vacationed in Florida. While in a pool in Florida, the ill infant had two fecal accidents.

August Florida Cr. parvum 5 Illness was associated with swimming in a pool, and one 2000 infected child reportedly swam while ill. Water quality information was not available.

August Minnesota Cr. parvum 4 This outbreak was detected through routine surveillance. All 2000 four were interviewed by the state health department and queried about history of illness and exposures to water, food, animals, and child care centers. All four persons reported swimming in the same municipal swimming pool. No cor-rective action was taken. The outbreak was reported after the pool had closed at the end of the swimming season.

Bacteria August Wisconsin Es. coli 5 Swimmers who had visited the same beach became ill. The 1999 O157:H7 popular beach featured a shallow, dammed area that was used for wading. Total and fecal coliforms were detected in water samples collected before and during the outbreak, although the levels did not exceed regulatory levels for microbiologic quality of water. One sample that was tested for Es. coli O157:H7 was negative.

40 MMWR November 22, 2002 State where Suspected Number Outbreak outbreak or confirmed of persons date(s) occurred etiologic agent affected Case description August Washington Es. O157:H7 36 A freshwater outbreak occurred among visitors to a state park.

1999 Swimming in the lake, getting lake water in the mouth, and swallowing water from the lake were all associated with ill-ness. In addition, Es. coli O157:H7 was laboratory-confirmed among the affected persons, including seven hospitalizations, and isolated from samples of sediment from the lake as well as the water.

May Wisconsin Legionella 20 A whirlpool was implicated in a serologically confirmed 2000 pneumophila outbreak of Pontiac fever among 20 persons who visited a hotel. L. pneumophila group 5 bacteria were isolated from the whirlpool but not from the swimming pool. Additional samples collected from sand filters, a water heater, and a shower also were negative for L. pneumophila. Disinfectant levels for both the whirlpool and pool were inadequate at the time of the outbreak. Analysis of responses to question-naires administered to motel visitors revealed that only whirl-pool and swimming pool exposures were substantial risk factors for illness. No other possible common exposures dem-onstrated associations with illness. The investigation con-cluded from both the epidemiologic and environmental data that the whirlpool was the likely vehicle of transmission.

July Connecticut Es. coli 11 The outbreak occurred among persons who lived in or vis-2000 O157:H7 ited the same community during the last 2 weeks of July. An environmental assessment of both well water and lake water was conducted. The well water system was found to be in compliance with drinking water standards, and samples from two wells and ice made from this water were all negative for total coliforms. Lake water sampled after the outbreak was negative for Es. coli and shiga toxins but was positive for total and fecal coliforms. No other environmental problem at the lake was identified, and animal feces collected from local wildlife tested negative for Es. coli O157 or any other shiga toxin-producing isolates. The environmental assess-ment, combined with epidemiologic evidence, implicated swimming at the lake and swallowing water as substantial risk factors for illness. During this same time period, a tod-dler with severe diarrhea reportedly had swum in the water during a 1-week period while ill.

July Minnesota Shigella sonnei 15 Fifteen persons reported illness after swimming at a fresh-2000 water lake. Of these, 13 tested positive for Sh. sonnei, one for Cr. parvum, and one for both organisms.

Vol. 51 / SS-8 Surveillance Summaries 41 State where Suspected Number Outbreak outbreak or confirmed of persons date(s) occurred etiologic agent affected Case description August Minnesota Sh. sonnei 25 The only shared exposure among the affected persons was 2000 swimming in a particular area of a public beach. Seventeen persons from >5 families who separately attended the beach on the same day became ill; these persons did not all attend the beach together and did not consume the same food prod-ucts at the beach. Eight other persons, who were identified through routine surveillance, also reported swimming at the beach on the date in question. Water samples collected

>2 days after the implicated day did not test positive for Shigella organisms. Because of the substantial number of chil-dren swimming at the lake, a fecal accident was the likely source of contamination.

Vir uses Viruses June New York Norwalk-like 168 The outbreak was associated with swimming in shallow water 1999 virus at a freshwater lake. Feces had been removed from the lake that day by lifeguards.

June Idaho Norwalk-like 25 The outbreak occurred at a resort and water park. The pool 1999 virus implicated in the investigation was untreated. The source of the pools water was a natural hot springs that was high in mineral content and was not chlorinated or filtered. The investigators noted that geothermal pools used for swim-ming are not required to be regulated by public health offi-cials in that locale. The same pool had been implicated in a previous outbreak of Norwalk-like virus in June 1996.

January Wisconsin Norwalk-like 9 A motel pool was linked to the outbreak. The affected per-2000 virus sons had attended a party where diapered infants were in the water; they became ill in <48 hours after attending the event.

