Massive microbiological groundwater contamination associated with a by osx43699


Massive Microbiological Groundwater Contamination Associated with a
Waterborne Outbreak in Lake Erie, South Bass Island, Ohio
Theng-Theng Fong,1 Linda S. Mansfield,2 David L. Wilson,3,4 David J. Schwab,5 Stephanie L. Molloy,6 and
Joan B. Rose 1,6
1Department   of Crop and Soil Science, 2Department of Microbiology and Molecular Genetics, 3National Food Safety and Toxicology
Center, and 4Department of Food Science and Human Nutrition, Michigan State University, East Lansing, Michigan, USA; 5National
Oceanic and Atmospheric Administration Great Lake Environmental Research Laboratory, Ann Arbor, Michigan, USA; 6Department
of Fisheries and Wildlife, Michigan State University, East Lansing, Michigan, USA

                                                                                                                   During the time of the outbreak, South
 BACKGROUND: A groundwater-associated outbreak affected approximately 1,450 residents and visitors            Bass Island used three main types of waste-
 of South Bass Island, Ohio, between July and September 2004.                                                 water disposal systems: a) the village of Put-
 OBJECTIVES: To examine the microbiological quality of groundwater wells located on South Bass                in-Bay operated a publicly owned treatment
 Island, we sampled 16 wells that provide potable water to public water systems 15–21 September               works (POTWs) that served the village;
 2004.                                                                                                        b) some of the businesses on the island were
 METHODS: We tested groundwater wells for fecal indicators, enteric viruses and bacteria, and pro-            served by small package wastewater treatment
 tozoa (Cryptosporidium and Giardia). The hydrodynamics of Lake Erie were examined to explore                 works with aeration; and c) on-site waste-
 the possible surface water–groundwater interactions.                                                         water treatment works, such as septic tanks,
 RESULTS: All wells were positive for both total coliform and Escherichia coli. Seven wells tested            mound systems, subsurface sand filters, and
 positive for enterococci and Arcobacter (an emerging bacterial pathogen), and F+-specific coliphage           holding tanks served most of the unincorpo-
 was present in four wells. Three wells were positive for all three bacterial indicators, coliphages, and     rated areas of the island. The small package
 Arcobacter; adenovirus DNA was recovered from two of these wells. We found a cluster of the most             (or semipublic) wastewater treatment plants
 contaminated wells at the southeast side of the island.
                                                                                                              (WWTPs) are privately owned facilities that
 CONCLUSIONS: Massive groundwater contamination on the island was likely caused by transport of               are regulated the same as POTWs by the
 microbiological contaminants from wastewater treatment facilities and septic tanks to the lake and the       Ohio EPA. All POTWs on South Bass Island
 subsurface, after extreme precipitation events in May–July 2004. This likely raised the water table,
                                                                                                              discharge treated effluent to Lake Erie and are
 saturated the subsurface, and along with very strong Lake Erie currents on 24 July, forced a surge in
 water levels and rapid surface water–groundwater interchange throughout the island. Landsat images           regulated by the Ohio EPA under a National
 showed massive influx of organic material and turbidity surrounding the island before the peak of the         Pollutant Discharge Elimination System
 outbreak. These combinations of factors and information can be used to examine vulnerabilities in            (NPDES) permit. Under Ohio NPDES per-
 other coastal systems. Both wastewater and drinking water issues are now being addressed by the              mits (Ohio EPA 2005), sanitary sewage treat-
 Ohio Environmental Protection Agency and the Ohio Department of Health.                                      ment systems are allowed to discharge with a
 KEY WORDS: Arcobacter, groundwater, microbiological contamination, outbreak, viruses, waterborne.            daily fecal coliform bacteria limit of 2,000
 Environ Health Perspect 115:856–864 (2007). doi:10.1289/ehp.9430 available via            colony-forming units (CFU)/100 mL with a
 [Online 6 February 2007]                                                                                     design flow of < 5,000 gal/day if they do not
                                                                                                              discharge directly into an Ohio river.
                                                                                                                   Description of the outbreak. On 2 August
Contaminated groundwater is the most com-             which is the largest community on the island;           2004, the Ottawa County Health Department
monly reported source of waterborne disease in        Put-in-Bay has a permanent population of 350            (OCHD) in Ohio received several telephone
the United States, associated with 64% of the         and up to 25,000 visitors/day during the                calls from persons reporting gastrointestinal ill-
drinking water outbreaks between 1989 and             tourist season. Potable water on the island is          ness after visiting South Bass Island. A food-
2002. In recent national figures (2001–2002),          provided through a number of public and pri-            borne disease outbreak investigation was
groundwater sources constituted 92% of the            vate water systems. A public water system was           initiated by the OCHD and the Ohio
outbreaks, which often occurred in small              defined by the Ohio EPA (2005) as a system               Department of Health (ODH). On 12 August
communities (Blackburn et al. 2004). A large          that has at least 15 service connections or regu-       2004 the Ohio EPA was informed about a
groundwater-associated outbreak in the                larly serves an average of at least 25 individuals      possible waterborne outbreak and began an
Great Lakes basin occurred between June               daily at least 60 days of the year.
and September 2004 on South Bass Island,                   The village of Put-in-Bay is served by a           Address correspondence to J.B. Rose, Department of
Ohio, affecting approximately 1,450 individ-          municipal public water system that uses pri-            Fisheries and Wildlife, 15 Natural Resources, Michigan
uals (both residents and visitors) [Ohio              marily treated surface water from Lake Erie.            State University, East Lansing, MI 48824-1222 USA.
Environmental Protection Agency (EPA)                 However, many businesses and the majority               Telephone: (517) 432-4412. Fax: (517) 432–1699.
2005]. The present study was undertaken to            of residents on the island use untreated                E-mail:
investigate the groundwater quality on the            groundwater pumped from wells on their                    We thank J. Linz for providing guidance in the
                                                                                                              culturing of Campylobacter and Arcobacter; the Ohio
island and the factors associated with the            premises as their primary source of potable             Environmental Protection Agency for their funding
contamination event.                                  water. There are approximately 13 transient             and assistance in sampling; and R. Ives, T. Jenkins,
     South Bass Island is located in Ottawa           noncommunity public water systems and                   and M. Wong for assisting with samples processing.
County, Ohio, off the southern coast of Lake          small businesses on the island that use wells to          This study was supported by the NOAA Center of
Erie, and approximately 5 mi from the                 meet their water needs. According to the Ohio           Excellence for Great Lakes and Human Health
Canadian border (Figure 1). South Bass Island         EPA (2005), transient noncommunity public               (CA4/III-08) and the Centers for Emerging Infectious
                                                                                                              Diseases and Microbial Pathogenesis at Michigan
is one of the main tourist destinations in the        water systems are water systems that do not             State University.
Midwest and has the nickname “Key West of             regularly serve at least 25 of the same persons           The authors declare they have no competing
the Midwest.” Most bars and restaurants on the        over 6 months of the year (e.g., restaurants,           financial interests.
island are located in the village of Put-in-Bay,      campgrounds, gas stations).                               Received 16 June 2006; accepted 6 February 2007.

