.. flan"
, .16!

, 2%.».
2.2.15:
.1 -ci\, a!

.2? x 1.1321...
iii-“v.2.nvsvnifiuh

loin 11v

. 1.}. in... 4.24..
. ..
).1155Inv_~!.0.
. . .x1l.ri».8.

3:2. ”one...
I ISI‘

.. I...» ..
n“. it...”
r l [l‘

.1..-

 

.( :.,~..;..r.:.3 ..
3...?75flu

it... 1.;

. $ 1... .5

 

 

LIBRARY
’Cq Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

Occurrence and Transport of Waterbome \firuses in Surface
Water in Michigan and Associated Public Health Risks

presented by
Theng Theng Fong

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Crop and Soil Sciences

\ - M <4”:
AQM \\ M\

,l Major Professor’s Signature
5" / o kl / 0?

Date

 

 

 

MSU is an Afiinnative ActiorVEquaI Opportunity Employer

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.

| MAY BE RECALLED with earlier due date if requested.

1

l

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K‘IPrQIAchrelelRC/Daleoue indd

 

OCCURRENCE AND TRANSPORT OF WATERBORNE VIRUSES IN SURFACE
WATER IN MICHIGAN AND ASSOCIATED PUBLIC HEALTH RISKS

by

Theng Theng Fong

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Crop and Soil Sciences

2009

ABSTRACT

OCCURRENCE AND TRANSPORT OF WATERBORNE VIRUSES IN SURFACE
WATER IN MICHIGAN AND ASSOCIATED PUBLIC HEALTH RISKS

By

Theng Theng Fong

Enteric viruses are excreted in high concentrations by humans and are frequently
detected in fecally-contaminated surface water. Enteric viruses, such as adenoviruses and
enteroviruses, are capable of causing a wide spectrum of diseases and are the main cause
of water-transmitted gastrointestinal diseases in swimmers. The overall objective of this
study was to develop an understanding of and quantify human health risks associated
with the presence of viral pathogens, particularly adenoviruses, contracted during
recreational activities at beaches affected by sewage inputs in the lower Grand River,
Michigan. In this research, it is hypothesized that human adenoviruses are present in high
concentrations in sewage and rivers receiving sewage effluents because wastewater
treatments do not efficiently remove these viruses and the risks for virus exposure at
Great Lakes beaches are elevated above an acceptable risk for recreational waters during
the swimming season. In addition, virus transport and attenuation in rivers were evaluated
in a tracer study using bacteriophage P22 as a surrogate for waterborne viruses.
Adenovirus distribution and loading in sewage contaminated water were evaluated by
determining concentrations of human adenoviruses (HAst) in the lower Grand River
during dry weather and after combined sewer overflow (CSO) events. Human
adenoviruses were detected from 6/20 river water samples collected during dry weather

with concentrations ranging between 8 x101 and 6.6 x 104 viruses/L (average: 7.8 x 103

viruses/L). Concentration of HAst in samples collected after CSO events ranged
between 6 x 104 and 1.3 x 106 viruses/L (average: 5.4 x 105 viruses/L). As a part of
exposure assessment and to model virus transport and inactivation afler CSO events, a
dual tracer study was conducted on a 40-km reach of the lower Grand River in Grand
Rapids. From the tracer study, it was concluded that bacteriophage P22 is a suitable tracer
for the complex surface water system with inactivation rates between 0.27 and 0.57 day'l.
A model for estimating virus transport and attenuation after discharge was developed
based on data from the tracer study. With the field survey data and virus transport model,
a quantitative microbial risk assessment was performed to evaluate the probability of
infection or gastrointestinal disease resulting from incidental ingestion of contaminated
recreational water at Lake Michigan beaches that receive input from the lower Grand
River. Monte Carlo simulations were used to characterize uncertainty associated with
different discharge scenarios. Uncertainty analysis showed that river discharge and
concentration of viruses in C80 discharge were the main contributing factors to
differences in risk and duration of elevated risk. The duration of C80 discharge was
identified as the key factor determining the duration of beach closure after a C80 event.
Risk analysis shown that even in the best case scenario (i.e. with the lowest CSO virus
concentration and C80 discharge volume), the risk of acute viral-induced gastrointestinal
illness with swimming in Lake Michigan beaches after a C80 event in the lower Grand
River in Grand Rapids is 2.4x10", which is higher than EPA tolerable level for fresh
recreational water of 8 x 10'3 . The high risk associated with recreating in water receiving
CSO discharge determined from this study emphasizes the importance of proper

wastewater treatment and retention.

ACKNOWLEDGMENTS

NOAA for provided partial funding of the research conducted for this dissertation.
I would like to extend my appreciation to my advisor, Dr Joan Rose, for financial
support, research ideas, patience and understanding through both research and writing of
this dissertation. I also thank my committee members, Dr James Smith, Dr Alvin
Smucker, Dr Hui Li and Dr Phanikumar Mantha, for their patience and guidance through
my research and graduate studies.