Stool specimens collected from these persons tested positive for Norwalk-like virus by reverse-transcriptase polymerase chain reaction and were negative for enteric bacteria.

Other June Illinois Unknown 25 Persons reported gastrointestinal illness after swimming in a 1999 community lake. A septic system failure had occurred, resulting in sewage contamination of the lake. Tests of the lake water on >2 dates before and after the first case was reported were all positive for fecal coliforms.

June California Unknown 23 An outbreak of suspected viral etiology occurred among per-1999 sons who attended three separate pool parties at an apart-ment complex. Initially, the outbreak was reported as a possible foodborne outbreak, but no food items had been

42 MMWR November 22, 2002 State where Suspected Number Outbreak outbreak or confirmed of persons date(s) occurred etiologic agent affected Case description shared among the persons who had attended the three par-ties. Moreover, persons who only ate and did not swim in the pool did not report illness. Stool specimens collected from two persons tested negative for bacterial and parasitic pathogens. No indication was provided that viral testing was done. The health department conducted a physical inspec-tion of the pool and examined pool maintenance records.

No water testing was performed by the health department, nor were pool maintenance records available for the dates immediately after the day of exposure, although reported chlorine levels seemed adequate. However, swimmers reported that on the day in question, the pool was over-crowded and that toddlers were swimming in the pool.

August Florida Sh. sonnei and 38 The outbreak occurred among persons who visited a beach 1999 Cr. parvum park, and it was attributed to both Sh. sonnei and Cr. parvum.

Illness was epidemiologically linked to playing in an inter-active fountain at the park, ingesting water, and consuming food and beverages at the fountain. The fountains recircu-lation and disinfection systems were not approved by the health department and were inadequate or not completely operational at the time of its use. Samples of the fountain water tested positive for coliforms but did not test positive for fecal coliforms. Nevertheless, the cause of the outbreak was determined to be the fountain, which was closed until the health departments concerns could be remedied.§ September Texas Unknown 12 Persons reported symptoms that included exhaustion, sore 1999 muscles, headache, chills, and fever after attending a confer-ence at a Texas guest ranch. One woman reported a miscar-riage during her illness. Exposure to a hot tub, defined as either immersion or being near the hot tub, was associated with illness. Although clinical specimens (urine, blood, spu-tum, and throat swabs) were tested for organisms, including Leg. pneumophila serogroups 1 and 6, influenza virus, parain-fluenza virus, and adenovirus, no infectious agent was iden-tified. No testing for biologic or chemical agents was performed on water samples because the hot tub had already been drained, refilled, and hyperchlorinated before the environmental investigation.

§ CDC. Outbreak of gastroenteritis associated with an interactive water fountain at a beachside parkFlorida, 1999. MMWR 2000;49:565-8.

Vol. 51 / SS-8 Surveillance Summaries 43 State where Suspected Number Outbreak outbreak or confirmed of persons date(s) occurred etiologic agent affected Case description February Vermont Bromine 3 Chemical keratitis resulted from exposure to bromine in a 2000 hotel swimming pool. Bromine levels were >5 ppm (accept-able bromine levels are 1-3 ppm), and the pH level was >8.5.

Patrons that spent time with their heads underwater with their eyes open were affected.

March Maine P. aeruginosa 11 An outbreak of dermatitis affected persons who had stayed 2000 at a hotel during a sports tournament. These persons reported having a rash that was accompanied by other symp-toms, including ear infection, cough, headache, and joint pain. P. aeruginosa was isolated from water samples from both the pool and hot tub. The investigator also indicated that disinfection might have been hindered by a high number of swimmers and the addition of water to the leaking pool.

May Florida Unknown 2 Persons became ill after swimming in a lake that was not 2000 licensed as a bathing area and had signs posted indicating that the area was unsafe for swimming. No water testing results were available before or after the incident. The lake had been closed for swimming for years, and routine samples had not been collected.

July Maine Unknown 32 Swimming at a beach in a campground was epidemiologi-2000 cally implicated in an outbreak of gastrointestinal illness.

Affected persons reported a combination of symptoms, including diarrhea, vomiting, nausea, cramps, and headache that lasted 5-48 days. Only one stool specimen was col-lected, which tested negative for bacterial pathogens; there-fore, the outbreak was of suspected viral etiology. However, no tests for parasitic organisms were reported. The environ-mental assessment and interviews indicated that the pres-ence of diapered children, a heavy bather load, and warm water temperatures were contributing factors to the outbreak.

Water samples tested negative for coliforms.

July Florida Unknown 4 Affected persons reported swimming in a freshwater spring 2000 that had a history of inadequate water quality. Water samples from the springs swimming areas were positive for total and fecal coliforms on multiple dates. Drinking water at this facility was also tested but was negative for coliforms.