856                                                                                    VOLUME   115 | NUMBER 6 | June 2007 • Environmental Health Perspectives
                                                                                                                                       Contaminated groundwater and waterborne disease

investigation of the drinking and wastewater                                     During 23–31 August 2004, the Ohio                                  Materials and Methods
systems. On 16 August 2004 the ODH                                          EPA and the CDC performed another study                                  Sampling sites. A total of 16 wells at Put-in-
reported to the Foodborne and Diarrheal                                     sampling for total coliform, Escherichia coli,                           Bay were selected by the Ohio EPA sampling
Diseases Branch at the Centers for Disease                                  and other enteric pathogens such as Campylo-                             team (15–21 September 2004) for study.
Control and Prevention (CDC; Atlanta, GA)                                   bactor at locations identified during the previ-                          Figure 3 shows the layout of the island and
that there were 70 cases of gastroenteritis with                            ous epidemiologic investigation (O’Reilly                                the sampling sites. The 16 wells were labeled
illness onsets between 1 June and 16 August                                 et al. 2007). They also performed macro-                                 according to the Ohio EPA labeling system
2004, including two confirmed cases of                                      scopic particulate analysis from those wells.                            (PB-1, PB-2, PB-3, PB-4, PB-5, PB-6A,
campylobacteriosis. All cases had a common                                  During 8–10 September 2004, the ODH                                      PB-6B, PB-6C, PB-7A, PB-7B, PB-8, PB-9,
history of visiting South Bass Island before                                conducted an extensive environmental assess-                             PB-11, PB-12, PB-14, PB-19). Water sam-
contracting gastroenteritis. The number of                                  ment and water quality monitoring investiga-                             ples were collected from various types of busi-
cases reported per day peaked around                                        tion of private wells on the island.                                     ness including cottages/home rentals (n = 6),
15 August 2004 when 75 cases were reported                                       The present study was initiated to assist the                       parks (n = 3), food services (n = 2), an apart-
(O’Reilly et al. 2007). On 26 August 2004                                   Ohio EPA, following the CDC’s initial investi-                           ment building, a public pool, a campground,
the ODH and the OCHD advised island resi-                                   gation, to further examine the extent of micro-                          a research facility, and a water treatment
dents with private well water to boil their                                 biological quality of the groundwater wells                              plant. Wells were not disinfected before the
drinking water or use bottled water, and the                                located on South Bass Island. Approximately                              outbreaks but some were chlorinated after the
Ohio EPA began ordering water-use advi-                                     1 month after the peak of the outbreak, sam-                             outbreak. At 12 of the sites, semipublic facili-
sories for public water systems with an indica-                             ples were collected from 16 wells that provide                           ties (package WWTPs) were used for waste-
tion of contamination.                                                      potable water to public and private water sys-                           water treatment. Most of the unincorporated
    By 5 September 2004, approximately                                      tems on the island in order to examine the                               areas of South Bass Island are served by on-
1,450 gastroenteritis cases had been reported.                              ongoing risk to the population and residual                              site wastewater treatment works. On-site sys-
Figure 2 shows those cases based on the epi-                                distribution of the contamination on the                                 tems typically consist of a septic tank and
demiologic study undertaken (O’Reilly et al.                                island. Groundwater analyses included conven-                            leach field. Two sites (PB-9 and PB-19) were
2007). The majority of cases occurred between                               tional indicators used for water, such as coli-                          served by on-site systems: PB-9 (airport) used
25 July and 17 August. The CDC detected a                                   form bacteria (including total coliform and                              a mounded system, and PB-19 [ODNR
mixture of pathogens, including Campylobacter                               E. coli), and the alternative indicators entero-                         (Ohio Department of Natural Resources)
spp., Norovirus, Giardia spp., and Salmonella                               cocci, coliphage, and Clostridium perfringens.                           Oak Point picnic area] used a subsurface sand
typhimurium in human fecal specimens                                        The wells were also tested for Campylobacter,                            filter. One site (PB-5; Horny Toad) was
(O’Reilly et al. 2007).                                                     Giardia, Cryptosporidium, and human enteric                              served by the village of Put-in-Bay full-scale
    A case–control study was conducted by                                   viruses. The U.S. EPA standards for drinking                             publicly owned sewage treatment plant (Ohio
the CDC and the ODH between 30 August                                       water for total coliform bacteria are < 1/100                            EPA 2005). A septage disposal site was
and 7 September 2004. A significant associa-                                 mL (U.S. EPA 1986, 2001d). Other fecal indi-                             located in the middle of the island, between
tion between gastroenteritis symptoms and                                   cators (E. coli, enterococci, and coliphage) and                         Catawba Rd. and Put-in-Bay Rd.
tap water consumption, as well as the                                       pathogens should not be present. In addition,                                 Bacterial indicators and coliphage analyses.
amount of tap water consumed, on the island                                 we explored the factors associated with the                              Grab samples were collected using sterile
was reported both from wells and the munic-                                 microbial contaminants and their transport                               bottles; sodium thiosulfate was added to neu-
ipal system (O’Reilly et al. 2007). Whereas                                 using climate and satellite data as well as a                            tralize water samples with detectable chlorine
recreational exposure (i.e., direct contact                                 hydrodynamics model of Lake Erie.                                        residuals. Samples were processed within
with the lake, swimming or wading in the
lake water, swimming in any pool on the                                                     80                                                                                                      3.5
island) was not shown to be statistically asso-                                                                 May monthly rainfall
                                                                                                                event is 200% above
ciated with illness, direct exposure to Lake                                                70                  50-year average
Erie did show an increased odds ratio (OR),
with 13% of the cases and 8% of the controls                                                60

                                                                                                                                                                                                          Average daily precipitation (cm)
showing illness [matched OR = 6.0; p = 0.1;                                                                                                                                                         2.5
95% confidence interval (CI), 0.7–276]
(O’Reilly 2005).                                                                            50                                                 Lake Erie current
                                                                                                                                               event, 7/24
                                                                             No. of cases


                                                           Sampling                         40
                                                                                                                                                                            Boil order, 8/26        1.5
                   Isle St. George                               vania                      30
                      Middle Bass Island        Indiana
                      South Bass Island                                                                                                                                                             1
                                      Kelleys Island

                     Catawba Island

            Port Clinton             Sandusky Bay
                                                                                            0                                                                                                       0
                                                                                             5/1   5/8 5/15 5/22 5/29 6/5 6/12 6/19 6/26 7/3 7/10 7/17 7/24 7/31 8/7 8/14 8/21 8/28 9/4 9/11 9/18
                                     Miles       5         10         15                                                                      Date
Figure 1. Map showing South Bass Island in Lake                             Figure 2. Gastroenteritis cases by date in 2004, average daily rainfall (cm), and key events over the duration
Erie and in relation to the State of Ohio (U.S.                             of the outbreak. Index cases were reported on 30 May 2004 (n = 1,450); the estimated number of cases was
Geological Survey 2006).                                                    obtained from O’Reilly et al. (2007).

Environmental Health Perspectives                      • VOLUME 115 | NUMBER 6 | June 2007                                                                                                              857
Fong et al.