Special thanks to my family, friends and lab mates for both their emotional
support and their physical assistance during difficult times through this long and tedious

process, I truly appreciate all you have done for me.

iv

TABLE OF CONTENTS

LIST OF TABLES .................................................................................. vii
LIST OF FIGURES ................................................................................... x
ABBREVIATIONS ......................................................................................................... xiii
CHAPTER 1. INTRODUCTION ........................................................................................ l
Waterbome Diseases and the Sources of Waterbome Pathogens ............................ 2
Transport and Survival of Viruses in Surface Water ............................................. 14
The Economic and Human Health Effects Caused by Deteriorating Water Quality
in the Great Lakes and in Michigan ....................................................................... 19
The Occurrence of Viruses in Waters of the Great Lakes ..................................... 24
Case Study: Investigation of a Waterbome Outbreak and Virus
Occurrence in the Great Lakes Region ...................................................... 25
Quantitative Microbial Risk Assessment (QMRA) ............................................... 36
Hypotheses and Objectives .................................................................................... 38
References .............................................................................................................. 41

CHAPTER 2. MONITORING OF HUMAN VIRUSES IN SURFACE WATER AND

WASTEWATER IN MICHIGAN ..................................................................................... 55
Abstract .................................................................................................................. 55
Introduction ............................................................................................................ 56
Material and Methods ............................................................................................ 58

Sample collection and concentration ......................................................... 58
Detection and quantification of human adenoviruses ................................ 60
Nested PCR and adenovirus DNA sequencing .......................................... 61
Sequencing analysis ................................................................................... 62
Virus recovery efficiency ........................................................................... 65
Results and Discussion .......................................................................................... 66
Wastewater ................................................................................................. 66
Surface water ............................................................................................. 76
C80 samples .............................................................................................. 80
Virus recovery efficiency ........................................................................... 81
Conclusions ............................................................................................................ 82
References .............................................................................................................. 84

CHAPTER 3. THE USE OF BACTERIOPHAGE P22 AS A TRACER FOR
STUDYING VIRAL TRANSPORT AND SURVIVAL IN THE LOWER GRAND

RIVER, MICHIGAN, USA ............................................................................................... 90
Abstract .................................................................................................................. 90
Introduction ............................................................................................................ 91
Material and Methods ............................................................................................ 94

Phage Identification and Isolation ............................................................. 94
Tracer Study ............................................................................................... 99

Study site ........................................................................................ 99

Rhodamine WT and bacteriophage P22 preparation ........... - ........ 100

Tracers release and sampling conditions ..................................... 101
Bacteriophage P22 detection and enumeration ............................ 102

Tracer data analysis ...................................................................... 102

Predicting virus concentration after a CSO event .................................... 103
Sensitivity analysis ................................................................................... 104
Results and Discussion ........................................................................................ 106
Sensitivity analysis ................................................................................... 1 15
Conclusions .......................................................................................................... 123
References ............................................................................................................ 125

CHAPTER 4. RISK OF GASTROINTESTINAL DISEASE ASSOCIATED WITH
EXPOSURE TO ENTERIC VIRUSES AT RECREATIONAL BEACHES
DOWNSTREAM OF CSO SITES ON THE LOWER GRAND RIVER, GRAND .

RAPIDS, MICHIGAN ..................................................................................................... 130
Abstract ................................................................................................................ 130
Introduction .......................................................................................................... 132
Materials and Methods ......................................................................................... 134

Hazard identification ................................................................................ 134

Dose response determination ................................................................... 135
Exposure assessment ................................................................................ 139

Risk characterization ................................................................................ 143

Point risk estimate ........................................................................ 143

Interval risk estimate .................................................................... 145

Recreational Risk on Days without a CSO Event ........................ 149

Sensitivity analysis ................................................................................... 15 1
Results .................................................................................................................. 151
Point risk estimate .................................................................................... 151
Interval risk estimate ................................................................................ 162
Discussion ............................................................................................................ 178
Conclusions .......................................................................................................... 183
References ............................................................................................................ 1 85
CHAPTER 5. SUMMARY AND CONCLUSIONS ....................................................... 188

vi

LIST OF TABLES

Table 1.1 Waterbome viral pathogens and emerging waterborne viral pathogens and
related diseases ......................................................................................................... 5

Table 1.2 Detection and quantification of human adenoviruses from different
environmental matrices ............................................................................................ 8

Table 1.3 Adenovirus serotypes isolated from environmental samples ............................ 11

Table 1.4 Primer sets for virus detection. The equivalent original volume of water
analyzed for each sample 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 for HEV ................ 30

Table 1.5 Samples are arranged in the order of contamination (from the highest to the
lowest bacteria indicator and virus counts) ............................................................ 31

Table 2.1 Primers and probes used in the detection of adenovirus DNA in water samples.
................................................................................................................................ 64

Table 2.2 Arithmetic mean and range of adenoviruses detected from each treatment step
and average loglo removal compared to virus concentration in raw sewage ......... 69

Table 2.3 Comparison between human adenovirus isolates detected in raw sewage and
primary effluent of East Lansing wastewater treatment plant and their closest
prototypes on GenBank. Adenovirus prototypes used in the comparison are:
species A: human adenovirus type 12 (AB330093), type 18 (DQ149610), type 31
(AB330112); species Bl: human adenovirus type 3 (EF494650), type 7
(AC000018); species 32: adenovirus type 35 (AC000019); species C: adenovirus
type 1 (AC000017), type 2 (EU867481); species D: adenovirus type 8
(AB361058), type 19 (AB330133), type 51 (AB330132); species E: adenovirus
type 4 (AB330085); and species F: adenovirus type 40 (AB330121) and type 41
(EF429128). .......................................................................................................... 71