August Florida Unknown 9 A motel pool that was cloudy and dirty at the time of expo-2000 sure was implicated. Persons who swam in the pool and did not share any other common exposure became ill with gas-troenteritis. Disinfectant residuals and operation of the fil-

44 MMWR November 22, 2002 State where Suspected Number Outbreak outbreak or confirmed of persons date(s) occurred etiologic agent affected Case description tration system at the time of the investigation were deficient.

Problems had also occurred with the equipment used for adjusting pH.

October Alaska P. aeruginosa 29 In an outbreak of dermatitis, the median number of persons 2000 in the implicated hotel pool and hot tub was greater than the maximal bather loads permitted. The maximum bather load for the pool was nine persons, and for the hot tub, six persons. On the basis of interviews, the median numbers of persons were determined to be 12 (range: 4-20) in the pool and 10 (range: 3-20) in the hot tub. The pool and hot tub were on separate filtration systems. Sand from the pool filter and water from the filter both tested positive for P. aeruginosa. The filtration system was suspected as having malfunctioned sometime during the day of exposure and had been unable to maintain adequate levels of disinfection to accommodate the excessive bather load.

Vol. 51 / SS-8 Surveillance Summaries 45 Glossary Action level A specified concentration of a contaminant in water. If this concentration is reached or exceeded, certain actions (e.g., further treatment and monitoring) must be taken to comply with a drinking water regulation.

Back-siphonage A reversal of the normal flow of water or other liquid caused by a negative-pressure gradient (e.g., within a water system).

Boil-water advisory A statement to the public advising that tap water must be boiled before drinking it.

Cercarial dermatitis Dermatitis caused by contact with the cercariae (larval stage) of certain species of schistosomes whose normal hosts are birds and nonhuman mammals.

Class Waterborne-disease outbreaks are classified according to the strength of the epidemiologic and water-quality data implicating water as the source of the outbreak (see Table 1).

Coagulation The process of adding chemicals to water to destabilize charges on naturally occurring particles to facilitate their subsequent aggregation and removal by flocculation or filtration.

Coliforms All aerobic and facultative anaerobic, gram-negative, nonspore-forming, rod-shaped bacteria that ferment lactose with gas formation within 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> at 95ºF (35ºC).

Community water A public water system that serves year-round residents of a community, subdivision, or mobile system home park that has >15 service connections or an average of >25 residents for >60 days/year.

Contact time The length of time water is exposed to a disinfectant (e.g., chlorine contact time).

Cross-connection Any actual or potential connection between a drinking water supply and a possible source of contamination or pollution (e.g., a wastewater line).

Cyst The infectious stage of Giardia intestinalis and certain other protozoan parasites that have pro-tective walls that facilitate their survival in water and other environments.

Disinfection Chemicals formed in water through reactions between organic matter and disinfectants.

by-products Distribution system Water pipes, storage reservoirs, tanks, and other means used to deliver drinking water to con-sumers or store it before delivery.

Excystation The release of the internal (i.e., encysted) contents (e.g., trophozoites or sporozoites) from cysts or oocysts.

Fecal coliforms Coliforms that grow and produce gas at 112.1ºF (44.5ºC) within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />.

Filter backwash Water containing the material obtained by reversing the flow of water through a filter to dis-lodge the particles that have been retained on it.

Filtration The process of removing suspended particles from water by passing it through one or more permeable membranes or media of limited diameter (e.g., sand, anthracite, or diatomaceous earth).

46 MMWR November 22, 2002 Finished water The water (e.g., drinking water) delivered to the distribution system after treatment, if any.

Flocculation The water-treatment process after coagulation that uses gentle stirring to cause suspended par-ticles to form larger, aggregated masses (floc). The aggregates are removed from the water by a separation process (e.g., sedimentation, flotation, or filtration).

Free, residual chlorine The concentration of chlorine in water that is not combined with other constituents, thus serv-level ing as an effective disinfectant.

Groundwater system A system that uses water extracted from the ground (i.e., a well or spring).

Groundwater under Any water beneath the surface of the ground with substantial occurrence of insects or other the direct influence macrooganisms, algae, or large-diameter pathogens (e.g., Giardia intestinalis or for subpart H of surface water systems serving >10,0000 persons only Cryptosporidium parvum, or substantial and relatively rapid shifts in water characteristics (e.g., turbidity, temperature, conductivity, or pH that closely correlate to climatological or surface water conditions. Direct influence must be determined for individual sources in accordance with criteria established by the state. The state determination of direct influence might be based on site-specific measurements of water quality or documentation of well construction characteristics and geology with field evaluation.