24 hr of collection. Total coliform and E. coli                                          An enrichment step was performed on             MI) at each sampling site. Filtration, elution,
were analyzed by a filtration/agar method                                           samples that were negative by DAL procedure          and recovery of parasites from filters were per-
(American Public Health Association et al.                                          to amplify the number of coliphage in the sam-       formed following U.S. EPA method 1623
1995) and by the Colilert Presence/Absence                                          ples. A 1-L sample was supplemented with             (U.S. EPA 2001c). Parasite oocysts/cysts were
test kit (IDEXX Laboratories, Inc., Westbrook,                                      12.5 mL 4 M MgCl2 and 5 mL host organism             then concentrated into a pellet by the Dynal
ME). Aliquots of each water sample (300 mL)                                         (E. coli C3000 or E. coli F amp) in log phase        Immunomagnetic Separation Technique
were filtered through a membrane filter and                                         and 50 mL 10X TSB. For enrichment with               (IMS; Dynabeads CG-combo Kit; Dynal
enumerated on mENDO (total coliform)                                                E. coli C3000, we also added 10 mL/L 1%              Biotech, Inc., Lake Success, NY). The concen-
media (catalog no. 273620) and EC-MUG                                               nalidixic acid solution. For E. coli F+amp as the    trated pellet was resuspended with MilliQ
(E. coli) media (catalog no. 222200), both                                          host organism, enrichments were supple-              water [obtained from a Nanopure Diamond
from Difco Laboratories, (Detroit, MI).                                             mented with 10 mL streptomycin/ampicillin            Analytical Ultrapure water system (Barnstead
Enterococci were enumerated by membrane                                             solution. Enrichments were then incubated for        International, Dubuque, IA)] into 20 mL and
filtration and cultured on mEI agar (catalog                                        16–24 hr at 37°C. Phage presence in the              split into two. One portion was screened for
no. 214885; Difco Laboratories) following                                           enrichments was confirmed via plaque forma-           the presence of Cryptosporidium oocysts and
U.S. EPA Method 1600 (U.S. EPA 2002).                                               tion using the overlay method described above.       Giardia cysts under Carl Zeiss Axioskop 2
MCP medium (catalog no. 7477A; Acumedia,                                                 Arcobacter/Campylobacter spp. analysis.         fluorescence microscope (Zeiss, Thornwood,
Baltimore, MD) was used for C. perfringens                                          Four liters of grab samples were collected from      NY) after staining with monoclonal antibodies
analysis (Bisson and Cabelli 1979). After incu-                                     each well initially for analysis of Campylobacter    (EasyStain; Biotechnology Frontiers, North
bation, yellow colonies that turned red or dark                                     spp. Concentrated Maximum Recovery                   Ryde, Australia) tagged with fluorescein isoth-
pink after being exposed to ammonium                                                Diluent (Oxoid, Basingstoke, UK) was then            iocyanate. The pellet was also stained with a
hydroxide were counted as C. perfringens. Each                                      added to each water sample at a ratio of 1:10.       0.4-μg/mL 4´6-diamidino-2-phenyl indole
sample was analyzed in triplicate.                                                  Samples were filtered through 0.45-μm mem-            (DAPI) solution. Before staining, IMS con-
     We analyzed coliphages using the double                                        brane filters (Gelman Sciences, Ann Arbor,           centrates were applied to Dynal Spot-On
agar layer (DAL) method or enrichment                                               MI). After filtration, the membranes were then        slides (Dynal Biotech, Oslo, Norway) and air
method as described in U.S. EPA Methods                                             placed on Bolton Selective Enrichment Agar           dried. After air-drying and fixing in methanol,
1601 and 1602 (U.S. EPA 2001a, 2001b). We                                           (BSEA; Oxoid) and flooded with 10 mL                 all samples were stained with 50 μL EasyStain,
used hosts E. coli F-amp (ATCC no. 700891;                                          Bolton broth supplemented with 5% defibri-            followed by 1 mL DAPI and were examined
American Type Culture Collection, Manassas,                                         nated sheep’s blood, cefoperazone (20 μg/mL),        by fluorescence microscopy. The remaining
VA) for detecting F+-specific coliphage and                                         vancomycin (10 μg/mL), and amphotericin B            pellet suspensions were stored for further
E. coli C3000 (ATCC no. 15597) for total                                            (2 μg/mL). The plates (and all other incuba-         analysis [i.e., cell culture if a (oo)cyst was
(somatic and F+-specific) coliphages. In brief,                                      tions) were placed in an anaerobic jar and incu-     detected by microscopy].
we prefiltered 20 mL of each water sample                                           bated at 37°C with 40 rpm agitation under an              Human enteric virus analyses. Approxi-
through a 0.22-μm filter to remove debris and                                        atmosphere of 10% CO2, 10% H2, and 80%               mately 1,000 L of water at each site was filtered
bacteria: 0.5 mL host and 2 mL filtered sample                                       N2. After 48 hr incubation, turbid broth cul-        through a 1 MDS cartridge filter (CUNO Inc.,
were then added to 3 mL trypticase soy broth                                        tures were diluted serially, plated onto BSEA,       Meriden, CT, USA). Virus elution and con-
(TSB) containing 1.5% agar before mixing and                                        and incubated as above without agitation. Five       centration was carried out by organic floccula-
pouring onto a tryptic soy agar plate. Five                                         isolated colonies from the enrichment cultures       tion as described by Fout et al. (1996). Viruses
replicates were assayed for a total of 10 mL.                                       that produced growth were passaged onto              were desorbed from the filters by two rounds
Overlays were incubated at 37°C for 24 hr and                                       BSEA to produce pure culture.                        of reverse passage of 1 L 1.5% beef extract
then assessed for plaque formation.                                                      We performed C. jejuni–specific poly-           solution (1.5% wt/vol beef extract, 0.05 M
                                                                                    merase chain reaction (PCR) analysis to rapidly      glycine, pH 9.0–9.5). Viruses were flocculated
                                                                                    screen colonies isolated on BSEA (Wilson et al.      by the addition of ferric (III) chloride to a final
                                                                                    2000). Colonies were transferred by sterile          concentration of 2.5 mM and by lowering the
                                                     19                             toothpicks to PuReTaq Ready-To-Go PCR                pH to 3.5 (Payment et al. 1984). Viral concen-
                                                                                                                                         trates were centrifuged at 2,500 × g for 15 min,

                                                                                    beads (Amersham Biosciences, Bucks, UK).
              re R

                                                                                    PCR was performed with primer concentra-             and the pellet was resuspended in 30 mL of

                           M                                        Delaware
                                  he                                                tions at 1 pmol/μL. Chromosomal DNA was              0.15 M sodium phosphate (final pH 9.0).


                                                                                    extracted with Easy DNA (Invitrogen,                 Viruses were purified by centrifugation at
                                                           5                        Carlsbad, CA) from selected BSEA isolates.           10,000 × g for 10 min; brought to a neutral
                                                                                    PCR-restriction fragment length polymor-             pH; supplemented with 100 U penicillin,
         14                     w                                              1
                                                                                    phism (PCR-RFLP) was then used to distin-            100 μg streptomycin, and 0.25 μg ampho-



        7a, 7b
                                                                                    guish between Campylobacter, Helicobacter, and       tericin B; and stored at –80°C until analysis.


                                                                                    Arcobacter spp. (Marshall et al. 1999). This              Culturable viruses were assayed on

          11           6a,                            12
                                                                                    PCR amplifies a 1,004-bp fragment within the          Buffalo Green Monkey cells with Eagle mini-
                       6b, 6c 3                                                     coding region of the 16S rRNA gene in                mum essential medium supplemented with
                                              9                                     Campylobacter, Arcobacter, and Helicobacter          2% fetal bovine serum and incubated at

                                                                                    spp. Thereafter, all of these genera can be iden-    36.5 ± 1°C for 14 days. Samples were exam-

                                                               Sampling locations   tified to the genus level with one restriction       ined daily for the development of cytopathic
                                                               Roads                enzyme.                                              effects. All samples underwent a secondary
                                                                                         Cryptosporidium spp. and Giardia spp.           passage after freeze-thaw. We used U.S. EPA
                                                                                    analysis. For parasite analysis (Cryptosporidium     MPN (Most Probable Number) software
Figure 3. Wells sampled on South Bass Island                                        spp. and Giardia spp.), approximately 100 L          (U.S. EPA, Cincinnati, OH) to calculate the
15–21 September 2004; map reproduced with per-                                      water was filtered through an Envirochek HV           MPN/100 L values and confidence limits for
mission from the Ohio EPA (2005).                                                   filter (Pall Gelman Laboratories, Ann Arbor,          viruses.