Table 2.4 Surface water samples positive for adenovirus DNA by nested PCR collected
during one-week intensive sampling of the lower Grand River, Michigan, June
19-June 24, 2005. Samples were arranged in the order of sampling sites, from
Lake Michigan to the most inland site on Grand River, Riverside Park, Grand
Rapids, MI. Refer to Figure 2.1 for location of sampling sites ............................ 78

Table 2.5 Nested PCR detection of HAdV (viruses /ml) from surface water samples from
the lower Grand River, MI, between 11/2/05 and 8/9/07 ...................................... 79

Table 3.1 Identity of bacteriophage “PRDl” stock received from Dr Charles Gerba,
University of Arizona, between 1993 and 2002 .................................................... 96

vii

Table 3.2 Comparison of bacteriophages PRDl, bacteriophage P22 and human
adenovirus, an important human pathogen ............................................................ 98

Table 3.3 Distance to the sampling sites fi'om the injection point, comparison of average
velocity and average recovery of bacteriophage P22 using a hydrological model
(Shen et a1. 2008) and sine curve fitting .............................................................. 107

Table 3.4 Bacteriophage P22 and Rhodamine WT arrival and peak time at each sampling
site ........................................................................................................................ 109

Table 3.5 Overall and seasonal mean, and standard deviation for river discharge, CSO
volume and CSO discharge duration. p-values for ANOVA results were listed in
column four. Significant differences in means were observed in seasonal river

discharge .............................................................................................................. 121
Table 4.1 Dose response model for several enteric viruses representing different

potency... ............................................................................................................. 138
Table 4.2 Mean concentration, standard deviation and range of adenovirus concentrations

in wastewater, river water and CSO discharges ................................................... 142
Table 4.3 Parameters and their associated values used in point risk estimation .............. 145

Table 4.4 Parameters and their associated distribution type, mean and standard
deviation used in interval risk estimation. ........................................................... 147

Table 4.5 Parameters and their associated distribution type, mean and standard deviation
used in risk estimation for recreation in surface water without a CSO event ...... 150

Table 4.6 Single event and annual risk of rotavirus and echovirus 12 infection,
morbidity, and mortality for swimming 25 km downstream of a CSO discharge
point. Duration of elevated risk refers to time after virus release. Annual risks
were calculated based on a swim season between May 31 and September 1,
with a maximum exposure of 12 days and one swim event per day .................... 158

Table 4.7 Single event and annual risk of rotavirus and echovirus 12 infection,
morbidity, and mortality for swimming 50 km downstream of a CSO discharge
point. Duration of elevated risk refers to time after virus release. Annual risks
were calculated based on a swim season between May 31 and September 1,
with a maximum exposure of 12 days and one swim event per day .................... 159

Table 4.8 Single event and annual risk of rotavirus and echovirus 12 infection, morbidity,

and mortality for wading 25 km downstream of a CSO discharge point. Duration
of elevated risk refers to time afier virus release. Annual risks were calculated

viii

based on an angling season between May and September, with a maximum
exposure of 20 days and one wading event per day ............................................. 160

Table 4.9 Single event and annual risk of rotavirus and echovirus 12 infection, morbidity,
and mortality for wading 50 km downstream of a CSO discharge point. Duration
of elevated risk refers to time after virus release. Annual risks were calculated
based on an angling season between May and September, with a maximum
exposure of 20 days and one wading event per day ............................................. 161

Table 4.10 Comparison of mean risk of infection for recreational activities using point

estimate, partial Monte Carlo and full Monte Carlo for rotavirus and echovirus 12
in contaminated surface water in the lower Grand River. ................................... 176

ix

LIST OF FIGURES

Figure 1.1 Wells sampled on South Bass Island, Lake Erie, Ohio (labeled PB-l through -
19 from 9/ 1 5-9/21/2004 (map reproduced from Ohio EPA 2005) ........................ 27

Figure 1.2 Gastroenteritis cases by date, average daily rainfall (cm) and key events over
the duration of the outbreak. Index cases were reported on May 30, 2004. (N =
1,450) ..................................................................................................................... 35

Figure 2.1 Sampling sites along the lower Grand River during the intensive study in June
2005, sites that was positive for adenoviruses (0), sites that were negative for
adenovirus (O) and sites selected for continued monitoring between 2005 and
2007 (A) ................................................................................................................. 59

Figure 2.2 Concentrations of HAdV (logro Viruses/L) in wastewater collected from an
advanced wastewater treatment plant in East Lansing, Michigan between August
2005 and August 2006 ........................................................................................... 68

Figure 2.3 Mean adenovirus concentrations (logm viruses/L) in raw sewage, primary
effluent, secondary-treated effluent and tertiary-treated effluent. ......................... 69

Figure 2.4 Adenovirus serotypes isolated from raw sewage and primary effluent samples
collected from East Lansing waste water treatment plant (n=35) .......................... 73

Figure 2.5 A neighbor-joining tree of adenovirus hexon amplicons detected in waste
water. Reference strains of human adenovirus were selected from GenBank under
the accession number indicated in the text. The scale indicates nucleotide
substitutions per position. Numbers above the branches indicate boostrap
percentage (>50) based on 1000 replicates. ........................................................... 74