Heterotrophic Microorganisms that use organic material for energy and growth.

microflora Infant water Bottled waters that are marketed for direct consumption by infants or use in mixing with infant formula.

Individual (or private) A water system not owned or operated by a water utility and that serves <15 residences or farms water system not having access to a public water system.

Maximum- The maximum permissible concentration (i.e., level) of a contaminant in water supplied to any contaminant level user of a public water system.

Nephelometric The units in which the turbidity of a sample of water is measured when the degree to which light turbidity units is scattered is assessed with a nephelometric turbidimeter.

Noncommunity water A public water system that 1) serves an institution, industry, camp, park, hotel, or business that system is used by the public for >60 days/year, 2) has >15 service connections or serves an average of >25 persons, and 3) is not a community water system.

Nontransient Public water systems that serve >25 of the same persons for >6 months/year (e.g., a factory or noncommunity school).

water systems Oocyst The infectious stage of Cryptosporidium parvum and certain other coccidian parasites with a protective wall that facilitates survival in water and other environments.

Public water system A system, classified as either a community water system or a noncommunity water system, that provides piped water to the public for human consumption and is regulated under the Safe Drinking Water Act.

Raw water Surface water or groundwater that has not been treated in any way.

Vol. 51 / SS-8 Surveillance Summaries 47 Reverse osmosis A filtration process that removes dissolved salts and metallic ions from water by forcing it through a semipermeable membrane. This process is also highly effective in removing microbes from water.

Source water Untreated water (i.e., raw water) used to produce drinking water.

Surface water The water in lakes, rivers, reservoirs, and oceans.

Total coliforms Nonfecal and fecal coliforms that are detected by using a standard test. Total coliforms are a group of closely related bacteria that are usually free-living in the environment, but are also normally present in water contaminated with human and animal feces. With certain exceptions, they do not cause disease. Specifically, coliforms are used as a screen for fecal contamination as well as to determine the efficiency of treatment and the integrity of the water distribution sys-tem. The presence of total coliforms in drinking water indicates that the system is either fecally contaminated or vulnerable to fecal contamination.

Transient Public water systems that regularly serve >25 of the same persons for >6 months/year (e.g.,

non-community highway rest stations, restaurants, and parks with their own public water systems).

water systems Turbidity The quality (e.g., of water) of having suspended matter (e.g., clay, silt, or plankton) that results in loss of clarity or transparency.

Untreated water Surface water or groundwater that has not been treated in any way (also called raw water).

Water quality A microbial, chemical, or physical parameter that indicates the potential risk for infectious dis-indicator eases associated with using the water for drinking, bathing, or recreational purposes. The best indicator is one whose density or concentration correlates best with health hazards associated with a type of hazard or pollution.

Water utility A water provider that distributes drinking water to a community through a network of pipes.

Watershed An area from which water drains to a single point; in a natural basin, the area contributing flow (i.e., water) to a place or point on a stream.

Watershed-control A program to protect a watershed from contamination or pollution.

program

48 MMWR November 22, 2002 All MMWR references are available on the Internet at http://www.cdc.gov/mmwr. Use the search function to find specific articles.

Use of trade names and commercial sources is for identification only and does not imply endorsement by the U.S. Department of Health and Human Services.

References to non-CDC sites on the Internet are provided as a service to MMWR readers and do not constitute or imply endorsement of these organizations or their programs by CDC or the U.S. Department of Health and Human Services. CDC is not responsible for the content of these sites. URL addresses listed in MMWR were current as of the date of publication.

MMWR The Morbidity and Mortality Weekly Report (MMWR) series is prepared by the Centers for Disease Control and Prevention (CDC) and is available free of charge in electronic format and on a paid subscription basis for paper copy. To receive an electronic copy each week, send an e-mail message to listserv@listserv.cdc.gov. The body content should read SUBscribe mmwr-toc. Electronic copy also is available from CDCs Internet server at http://www.cdc.gov/mmwr or from CDCs file transfer protocol server at ftp://ftp.cdc.gov/pub/publications/mmwr. To subscribe for paper copy, contact Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402; telephone 202-512-1800.

Data in the weekly MMWR are provisional, based on weekly reports to CDC by state health departments. The reporting week concludes at close of business on Friday; compiled data on a national basis are officially released to the public on the following Friday.

Address inquiries about the MMWR series, including material to be considered for publication, to Editor, MMWR Series, Mailstop C-08, CDC, 1600 Clifton Rd., N.E., Atlanta, GA 30333; telephone 888-232-3228.

All material in the MMWR series is in the public domain and may be used and reprinted without permission; however, citation of the source is appreciated.

U.S. Government Printing Office: 2003-533-155/69070 Region IV