858                                                                                                               VOLUME   115 | NUMBER 6 | June 2007 • Environmental Health Perspectives
                                                                                                   Contaminated groundwater and waterborne disease

     Enteric viruses were also detected by PCR.       integrated currents for 8 days during the out-                of < 1 NTU (U.S. EPA 1975). Ten of 16 wells
Concentrated water samples (6 mL) were fur-           break period in 2004. Average monthly precip-                 (PB-1, PB-3, PB-4, PB-5, PB-6B, PB-6C,
ther purified, concentrated, and desalted with         itation data for 2004 was obtained from the                   PB-7B, PB-8, PB-11, and PB-19) were found
Centriprep YM-50 concentrator columns                 ODH (2005). True-color LandSat 7 images of                    to have TDS levels that exceeded the National
(Millipore). The final volume of concentrated          western Lake Erie were obtained from                          Secondary Drinking Water Regulations of
eluate recovered was approximately 750 μL.            OhioLINK (2004) during the period of sus-                     500 mg/L (U.S. EPA 1991).
Concentrates were stored at –80°C until analy-        pected contamination.                                              Table 3 shows the results of the micro-
sis. We extracted and purified viral RNA from                                                                        biological analyses arranged from the most
concentrates using a QIAamp Viral RNA Mini            Results                                                       contaminated wells to the least contaminated.
Kit (Qiagen, Valencia, CA) following the              Groundwater quality. Four water samples were                  Total coliforms, E. coli, enterococci, C. per-
manufacturer’s protocol. Purified viral RNA           collected on 15, 16, 20, and 21 September                     fringens, total coliphages, and F±-specific coli-
was eluted in 60 μL of RNase-free water. Each         2004 for a total of 16 samples. The dates and                 phages were monitored as evidence of fecal
concentrated and purified sample was serially          the physical and chemical analyses of ground-                 contamination. Eleven wells were positive for
diluted to a concentration of 10–1 and stored at      water samples, including free Cl2, total Cl2,                 total coliforms (range, 2.2–90 CFU/100 mL;
–20°C. Both concentrations were assayed to            pH, turbidity, and total dissolved solids (TDS),              mean, 12.55 CFU/100 mL) and eight wells
examine inhibition. We performed PCR ampli-           are shown in Table 2. Four samples (PB-3,                     were positive for E. coli (range, 0.1–4
fication for norovirus (NV), human enterovirus         PB-4, PB-11, and PB-14) had free Cl2 and total                CFU/100 mL; mean, 0.64 CFU/100 mL) by
(HEntV), and human adenovirus (HAdV)                  Cl2 ranging from 0.08 to 0.49 mg/L and from                   the membrane filtration method. All wells
using the primer sets shown in Table 1. HEntV         0.04 to 1.00 mg/L, respectively. The pH was                   were positive for both total coliform and
primer sets were able to detect at least 25 differ-   fairly consistent, ranging from 6.8 to 7.4. High              E. coli by the Colilert Presence/Absence test
ent HEntVs; echovirus 22 was not detected.            levels of turbidity were found in PB-8, PB-9,                 kit. Seven wells tested positive for enterococci,
We used a nested primer set designed by Allard        and PB-19, with turbidity of 6.5, 4.1, and 3.9                with counts ranging from 0.1 CFU/100 mL to
et al. (1992) to amplify HAdV. The primers            nephelometric turbidity units (NTU), respec-                  6.6 CFU/100 mL (mean, 1.38 CFU/100 mL).
were able to identify 47 HAdV serotypes,              tively. The National Primary Drinking Water                   C. perfringens was not detected in any well.
including the more common HAdV types 2,               Regulations require a maximum turbidity level                 Overall, five wells (PB-3, PB-9, PB-6B,
40, and 41 (Allard et al. 1992; Puig et al.
1994). Reverse transcriptase (RT)-PCR for             Table 1. Primer sets for virus detection. The equivalent original volume of water analyzed for each sample
noroviruses and enteroviruses was performed           ranged between 4.58 L and 6.46 L for NV, between 1.37 L and 1.94 L for HAdV and between 2.29 L and 3.23 L
using GeneAmp Gold RNA PCR Core Kit                   for HEntV.
(Applied Biosystems, Foster City, CA) accord-         Virus group, primer                    Sequence (5´ to 3´)                  Amplicon (bp)              Reference
ing to manufacturer’s recommendations (Le             NV
Guyader et al. 1996). Primers NVp110 and               NVp110                    AC(A/T/G)AT(C/T)TCATCATCACCATA                           398          Le Guyader et al. 1996
NVp36 were used. Noroviruses G1 and G2                 NVp36                     ATAAAAGTTGGCATGAACA
were used as positive controls (courteously pro-      HAdV
vided by J. Massey, Michigan Department of             AV-A1                     GCCGCAGTGGTCTTACATGCACATC
Community Health, Lansing, MI). We per-                AV-A2                     CAGCACGCCGCGGATGTCAAAGT                                  300          Allard et al. 1992
                                                       AV-B1a                    GCCACCGAGACGTACTTCAGCCTG
formed RT-nested-PCR for HEntV following a             AV-B2a                    TTGTACGAGTACGCGGTATCCTCGCGGTC                            143          Allard et al. 1992
modified protocol by Fong et al. (2005). The           HEntV
final amplicon size was 154 bp. Poliovirus 1,           ENT-up-1                  GTAGATCAGGTCGATGAGTC                                                  Fong et al. 2005
LSc strain (ATCC no. VR-59) was used as a              ENT-down-1                AC(T/C)GG(A/G)TGGCCAATC                                  330          De Leon et al. 1990
positive control; MilliQ water was used as a           ENT-up-2a                 CCTCCGGCCCCTGAATG                                                     De Leon et al. 1990
negative control for all PCR assays.                   ENT-down-2a               ATTGTCACCATAAGCAGCC                                      154          Fong et al. 2005
     Lake Erie hydrodynamic modeling. We              aPrimers used for the second round of PCR.

used a coastal ocean circulation model
(Princeton Ocean Model; Blumberg and                  Table 2. Physical and chemical data of groundwater collected in September 2004.
Mellor 1987) to simulate the currents in Lake                                                        Date          Free Cl2   Total Cl2                Turbidity          TDS
Erie during 2004. The model uses observed             Site                         Sample ID       collected        (mg/L)     (mg/L)           pH      (NTU)            (ppm)
winds from weather stations around the lakes          Put-in-Bay well                 PB-1          16 Sep          ND          0.04        7.12     0.08 (0.32)           410
and buoys in the lake to estimate currents and        Bird’s Nest Resorta             PB-2          15 Sep          ND          ND          7.2      0.1                   480
thermal structure on a horizontal grid with           Clinsters                       PB-3          21 Sep          0.1         0.39        7.4      0.01                  630
2-km grid spacing. This model has been used           Fox’s Dena                      PB-4          15 Sep          0.08        0.1         6.8      0.12                  520
extensively in the Great Lakes for operational        Horny Toad                      PB-5          16 Sep          ND          ND          6.8      0 (0.37)              490
                                                      Island Club well 1             PB–6A          21 Sep          ND          ND          7        0.05                  370
coastal forecasting (Kelley et al. 1998; Schwab       Island Club well 2             PB–6B          21 Sep          ND          ND          7.4      0.01                  560
and Bedford 1999) and has generally proven to         Island Club well 3             PB–6C          21 Sep          ND          ND          7.07     0                     620
be quite accurate (Beletsky and Schwab 2001;          ODNR State Park well 1         PB-7A          20 Sep          ND          ND          7.11     0.05 (0.31)           510
Beletsky et al. 2003). Storm surges in Lake Erie      ODNR State Park well 2         PB-7B          20 Sep          ND          ND          6.92     0.05 (0.07)           820
have previously been shown to influence water          OSU Stone Lab                   PB-8          20 Sep          ND          ND          7.1      6.48 (5.53)         1,370
levels (O’Connor et al. 1999). Currents were          Airporta                        PB-9          15 Sep          ND          ND          6.92     4.14                  500
                                                      Saunders South                 PB-11          20 Sep          0.13        0.2         7.4      0.26 (0.31)           570
calculated on an hourly basis and show consid-        Skyway Loungea                 PB-12          16 Sep          ND          ND          7.2      0.25 (0.43)           490
erable hour-to-hour and day-to-day variability,       Victory Park Resortb           PB-14          15 Sep          0.49        1           7        0.08                  960
mainly due to variability in the wind field. As        ODNR Oak Point                 PB-19          16 Sep          ND          ND          7.09     3.86 (10.74)          600
an example of current variability around the
                                                      ND, not detected. The turbidity of all samples was measured in the field by the Ohio EPA; the turbidity readings taken at
South Bass Island area, we developed and              the Michigan State University laboratory are shown in parentheses.
examined plots of daily averaged, vertically          aChlorinator was turned off before sampling. bBleach was added directly into the well by the owner for decontamination.