Figure 2.6 Human adenovirus concentrations (loglo Viruses/L) in CSO discharge samples
collected from Market St retention basin in Grand Rapids, Michigan. ................. 80

Figure 3.1 Map of the Grand River showing the spatial distribution of land use in the
watershed and the three sampling locations ......................................................... 100

Figure 3.2 Bacteriophage P22 concentrations (pfu/ml) at sampling sites versus time after
dye injection (h). (a) At site 1, Wealthy St bridge, average peak bacteriophage
concentration was 1477 pfu/ml; (b) At site 2, 28th St bridge, average peak
bacteriophage concentration was 388 pfu/ml; (c) At site 3, Lake Michigan Drive
bridge, average peak bacteriophage concentration was 74 pfir/ml. ..................... 110

Figure 3.3 Bacteriophage P22 concentrations (loglo mel) at all three sampling sites
versus time after release (loglo min). ................................................................... 113

Figure 3.4 Mean annual river discharge for the lower Grand River at Grand Rapids, MI,
between 1930 and 2008 ....................................................................................... 116

Figure 3.5 Annual CSO volume discharging into the lower Grand river in Grand Rapids,
MI, between year 2000 and 2008 ......................................................................... 117

Figure 3.6 Mean and total CSO discharge volume into the lower Grand river fi'om CSO
outlets in Grand Rapids, MI, by month. Mean and total CSO volume generally
followed the same trend except in January, when mean CSO volume was the
highest. ................................................................................................................ 118

Figure 3.7 Mean river discharge (m3/s) versus mean CSO discharge volume (m3).
Correlation analysis did not show any significant correlation between river

discharge and CSO discharge .............................................................................. 119
Figure 3.8 Average CSO discharge duration (h) by month, no apparent trend was

observed for CSO discharge duration by month .................................................. 120
Figure 3.9 Mean river discharge (m3/s) and mean CSO volume (m3) by season ............ 122
Figure 3.10 Mean CSO discharge duration (h) by season ............................................... 122

Figure 4.1 Probability of rotavirus infection versus time after CSO discharge 25 km (Fig
4.1 a) and 50 km (Fig 4.1 b) downstream of discharge point. CSO virus
concentrations of (i) 60 viruses/ml and (ii) 1322 viruses/ml were compared ...... 154

Figure 4.2 Distribution of rotavirus probability of infection by its relative probability
(occurrence), 24-48 h after release, 50 km downstream of discharge point. Median
probability of infection is lower than mean probability of infection ................... 164

Figure 4.3 Mean risk of rotavirus infection for recreational activities within hours after
CSO discharge, 25 km downstream from discharge point. Mean risk of infection

peaked between 18th and 30th h after release. Mean infection risk was higher than
1.6 x 10'2 (calculated from US EPA tolerable recreational illness risk level of 8 x
103) for at least 96 h after virus release ............................................................... 165

Figure 4.4 Mean risk of rotavirus infection for recreational activities within hours after
CSO discharge, 50 km downstream from discharge point. Mean infection
probability peaked between 30th and 48tln h after release. Mean infection risk was
higher than 1.6 x 10'2 (calculated from US EPA tolerable recreational illness risk
level of 8 x 103) for at least 140 h after virus release .......................................... 166

Figure 4.5 Probability of rotavirus infection versus time after release (h) 25 km
downstream from discharge point. Elevated risk event could happen after more
than 100 h after virus release. River discharge and CSO variables in summer were
used for this simulation ........................................................................................ 168

xi

Figure 4.6 Probability of rotavirus infection verses time after release (h) 50 km
downstream fi'om discharge point. Elevated risk event could happen after
approximately 300 h after virus release. River discharge and CSO variables in
summer were used for this simulation ................................................................. 169

Figure 4.7 Scatter plot of probability of rotavirus infection versus river discharge (m3/s)
25 km downstream of virus discharge point between 30 and 40 hr after virus
release. The probability of infection peaks when river discharge is approximately
35m3/s .................................................................................................................. 171

Figure 4.8 Mean risk of rotavirus infection for recreational activities within hours after
CSO discharge, 50 km downstream from discharge point, simulated with the
overall river discharge values (including all seasons, mean: 106.64 m3/s, std dev:

33.50). Mean infection probability peaked between the 14th and 16th h after
release. Peak mean rotavirus infection risks were 2.71x10'1, 3.28x10'l and
24711101 for adult swimmers, non-adult swimmers and waders, respectively.
Mean infection risk was higher than 1.6 x 10'2 (calculated from US EPA tolerable
recreational illness risk level of 8 x 103) for less than 40 h after virus release...172

Figure 4.9 Relative probability of risk level versus probability of rotavirus infection for
adult swimmers in the full Monte Carlo analysis shows distribution of rotavirus
infections in surface water without a CSO event. Mean: 6.7x10'1, median:
7.0x10", mode: 7.5x10“. ..................................................................................... 177

Figure 4.10 Relative probability of risk level versus probability of echovirus l2 infection
for adult swimmers in the full Monte Carlo analysis shows distribution of
echovirus 12 infections in surface water without a CSO event. Mean: 1.2x10'l,
median: 6.8x10'2, mode: 6.0x10'3 ........................................................................ 178

xii

ABBREVIATIONS

Acoustic Doppler Current Profiler- ADCP
Analysis of variance- ANOVA

Areas of Concem- AOCs

Centers for Disease Control and Prevention- CDC
Colony forming unit- cfu

Combined sewer overflow — CSO

Combined sewer system — CSS

Contaminant candidate list — CCL

Cubic feet per second- cfs

Dissolved organic carbon- DOC

Dissolved organic matter- DOM

Double stranded DNA (/RNA)- dsDNA (fRNA)
Human adenoviruses- HAst

Human enteroviruses- HEVs

Humic acid- HA

International Committee on Taxonomy of Viruses- ICTV
Long-Term Control Plans - LTCPs

Michigan Department of Environmental Quality- MDEQ
Michigan State University- MSU
Microbiological source tracking- MST

National Research Council- NRC

Norovirus- NV

xiii

Ohio Department of Health- ODH

Ohio Environmental Protection Agency- OH EPA
Plaque forming unit- pfu

Polymerase chain reaction- PCR

Quantitative microbial risk assessment- QMRA
Reverse Transcriptase- RT

Rhodamine WT- RWT

Sanitary sewer overflow- SSO

Sanitary sewer system— SSS

Single stranded DNA (/RNA)- ssDNA (/RNA)
Soil and Water Assessment Tool- SWAT

Total suspended solid- TSS

Transient storage- TS

Tryptic soy agar- TSA

Trypticase soy broth- TSB

Ultraviolet- UV

United States Environmental Protection Agency — US. EPA
United States Geological Survey- USGS

Waste water treatment plant- WWTP

World Health Organization— WHO

xiv

CHAPTER 1

INTRODUCTION

Enteric viruses have been detected in various aquatic environments, i.e. ground
water, surface water, estuarine, tap water and marine water (Girones et al. 1995;
Castignolles et a1. 1998; Chapron et a1. 2000; J iang et a1. 2001; Xagoraraki et al. 2007).
Because enteric viruses are excreted in high concentration by infected individuals and
generally are transmitted through the fecal-oral route, water contaminated by fecal
sources remains an important vehicle for the transmission of virus related waterborne
diseases. Virus survival in the water environment is mainly controlled by temperature,
microbial activity, and solar radiation (Hurst et al. 1980; Yates et a1. 1987; Schijven and
Hassanizadeh 2000; Song et a1. 2005). Other factors such as attachment of viruses to
solids, dissolved organic matters and pH may contribute to the survival and transport
behavior of viruses, thus influencing human exposure via contaminated water.

The Great Lakes are important sources for drinking water and recreational
activities. Enteric viruses including rotavirus, adenovirus and enteroviruses have been
isolated from both surface water and groundwater of the Great Lakes (F ong et a1. 2007;
Xagoraraki et a1. 2007); however, waterbome outbreaks related to these viruses may be
underreported because they are not routinely looked for during outbreak investigations. In
older communities of the Great Lakes and Northeast regions, overflows from combined
sewer system (CSS) after heavy precipitation events are major sources of sewage/fecal
pollution, deteriorating surface water quality and beach closure. The cost of beach
closures is high; each beach closure in Michigan was estimated to cost would-be

swimmers between between $1274 and $37030 per day (Rabinovici et a1. 2004). The

actual health impact of a combined sewer overflow (CSO) event associated with illnesses
in swimmers is difficult to assess; however, this may be evaluated through a new
approach using a quantitative microbial risk assessment model.

Quantitative risk assessment is the process of integrating scientific and assessment
data regarding an environmental hazard into a framework to address the risk of exposure
and the potential health impacts (Rose and Grimes 2001). Risk assessment frameworks
have been used extensively to examine human health risks associated with exposure to
toxic chemicals in the environments. Quantitative risk assessment framework may also be
used to assess risk related to exposure to microbial pathogens provided that information
regarding microbial pathogenicity and exposures are available. The initial evaluation
explores the hazards of waterborne disease and methods for assessing exposure,

particularly for the Great Lakes.

Waterbome diseases and the sources of waterborne pathogens

Waterbome diseases have received widespread attention in recent years because
of expanding spectrum of pathogens involved and increasing number of cases reported.
Waterbome diseases may be transmitted through consumption of contaminated water,
inhalation of water vapors/aerosols, dermal contact as well as ingestion during bathing or
recreational activities. Waterbome pathogens are usually spread by the fecal-oral route
with water as an intermediate; for example, an infected host may excrete for example,
10ll virus particles/g of feces for several weeks following initial infection (Wadell et a1.
1987). In the United States, waterbome disease associated with fecal contamination of
drinking water and community supplies as well as recreational waters remains a major

public health problem. More than 50 % of waterborne disease outbreaks reported in the

United States between 1948 and 1994 were determined to be “acute gastrointestinal
illness”, viruses have often been ascribed to these illnesses (Curriero et al. 2001). In the
Great Lake region, few waterborne outbreaks associated with viruses and other pathogens
were reported, but the actual number of viral related outbreaks may be higher as they
often go unreported because the symptoms are usually mild and without long-term
complications (Mac Kenzie et a]. 1994; Hrudey et a1. 2003; Fong et a1. 2007; O'Reilly et
a1. 2007). Waterbome pathogens, especially viruses, may cause a wide range of diseases.
Major symptoms include but are not limited to diarrhea, vomiting, fever, respiratory
infection and conjunctivitis (Bosch 1998). Although the complications of waterborne
infections are usually mild and acute, they can cause severe illness and long-term
complications in young children, immune-sensitive and older populations. In addition,
some viral infections are easily transmissible and the resulting waterborne outbreaks
often have huge economic effects. Norovirus, for example, has a secondary transmission
rate of as high as 90% (Hoebe et a1. 2004).