Environmental Health Perspectives     • VOLUME 115 | NUMBER 6 | June 2007                                                                                                859
Fong et al.

PB-7A, and PB-12) were positive for three                the presence of enteric virus DNA or RNA via            moderately contaminated with coliform bac-
bacterial indicators (total coliform, E. coli, and       PCR. Adenovirus DNA was detected at two                 teria and E. coli. PB-19 was also positive for
enterococci), and PB-3 had the highest counts            sites, PB-9 and PB-12. The HAdV primers                 enterococci.
for all three indicators.                                used were able to identify 47 HAdV serotypes,               Hydrodynamic modeling. We examined
    Three wells were positive for total coli-            including both respiratory and enteric HAdV             the hydrodynamics of Lake Erie during the
phages (PB-5, PB-6B, and PB-12) and four                 (Allard et al. 1992). Neither enterovirus nor           outbreak to explore the possible surface water–
samples were positive for F + -specific coli-            norovirus was detected in any of the samples.           ground water interactions. Water levels forced
phages (PB-6A, PB-6B, PB-6C, and PB-9)                       Relative microbial contamination.                   by observed winds (O’Connor et al. 1999) and
when analyzed by the two-step enrichment                 Table 3 shows contamination of sites in the             storm surges in the lake are likely to influence
method. One well (PB-6B) was positive for                order of contamination level. PB-9, PB-12,              the subsurface aquifers. Figure 5 shows the
both somatic and F + -specific coliphages.               PB-5, and PB-6B were the most contami-                  daily averaged, vertically integrated modeled
PB-6B was also positive for all three bacterial          nated, followed by PB-3, PB-6A, and PB-6C;              currents in the South Bass Island area plotted
indicators. Although both types of phage are             then PB-19 and PB-7A; and PB-11 and                     for 8 days during the 2004 outbreak period.
indicators of fecal contamination, it has been           PB-8. PB-14, PB-2, PB-4, PB-1, and PB-7B                On the south side of the island, currents were
suggested that the F+-specific coliphages may             were the least contaminated. They had only              most commonly westward with moderated
represent the transport and survival of the              some evidence of coliform and E. coli conta-            speeds ranging from 5 to 10 cm/sec (30 May,
RNA human enteric viruses more adequately                mination via Colilert test. Although the cont-          25 June, 11 July, and 15 August 2004). There
(International Association on Water Pollution            amination was widespread across the island              were also two examples of eastward flow
Research and Control Study Group on Health               (Figure 3), the most contaminated sites were            (1 August and 9 September 2004), one of
Related Water Microbiology 1991).                        along the southeast side of the island. The             strong (> 20 cm/sec) westward flow (24 July
    A preliminary investigation of about                 cleaner sites were all north of these, with the         2004), and one of almost stagnant flow
50% of the pure cultures isolated on the                 exception of PB-5. All (100%) of the wells              (22 August 2004). Currents are almost always
Campylobacter-selective media recovered                  were positive for both total coliform and               weakly southward on the east side of the
Arcobacter spp., which genetically and morpho-           E. coli by the Colilert method, but in spite of         island. On the west side, currents are about
logically resembles Campylobacter. Arcobacter            some chlorine residual, only about 65% and              evenly divided between northward and south-
spp. were detected in seven wells (PB-3, PB-5,           47% were positive, respectively, when culti-            ward on these particular days. On 24 July
PB-6A, PB-6B, PB-6C, PB-9, and PB-12).                   vated by membrane filtration. Enterococci               2004, immediately before the beginning of the
The morphology of these cells was confirmed               and Arcobacter were found in 41% of the                 largest peak in cases during the outbreak, the
under a darkfield microscope. Arcobacter spp.             wells, whereas phage and HAdV DNA were                  current pattern shows clockwise circulation
has now been identified as an emerging cause              found in 37% and 12% of the wells, respec-              around the island, with current speeds exceed-
of diarrhea in humans and is closely related to          tively. Three of the five cleanest sites all car-        ing 20 cm/sec on the south shore. Water
Campylobacter (Wesley 1997). Species-specific             ried chlorine residuals and had low turbidity.          movement was almost stagnant on 22 August
PCR screening for C. jejuni DNA from pre-                Interestingly, site PB-3, despite having chlo-          2004, around the time of the sudden decrease
sumptive Campylobacter isolates was negative             rine residuals, was moderately contaminated;
(Figure 4).                                              this site is located between the two most cont-
    We detected no parasites in the 16 water             aminated sites: PB-9 and PB-12. No chlorine
samples collected. The volume of water ana-              residual was found in the four most contami-
lyzed for parasites ranged between 40 and                nated sites, and PB-9 had high turbidity.
75 L. No cultivatable viruses were detected in           PB-8 and PB-19 both produced water with
the approximately 500 L of groundwater                   high turbidities. These sites were located at
assayed. The samples were also analyzed for              the far northern tip of the island and were

Table 3. Contamination in samples shown in order from the highest to the lowest bacterial indicator and
virus counts.
                                     Bacteria (CFU 100/mL)
              Total       Total                                                        Coliphages      Enteric
             coliform   coliform     E. coli    E. coli                              (enrichment/L)    viruses
Sample ID      (MF)     (Colilert)    (MF)     (Colilert) Enterococci Arcobacter   Total     F-specific HAdV
PB-9             7.8        +          1.3         +           1.9        +         <1          +         +
PB-12            7.7        +          0.3         +           0.6        +          +         <1         +
PB-6B          38           +          0.4         +           0.1        +          +          +         –
PB-5             3.4        +        < 0.1         +           2          +          +         <1         –
PB-3            90          +          4           +           6.6        +         <1         <1         –
PB-6C          26           +          0.1         +         < 0.1        +         <1          +         –
PB-6A            5.9        +        < 0.1         +         < 0.1        +         <1          +         –
PB-19           12.8        +        < 0.1         +           5.9      < 0.1       <1         <1         –
PB-7A            3.7        +          0.7         +           5        < 0.1       <1         <1         –
                                                                                                                 Figure 4. 16S rDNA PCR-RFLP patterns of chromo-
PB-11            2.2        +          0.9         +         < 0.1      < 0.1       <1         <1         –
                                                                                                                 somal extracts from bacterial isolates recovered in
PB-8             3.3        +          2.6         +         < 0.1      < 0.1       <1         <1         –
                                                                                                                 South Bass Island groundwater after DdeI digest.
PB-14          < 0.1        +        < 0.1         +         < 0.1      < 0.1       <1         <1         –
                                                                                                                 Lane 1, 100-bp marker; lane 2, C. jejuni control
PB-2           < 0.1        +        < 0.1         +         < 0.1      < 0.1       <1         <1         –
                                                                                                                 (ATCC no. 11168); lanes 3–9, samples isolated from
PB-4           < 0.1        +        < 0.1         +         < 0.1      < 0.1       <1         <1         –
                                                                                                                 South Bass Island groundwater sites PB-5, PB-12,
PB-1           < 0.1        +        < 0.1         +         < 0.1      < 0.1       <1         <1         –
                                                                                                                 PB-3, PB-6A, PB-6B, and PB-6C, respectively. DdeI
PB-7B          < 0.1        +        < 0.1         +         < 0.1      < 0.1       <1         <1         –
                                                                                                                 restriction fragments of 421, 353, and 183 bp are
Abbreviations: +, positive; –, negative; MF, membrane filtration. All samples were negative for C. perfringen,   indicative of Arcobacter as specified by Marshall
Cryptosporidium spp., Giardia spp., noroviruses, and human enteroviruses.                                        et al. (1999).