Globally, waterborne disease is also a major problem, especially in less developed
regions. According to the World Health Organization (WHO), water-related diarrhea]
diseases cause more death than cancer, and are responsible for the deaths of 1.8 million
people every year in less developed regions of the world (WHO 2004; WHO 2005). In
developing countries, up to 13 million deaths annually may be attributed to consumption
of contaminated water (WHO 1996). Of the three important categories of waterborne
pathogens (viruses, bacteria, and protozoa), viruses are among the most resilient and
infectious, yet difficult to isolate because of the lack of sensitive and standard monitoring

protocols. In the United States, viral gastroenteritis is identified as the second most

common cause of illness (NIH 2006). Despite this, viral outbreaks are likely to be
underreported because improved technology for detection of viruses in stool and water
samples is still not widely practiced and in most cases, outbreak samples are only
screened for specific types of viruses.

More than 140 types of pathogenic viruses may be present in high concentrations
in sewage-contaminated waters. Four groups of waterborne viral pathogens that have
been most extensively studied and remain a public health concern are adenoviruses,
noroviruses and enteroviruses and rotaviruses (Moe et a1. 1994; Gerba et a1. 1996;
Crabtree et al. 1997; Parshionikar et a1. 2003; Widdowson 2004; Turcios et a1. 2006).
Recent advancement in molecular detection assays for viruses has brought to the public’s
attention some emerging potential waterborne viruses, viruses that recently have been
isolated from waters, such as polyomaviruses, torovirus, coronaviruses, Torque Teno
virus (TT virus) and picobimaviruses (Theron and Cloete 2002; Vaidya et al. 2002;
Myrmel et a1. 2004; Haramoto et a1. 2005). Although their transmission through water
has not been fully elucidated, these viruses have been shown to persist through
wastewater treatment processes and in aquatic environments. A list of some of the
important waterborne viral pathogens, their characteristics and implications to human

health are summarized in Table 1.1.

 

NocN Sieve—«N

 

v5 :oflaflékoom 3538ha 298%: SAME mESSoM
.6053 :28
Ade/EV mfianogqoocooxsfl
£3— .3 6 30885 539$an “we: 3 “Samoa 308228 0332on momESanbom
newcomer?— “Eta
couoemfi 3 $5.58 “me:
83:55 6.. 8 3 were»;
83 .3 8 DBEm EB wefiopa @5352? 50 55.“ data—805mm <72 5N mamas—$802...
$333885
883% dogmas boa—Emma @2383: £33200...
3ng we owefi 8.5» a 3:8 .58 £22838 .mutoueoobmmm .mEEeU—oamxoou
8cm nos—mmékoov— 35 3? mo 3on 883v < .heé .mefiaoE .mmmbflam SAME cmém .mgozofi moma$o§am
meme?» <ZM 5:: 5323.5 3588me
coca m5: >3 8 E832 Bee mucus ow .288:er .muguogfieou «2th 3-8 memgogv<s
moon—Bowed mEooeeU 8.885 20¢. 203:2 A55 8% meme;

 

.3338 332 93 meowofiam 3.6, 2:383? mafiuoEo can mcowofimm .83 cocoa—noun? : 2an

Of all waterborne enteric viruses, human adenoviruses are recently receiving
renewed attention and study. Human adenoviruses are frequently isolated in high
concentrations in environmental waters; ranging between 104 and 108 viruses/liter in
sewage, 100 and 104 viruses/liter in river, and 10'1 to 103 viruses/liter in seawater (Pina et
a1. 1998; Jiang et a1. 2001; Choi and Jiang 2005; He and Jiang 2005; Bofill-Mas et a1.
2006). Quantification of human adenoviruses from different environmental matrices are
summarized in Table 1.2. There are 51 serotypes of adenoviruses (tentative serotype 52
has just been isolated and sequenced) that cause a wide spectrum of illnesses in human,
from conjunctivitis, respiratory infection, gastroenteritis to haemorrhagic cystitis (Gu et
a1. 2003; Jones et a1. 2007). The 51 serotypes of human adenoviruses are grouped into 6
species (A—F) based on their oncogenicity in newborn hamsters, hemagglutinating, and
DNA sequence properties (Walls et al. 2003). Adenoviruses species F (serotypes 40 and
41), which attack the gastrointestinal system of hosts are the adenovirus species most
frequently isolated from environmental waters (Cho et a1. 2000; Lee and Kim 2002; van
Heerden et a1. 2005; Bofill-Mas et a1. 2006; Xagoraraki et a1. 2007) (Table 1.3).
Adenoviruses are the second most important etiologic agents of childhood gastroenteritis
after rotavirus; together, adenoviruses type 40 and 41 accounted for 50% of adenoviruses
isolated from stool specimens (Wigand et a1. 1983; Brown 1990; Soares et a1. 2002; Li et
al. 2004). In 1998, adenoviruses were identified by the US. EPA as one of nine
microorganisms and one of four viruses (the three others are caliciviruses,
coxsackieviruses and echoviruses) on the Drinking Water Contaminant Candidate List

(CCL) (U .8. EPA 1998). The CCL identifies contaminants of concern, which may require

regulation in the future. One key objectives of the CCL is the need to identify the
occurrence of the contaminant in water.

Adenoviruses belong to the family Adenoviridae. Adenovirus has double-stranded
DNA and ranges in size from 60 to 90 nm (Embrey et a1. 2002). The viruses are shed for
extended periods (up to months or years) in feces, urine and respiratory secretions of
infected persons (Crabtree et a1. 1997). Adenoviruses have been found to be more
resistant to UV disinfection, fluctuations in temperature and humidity than other enteric
viruses (Wasserrnan 1962; Meng and Gerba 1995; Mahl and Sadler 1975; Hara et a1.
1990). They are extremely resistant to UV irradiation because of their high molecular
weight and their undamaged DNA strand may serve as a template for repair by host
enzymes (Roessler and Severin 1996). They have been shown to be up to 60 times more
resistant to UV irradiation than RNA viruses, such as enteroviruses and hepatitis A virus
(Gerba et a1. 2002). Several studies have suggested that adenoviruses outnumbered other
enteric viruses in sewage-contaminated waters and may survive longer than other viruses
in water (Enriquez et a1. 1995; Pina et a1. 1998). The number of waterborne outbreaks
caused by adenoviruses might have been underestimated because they are not being

routinely screened for in outbreak samples.

boom

 

 

 

.3 a 385mm 5.5 .33 mon: .v :25 $33 cream _ :5
Bow
.3 so 38an 53 6an moi} cease ~2on .2 _ B :3
moon
.3 8 man: 93 0: <mD .<U «Son: noise =< 8.0 3v
26:50
Eat—89$
Bow
.3 yo man: 93 on <mD JO mods; “0366 =< 5d Qm
2.2—Eu
mun—5.:—
meow
.3 Ho 2:832 33: .oEom meta; mv35.— FSxBNN =< mod 23:
88 saw
.3 B 32-_Eom .ano303m ocflxwudafiwmm sofixnwd =< wed o
3cm
.3 Ho Begum :33 .993... coca; moths. vocab 0.53m _.o 5::
Sou
.3 8 Saga 525 6»on 53%; $326 «Scam =< _.o :\E
coca .3 B
35853352? :33 hofixmmd $3256 5251 =< 36 m3
custom
naus— Sbuv 3 .8388
3:98qu 393.3530 :2 33$ 52 333 :32 cabeuom 2.5.3» «o .328: Z

 

.moofiafi 3825—835 €28.36 Bod mom—53:83 58:: mo 53393:". 3.3 :ouooaofl N; 033.

 

 

 

 

88 8.5....
.3 3 838: 3., .m 33323 32383 .333. V 3253.3 5 23 33:2
.333 9.3.55 .3323
poem
.3 3 9833: .39: .933 MOE 3233 m25v 233m 3 Sam
“com
.33 28333 .39: .938. 303A $33.? 32:2 3. no 38
88 8.3.2
.3 3 33333 .35 .m 33035 .238; 3232.? .2333 5.. mm 232
88 .3 3 3.6 <5 .5 32 NS 3. 2 3 S:
88 .3 3 33m <8 .5 3233 032.5 2: s:
88 .3 3 3.352 33% moi: .25: 3233 5.. 838 3
88 .3 3 3.332 33% 3233.3 332.3 3233 5.. 333 3
32¢
88 33%
.33 32338 .33283m 3.23.3.3. 9.23%; 3. 3
anew—0mg“
88 33%
.3 3 32-_Eom 38333m .2533 @325; =< m
88 .3 3 3352 33% 32363 3232 3233 =< 3 an
.333
88 33%
.3 3 32338 .2283m 32333.3 323%.”. .2 3.35 3

3.38 2 3333

63398 395% mo “gazebo—gnaw 2560a mo bag: *

 

 

 

88 .130 2:852 ha: .082 “Exam «25A ”.253 =< 2 S
.63”— oar—«5mm
moon .33 2:832 £81083“ «cans ECG Nata: =< 2 ofiom
. Sou
.30 308$: agaiofifi mound 33a N253 265 ~ 25
Sea
.30 sonaam Emmioioe ”26.0 336 No3? =< _ 25
houagaom
boom
.36 gammx <mD.7m.H2 «236 big: otofim ommémm wfim
Son
.30 gomax $3.7m dz moi? ooze. =< 323 mg
Boa
.38 EEEomax «$3.7: dz «35% EUR; =< cmmému on
€853 .195 ‘33— 3:1—

 

PE8 2 0:3

10

 

88 .3 3 .333

 

 

 

 

 

 