860                                                                                       VOLUME   115 | NUMBER 6 | June 2007 • Environmental Health Perspectives
                                                                                                                   Contaminated groundwater and waterborne disease

in cases of the disease; the boil order was not       and 31% of the wells were positive for total                               investigation (15–21 September); 100% were
initiated until 26 August 2004. The pattern of        coliform and E. coli, respectively. About 63%                              positive for E. coli in our analysis using the
the current on 9 September 2004 before our            of the 22 wells at depths between 0–50 feet                                Colilert method, which has been known to
sampling shows the strong (> 20 cm/sec) cur-          and 18.5% of 54 wells at depths of 51–162 ft                               recover chlorine-injured bacteria (McFeters
rents that can occur during fall storms.              tested positive for E. coli. Some of the wells                             et al. 1993).
     The monthly rainfall in 2004 was on              were sampled approximately 1 month after the                                    Campylobacter was isolated from 16 stool
average greater than the 50-year monthly              outbreak during our more intensive microbial                               samples and one well sample during the
averages for the area (Figure 6) and in fact
was > 200% for May and 120% for June.                                            9
Three true-color LandSat images of western                                                                                                 Monthly precipitation data
Lake Erie from 10 May 2004, 27 June 2004,                                                                                                  Monthly average precipitation data (1951–2000)
and 15 September 2004 are shown in                                               8
Figure 7 (OhioLINK 2004). Biological activity
(greenish color) is indicated on the northeast of
the Island in the Figure 7A (10 May), but most                                   7

of the turbidity appears to be inorganic (white
and gray). By 27 June 2004 (Figure 7B), the
reddish colors in the lake indicate a dramatic
increase in biological activity or suspended
                                                        Precipitation (inches)

inorganic material (clay). Finally, by
15 September 2004 (Figure 7C), the activity
(or turbidity) diminished considerably except
in Sandusky Bay.                                                                 4

The South Bass Island waterborne outbreak,                                       3
which took place between June and August
2004, is one of the largest documented in the
Great Lakes in the last decade. Our investiga-                                   2
tion indicated that the fecal contamination was
massive and widespread throughout the
groundwater on South Bass Island. The multi-                                     1

tude of pathogens detected in the population,
including Camplylobacter and norovirus, sug-
gests that human wastes were the source of the                                       Jan   Feb   Mar         Apr    May   Jun       Jul     Aug         Sep      Oct       Nov        Dec
contamination originating from wastewater                                                                                 2004
facilities, septic tank effluent discharges, or pos-   Figure 6. Monthly precipitation (inches) in 2004 and monthly average precipitation data (based on 1951–2000)
sibly septage. The ODH reported that 78%              for the North Central Climatic region of Ohio in 2004. Reproduced with permission from ODH (2005).

  A                                       B                                                        C                                           D

 5/30/04                                  6/25/04                                                 7/11/04                                     7/24/04

10 cm/sec                                10 cm/sec                                               10 cm/sec                                  10 cm/sec

  E                                       F                                                        G                                           H

  8/1/04                                  8/15/04                                                 8/22/04                                      9/9/04

10 cm/sec                                10 cm/sec                                               10 cm/sec                                   10 cm/sec

Figure 5. Modeled currents around South Bass Island in Lake Erie before [(A) 30 May, (B) 25 June], during [(C) 11 July, (D) 24 July, (E) 1 August, (F) 15 August,
(G) 22 August], and after [(H) 9 September] the outbreak in 2004.

Environmental Health Perspectives     • VOLUME 115 | NUMBER 6 | June 2007                                                                                                               861
Fong et al.

investigation by the CDC in late August (> 90%) remove the bacteria and viruses. may have inactivated the viruses. Chlorinators
(O’Reilly et al. 2007) but not in our well sam- Interestingly O’Reilly et al. (2007) reported no were turned off just before collection of sam-
ples. In our study, seven water samples that significant difference between attack rates for ples. Adenovirus DNA was detected in wells
were presumptively positive for Campylobacter residents using well water or municipal water PB-9 and PB-12, which shows that these wells
spp. were identified as being contaminated (treated Lake Erie water). Although not statisti- were vulnerable to human wastes (both wells
with Arcobacter. Arcobacter spp. can be differ- cally significant, an increased OR was found by were < 51 ft deep) and that the use of multiple
entiated from other Campylobacter-like bacte- O’Reilly (2005) for cases with any contact with cell lines may be necessary in cell culture analy-
ria by two distinctive features: They grow at Lake Erie (13% of the cases and 8% of the sis (ODH 2005). The presence of human
15°C, and they are aerotolerant (Wesley controls showed illness, with a matched OR of adenoviral DNA indicated that groundwater
1997). Recent studies suggest that Arcobacter, 6.0; p = 0.1; 95% CI, 0.7–276).                          on South Bass Island was affected by human
especially Arcobacter butzleri, is associated with  Cell culture analysis showed no culturable sewage. The indicator viruses, in this case
persistent, watery diarrhea and bacteremia in virus; chlorination of the wells before sampling coliphages, were also detected (in 37.5% of
infected patients (Vandenberg                                                                                      the wells) and are highly indica-
2004). Little is known about the           A                                                                       tive of sewage contamination.
mechanisms of pathogenicity,                                                                                       They are also useful as viral indi-
potential virulence factors, or clini-                                                                             cators of groundwater contami-
cal importance of Arcobacter spp.                                                                                  nation because of their colloidal
because these organisms are often                                                                                  nature and ability to move
misidentified as a Campylobacter                                                                                   through the subsurface.
spp. if specific testing to species                                                                                    Overall, wells PB-5 (Horny
level is not performed (Diergaardt                                                                                 Toad), PB-6B (Island Club
et al. 2004). It is uncertain how                                                                                  well 2), PB-9 (airport), and PB-12
frequently Arcobacter can be                                                                                       (Skyway Lounge) were among the
found in sewage, surface water, or                                                                                 most contaminated sites; these
groundwater. The high prevalence                                                                                   wells were positive for all three
of Arcobacter in these water sam-                                                                                  bacterial indicators, coliphages,
ples and the ability of the bacteria                                                                               and Arcobacter spp. PB-9 and
to grow at cool temperatures fur-                                                                                  PB-12 were positive for adeno-
ther support its potential as a                                                                                    virus DNA. At PB-9, an on-site
waterborne pathogen. Arcobacter                                                                                    septic system was used for waste-
spp. should be considered as one           B                                                                       water treatment; this could have
of the emerging waterborne bac-                                                                                    been the possible source of the
terial pathogens, and waters                                                                                       contamination in the well. Both
should be further monitored for                                                                                    PB-12 and PB-6B were connected
this bacterium.                                                                                                    to a privately owned treatment
     We did not detect Crypto-                                                                                     facility with discharge to Lake
sporidium oocysts or Giardia cysts                                                                                 Erie.
in the present study using U.S.                                                                                        The possible sources of human
EPA method 1623 (U.S. EPA                                                                                          fecal wastes on this island included
2001c). These organisms are gen-                                                                                   149 household sewage treatment
erally detected in surface waters                                                                                  systems, 81 of which discharge to
and in drinking waters with inade-                                                                                 the subsurface and 68 of which
quate filtration (Betancourt and                                                                                   discharge to Lake Erie. Also, sep-
Rose 2004). This suggests that the                                                                                 tage was also applied to the land
groundwater on South Bass Island                                                                                   in the center of the western por-
was not under the direct influence                                                                                  tion of the island, between
of surface water because the pres-         C
                                                                                                                   Catawba and Put-in-Bay Rd.; the
ence of these parasites in water is                                                                                application was discontinued dur-
normally the result of surface                                                                                     ing this outbreak. On-site waste-
runoff (Hancock et al. 1998). The                                                                                  water disposal systems, which
microscopic particulate analysis                                                                                   discharge to the subsurface, have
test run by the Ohio EPA also cor-                                                                                 been shown to readily contami-
roborated these results, returning a                                                                               nate ground and surface waters.
low risk of groundwater under the                                                                                  Virus transport in karst and
“direct” influence of surface                                                                                      porous aquifers can be rapid,
waters. Although a few cases of                                                                                    ranging from 8 to 54 m/day (Paul
parasitic infections were reported                                                                                 et al. 1997). Viruses, bacteria, and
(three cases of giardiasis), these                                                                                 parasites have been reported at
could have been acquired from                                                                                      concentrations of 102–105/L in
contact with the lake. The                                                                                         sewage depending on the treat-
hydraulic conductivity between the                                                                                 ment scheme (type of secondary
lake and the groundwater, as well                                                                                  treatment) and level of disinfec-
as the type of glacial geology in Figure 7. Landsat satellite images of Lake Erie around South Bass Island 10 May tion (National Research Council
this area, would not effectively 2004 (A), 27 June 2004 (B), and 15 September 2004 (C) (OhioLINK 2004).            1998; Rose et al. 2001). The high

862                                                                              VOLUME   115 | NUMBER 6 | June 2007 • Environmental Health Perspectives
                                                                                                     Contaminated groundwater and waterborne disease

concentrations of bacteria allowed as a part of      25 July and 17 August 2004 (Figure 2). Rapid                         associated with drinking water—United States, 2001–2002.
the NPDES in sewage effluent discharges              transport and interchange likely occur between                       MMWR Surveill Summ 53(8):23–45.
                                                                                                                      Blumberg, AF, Mellor GL. 1987. A description of a three-dimen-
(2,000 CFU/100 mL) suggest that disinfection         the lake and groundwater as the water levels                         sional coastal ocean circulation model. In: Three-
of the wastewater was limited (Ohio EPA              rise as a result of storm surges.These combi-                        Dimensional Coastal Ocean Models, Coastal Estuarine
2005). The porous aquifer on the island pro-         nations of factors and information can be                            Science, Vol 5 (Heaps NS, ed). Washington, DC:American
                                                                                                                          Geophysical Union, 1–16.
vided little to no natural filtration that might      used to examine vulnerabilities in other                         Curriero FC, Patz JA, Rose JB, Lele S. 2001. The association
normally occur during water movement                 coastal systems.                                                     between extreme precipitation and waterborne disease out-
through the soil. In addition, fractures in the          Since the outbreak, the current goal has                         breaks in the United States, 1948–1994. Am J Public Health
limestone aquifer would have allowed for the         been to supply the entire island with fully                      De Leon R, Shieh C, Baric RS, Sobsey MD. 1990. Detection of
transport of bacteria and viruses throughout         treated drinking water from Lake Erie. To                            enteroviruses and hepatitis A virus in environmental sam-
the subsurface. Many existing on-site septic         protect public health, routine monitoring and                        ples by gene probes and polymerase chain reaction. In:
                                                                                                                          Proceedings of the 1990 Water Quality Technology
systems were installed in areas of thin to absent    disinfection of groundwater for potable use                          Conference, 11–15 November, San Diego, CA. Washington,
soils, whereas other wastewater systems dis-         on the island was made mandatory and sep-                            DC:American Water Works Association, 833–853.
charge to Lake Erie (Ohio EPA 2005).                 tage disposal was discontinued. Although lit-                    Diergaardt SM, Venter SN, Spreeth A, Theron J, Brozel VS.
Moreover, the microscopic particulate analysis       tle attention is often given to wastewater                           2004. The occurrence of campylobacters in water sources
                                                                                                                          in South Africa. Water Res 38(10):2589.
test performed by the Ohio EPA on three of           discharges and septic drainfield effects, these                   Fong T-T, Griffin DW, Lipp EK. 2005. Molecular assays for target-
the island wells found no surface water indica-      issues are being addressed by the health                             ing human and bovine enteric viruses in coastal waters and
tors, suggesting that groundwater contamina-         department and the Ohio EPA; an island-                              their application for library-independent source tracking.
                                                                                                                          Appl Environ Microbiol 71(4):2070–2078.
tion through the infiltration of lake water and       wide sewer is also a goal. In spite of having                    Fout GS, Schaefer FW III, Messer JW, Dahling DR, Stetler RE.
poorly installed sewage systems is a more likely     fully treated and disinfected water, the                             1996. ICR Microbial Laboratory Manual. EPA/600/R-95/178.
source of contamination (Ohio EPA 2005).             multiple-barrier concept should be used for                          Washington, DC:U.S. Environmental Protection Agency.
    Extreme precipitation events have also           all waters. Source protection is still needed for                    [accessed 20 June 2006].
been shown to be related to waterborne out-          both groundwater and surface water in order                      Hancock CM, Rose JB, Callahan M. 1998. Cryptosporidium and
breaks (Curriero et al. 2001; Morris and Levin       to fully protect both drinking and recreational                      Giardia in US groundwater. J Am Water Works Assoc
1995; Rose et al. 2000). Curriero et al. (2001)      waters.                                                          International Association on Water Pollution Research and
observed a 2-month lag statistical association           Outbreaks represent observable epidemics                         Control Study Group on Health Related Water Microbiology.
between extreme rainfall events and disease          and are generally considered to be under-                            1991. Bacteriophages as model viruses in water quality
                                                                                                                          control. Water Res 25:529–545.
outbreaks from groundwater. This is likely due       reported. Endemic waterborne disease may still                   Kelley JGW, Hobgood JS, Bedford KW, Schwab DJ. 1998.
to the transport time for pathogen movement          be occurring without being appropriately                             Generation of three-dimensional lake model forecasts for
from the source (i.e., sewage) to the exposure       documented. Because of the multitude of                              Lake Erie. Weather Forecasting 13(3):659–687.
site (i.e., groundwater wells) and the time          pathogens involved and infection with bacteria                   Le Guyader F, Neill FH, Estes MK, Monroe SS, Ando T, Atmar RL.
                                                                                                                          1996. Detection and analysis of a small round-structured
required for disease incubation and disease          such as Campylobacter (associated with Guillain                      virus strain in oysters implicated in an outbreak of acute
reporting. The average monthly precipitation         Barré disease) (Mishu et al. 1993), the possible                     gastroenteritis. Appl Environ Microbiol 62(11):4268–4272.
in the north central region of Ohio was at a         chronic outcomes should be followed in the                       Marshall SM, Melito PL, Woodward DL, Johnson WM,
                                                                                                                          Rodgers FG, Mulvey MR. 1999. Rapid identification of
record high in late May 2004 (ODH 2005).             population exposed in this outbreak. In addi-                        Campylobacter, Arcobacter, and Helicobacter isolates by
Suspected disease cases were first noted by          tion, the role of Arcobacter as a waterborne                         PCR-restriction fragment length polymorphism analysis of
30 May, followed by a small increase in June;        pathogen should be further evaluated. The pre-                       the 16S rRNA Gene. J Clin Microbiol 37(12):4158–4160.
                                                                                                                      McFeters GA, Pyle BH, Gillis SJ, Acomb CJ, Ferrazza D. 1993.
by July (2 months later) a 4-fold increase in        sent study supports the use of remote-sensing                        Chlorine injury and the comparative performance of
cases was reported. The rainfall probably con-       information, climatic data, and hydrodynamic                         ColisureTM, ColiLertTM and ColiQuikTM for the enumeration
tributed to a higher water table, as well as         modeling to examine high-risk water contami-                         of coliform bacteria and E. coli in drinking water. Water
                                                                                                                          Sci Technol 27(3–4):261–265.
increased run-off and flushing of contaminants        nation periods, particularly for islands and                     Mishu B, Ilyas AA, Koski CL, Vriensendorf F, Cook SD,
into the subsurface and Lake Erie.                   coastal systems in the Great Lakes.                                  Mithen FA, et al. 1993. Serological evidence of previous
    We suggest that massive groundwater con-                                                                              Campylobacter jejuni infection in patients with the Guillain
                                                                                                                          Barré Syndrome. Ann Intern Med 118(12):947–953.
tamination on the island was likely caused by                               REFERENCES                                Morris RD, Levin R. 1995. Estimating the incidence of water-
transport of microbiological contaminants                                                                                 borne infectious disease related to drinking water in the
from sewage discharges to the lake and to the        Allard A, Albinsson B, Wadell G. 1992. Detection of adeno-           United States. In: Assessing and Managing Health Risks
                                                         viruses in stools from healthy persons and patients with         from Drinking Water Contamination: Approaches and
subsurface from wastewater treatment facilities          diarrhea by two-step polymerase chain reaction. J Med            Applications (Reichard EG, Zapponi GA, eds). Wallingford,
and septic tanks after extreme precipitation             Virol 37(2):149–157.                                             UK:International Association of Hydrological Sciences,
events in May, June, and July 2004. This may         American Public Health Association, American Water Works             75–88.
                                                         Association, Water Environment Federation. 1995. Section     National Research Council. 1998. Issues in Potable Reuse: The
have raised the water table, saturated the sub-          9222: Membrane filter techniques for members of the col-          Viability of Augmenting Drinking Water Supplies with
surface, and—along with very strong Lake Erie            iform group. In: Standard Methods for the Examination of         Reclaimed Water. Washington, DC:National Academy Press.
currents on 24 July—forced a surge in water              Water and Wastewater. 19th ed. Washington, DC:American       O’Connor WP, Schwab DJ, Lang GA. 1999. Forecast verification
                                                         Public Health Association.
levels and rapid surface water–groundwater           Beletsky, D, Schwab DJ. 2001. Modeling circulation and thermal
                                                                                                                          for Eta Model winds using Lake Erie storm surge water lev-
                                                                                                                          els. Weather Forecasting 14(1):119–133.
interchange throughout the island. Our micro-            structure in Lake Michigan: annual cycle and interannual     ODH. 2005. Investigation of the Ground Water Quality of South
biological monitoring data showed that the               variability. J Geophys Res 106(C9):19745–19772.                  Bass Island, Ottawa County, Ohio. Columbus, OH:Ohio
                                                     Beletsky D, Schwab DJ, Roebber RP, McCormick MJ, Miller GS,
highest amount of contamination was present              Saylor JS. 2003. Modeling wind-driven circulation during
                                                                                                                          Department of Health.
                                                                                                                      Ohio EPA. 2005. South Bass Island, Ottawa County Gastro-
on the southwestern side of the island. The              the March 1998 sediment resuspension event in Lake               intestinal Illness Summer 2004. Columbus, OH:Ohio
Landsat images and the hydrodynamics of the              Michigan. J Geophys Res 108(C2):20-1–20-14.                      Environmental Protection Agency.
                                                     Betancourt WQ, Rose JB. 2004. Drinking water treatment
lake at that time also show that the southwest-          processes for removal of Cryptosporidium and Giardia. Vet
                                                                                                                      OhioLINK. 2004. OhioLINK LANDSAT 7 Satellite Image Server.
                                                                                                                          Available: [accessed
ern end of the island was affected. The amount           Parasitol 126:219–234.                                           10 May 2006].
and timing of the rainfall (2-month lag) and         Bisson JW, Cabelli VJ. 1979. Membrane filter enumeration         O’Reilly CE. 2005. Epi-Aid #2004-076 Final Trip Report: Outbreak
the strong southwestward currents on 24 July             method for Clostridium perfringens. Appl Environ Microbiol       of Gastroenteritis with Multiple Etiologies among Resort
                                                         37(1):55–66.                                                     Island Visitors and Residents—Ohio, 2004. Atlanta,
occurred directly before the beginning of the        Blackburn BG, Craun GF, Yoder JS, Hill V, Calderon RL, Chen N,       GA:Centers for Disease Control and Prevention.
peak of disease cases, which occurred between            et al. 2004. Surveillance for waterborne-disease outbreaks   O’Reilly CE, Bowen AB, Perez NE, Sarisky JP, Shepherd CA,

Environmental Health Perspectives    • VOLUME 115 | NUMBER 6 | June 2007                                                                                                       863
Fong et al.

    Miller MD, et al. 2007. A waterborne outbreak of gastroen-      Schwab DJ, Bedford KW. 1999. The Great Lakes forecasting sys-             Water by Filtration/IMS/FA. EPA-821-R-01-025. Washington,
    teritis with multiple etiologies among resort island visitors        tem. In: Coastal Ocean Prediction, Coastal and Estuarine             DC:U.S. Environmental Protection Agency.
    and residents: Ohio, 2004. Clin Infect Dis 44:506–512.               Studies 56 (Mooers CNK, ed). Washington, DC:American            U.S. EPA. 2001d. Total Coliform Rule: A Quick Reference Guide.
Paul JH, Rose JB, Jiang SC, Zhou X, Cochran P, Kellogg C,                Geophysical Union, 157–173.                                          EPA 816-F-01-035. Washington, DC:U.S. Environmental
    et al. 1997. Evidence for groundwater and surface marine        U.S. EPA. 1975. National primary drinking water regulations:              Protection Agency.
    water contamination by waste disposal wells in the Florida           maximum contaminant levels for turbidity. Fed Reg               U.S. EPA. 2002. Method 1600: Enterococci in Water by
    Keys. Water Res 31:1448–1454.                                        40:59570.                                                            Membrane Filter Using Membrane Enterococcus Indoxyl-B-
Payment P, Fortin S, Trudel M. 1984. Ferric chloride flocculation    U.S. EPA. 1986. Bacteriological Ambient Water Quality Criteria for        D-Glucoside Agar (mEI). EPA-821-R-02-022. Washington,
    for nonflocculating beef extract preparations. Appl Environ           Marine and Fresh Recreational Waters. EPA 440/5-84-002.              DC:U.S. Environmental Protection Agency.
    Microbiol 47(3):591–592.                                             Cincinnati, OH:U.S. Environmental Protection Agency, Office      U.S. Geological Survey. 2006. The National Map Viewer.
Puig M, Jofre J, Lucena F, Allard A, Wadell G, Girones R. 1994.          of Research and Development.                                         Available: [accessed 1 June
    Detection of adenoviruses and enteroviruses in polluted         U.S. EPA. 1991. National secondary drinking water regulations:            2006].
    waters by nested-PCR amplification. Appl Environ Microbiol            secondary maximum contaminant levels. Fed Reg 56:3597.          Vandenburg O. 2004. Arcobacter species in humans. Emerg
    60(8):2963–2970.                                                U.S. EPA. 2001a. Method 1601: Male-specific (F+) and Somatic               Infect Dis 10:1863–1867.
Rose JB, Daeschner S, Deasterling DR, Curriero FC, Lele S,               Coliphage in Water by Two-step Enrichment Procedure.            Wesley IV. 1997. Helicobacter and Arcobacter : potential
    Patz J. 2000. Climate and waterborne disease outbreaks.              EPA 821-R-01-030. Washington DC:U.S. Environmental                   human foodborne pathogens? Trends Food Sci Technol
    J Am Water Work Assoc 92:77–87.                                      Protection Agency.                                                   8(9):293–299.
Rose JB, Huffman DE, Riley K, Farrah SR, Lukasik JO,                U.S. EPA. 2001b. Method 1602: Male-specific (F+) and Somatic          Wilson DL, Abner SR, Newman TC, Mansfield LS, Linz JE. 2000.
    Harman CL. 2001. Reduction of enteric microorganisms at              Coliphage in Water by Single Agar Layer (SAL) Procedure.             Identification of ciprofloxacin-resistant Campylobacter
    the Upper Occoquan sewage authority water reclamation                Washington DC:U.S. Environmental Protection Agency.                  jejuni by use of a fluorogenic PCR assay. J Clin Microbiol
    plant. Water Environ Res 73(6):711–720.                         U.S. EPA. 2001c. Method 1623: Cryptosporidium and Giardia in              38(11):3971–3978.

864                                                                                                          VOLUME   115 | NUMBER 6 | June 2007 • Environmental Health Perspectives

To top