330233333 33% 3:35 33.2.5.2 R .3 .N .2 383“ 33835.32.
88 5.
.3 3 0:382 .23: 3:33 33203595 32 :2 3233332
woom
.3 3 2382 23: .3832 333203383 32 z. .N 322 33.3%
88
.3 3 333332 <3 .7: .22 38 83.332 3.9. ”a: 33m 332
Noom fig 35». 3; 8.28m 58>? 93,—.
88 .3 3 26 .332 33863.32.2 303 S .3. .m .2. 333 9:.
88 33B 3.335
.3 3 .3333: a; 3.3:. m .3835 333203393 mom a 33% .3. .9. .N M32: 2333:.
88
.3 3 .3333: 33> 8&3. m 33033 333203883 303 a 323% .2... .2. .N 32 352
Boa .3 3 3.33 <3 .5 303 383-332 59. G: 352
32605 Mom m :2 ~53me
33m 33:. <5 .22 3 3323333 325 z. .2 33833
$2689 mum m :2
33m 33: <5 .22 3 3:33:33. 325 z. .9. .2 .m 3.33%
.38 OS:
.3 3 2:33: .3: .3832 333063.383 mom Ga 33.3%
88 S
.3 3 32238 33% 3:320 .3 .3 .3 .R .2 .: 33.3
8383 mo .0:
nomwom 3830—95.».
oocobmom Bofiqflmooo >333 wfinbobm 258% o>+ mo .02 80532

 

329:3 32:08:83,220 89¢ 38202 89388 2230533 m; 2an

ll

Other key viruses of concern in water are the norovirus and enteroviruses.
Noroviruses cause gastroenteritis and are associated with multiple recreational and
drinking water outbreaks (Parshionikar et al. 2003; Hoebe et a1. 2004; Widdowson 2004;
Maunula L et a1. 2005). Between 2003 and 2004, the Centers for Disease Control and
Prevention (CDC) estimated that the norovirus was the etiologic agent in 4.8% and
16.7% of gastrointestinal illness outbreaks of drinking water and recreational water in the
US, respectively (Dziuban et a1. 2006; Liang ct a1. 2006). Noroviruses (NV s) are
formerly classified as Norwalk-like viruses. The virus was first identified in a
gastroenteritis outbreak in Norwalk, OH in 1968. Norovirus consists of small, circular
and single-stranded RNA. They belong to the family of Caliciviridae and are
approximately 23-35 nm in diameter (Embrey et a1. 2002). Noroviruses have low
infectious dose and can cause prolonged asymptomatic shedding in infected individuals
for up to two weeks (Murata et al. 2007).

Noroviruses are extremely stable in the environment, they are stable in less than
10 parts per million (ppm) chlorine and can withstand freezing and heating to 60 °C
(Nwachcuku and Gerba 2004). Substantial strain diversity leads to short-lived host
immunity to infection, permits re-infection and makes the development of a vaccine that
offers lifelong protection impossible (Glass et al. 2001). Recent data suggested that
humans may not be the only host group for NV, reservoir hosts of NVs may include
calves and pigs (van der Poel et a1. 2000).

Enteroviruses comprise of a large group of viruses that include poliovirus,
coxsackieviruses, echoviruses, and the numbered enteroviruses. At least 89 serotypes of

enteroviruses have been identified and ratified by the Executive Committee of the

12

International Committee on Taxonomy of Viruses (ICTV) (Bfichen-Osmond 2002). Two
members of the enverovirus group - coxsackievirus and echovirus have also been
included in the contaminant candidate list (CCL) of the Safe Drinking Water Act
Amendments of 1996 by the US Environmental Protection Agency (U .8. EPA) (Roessler
and Severin 1996).

Enteroviruses are single-stranded RNA viruses with an icosahedral capsid ranging
from 20 to 30 nm in diameter. Because of their ability to grow in cell culture and be
quantified as plaque forming units (PFU), enteroviruses are the most studied group of
viruses in water and had been included by the European Union regulations governing
water quality as a parameter for evaluating viral pollution in waters (Rao 1986; Pina et al.
1998). Enteroviruses can cause a wide spectrum of diseases in humans. Polioviruses .
usually infect their host by attacking the central nervous system and cause paralysis in
victims (poliomyelitis). Coxsackieviruses have not only been associated with respiratory
system infections and gastroenteritis, but also insulin-dependent diabetes and heart
diseases, such as myocarditis and pericarditis (Kocwa-Haluch 2001). Echoviruses are
usually associated with the common cold and respiratory diseases. The numbered
enteroviruses (type 68 — 71) have not been studied extensively but have been isolated
from patients with bronchiolitis, conjunctivitis, meningitis and paralysis resembling
poliomyelitis (Kocwa-Haluch 2001).

Viral pathogens, especially those that have high prevalence in human populations,
have been suggested as one of the most promising alternative tools to evaluate fecal
pollution as well as public health risk. Viral indicators may be used in conjunction with

bacterial indicators to assess risks relating to water consumption and recreational

13

activities (Pina et al. 1998; McQuaig ct al. 2006). As for tracking the source of fecal
contamination, the host specificity of viruses makes them ideal candidates for source
tracking purposes. In a comparative study of microbiological source tracking (MST)
methods, molecular detection of human enteric viruses reliably identified sewage, though
yielded some false negative results When analyzing individual human feces (Griffith et al.
2003). Human and animal enteric viruses have been tested in several studies to track fecal
contamination of difference origins (Ley et al. 2002; Maluquer de Motes et al. 2004;
Hundesa et a1. 2006).

Of all waterborne human enteric viruses, adenoviruses, with their high prevalence
and stability, and little seasonal variation in shedding, stand out as the most promising
index virus for assessing viral contamination in water (Allard et al 1990, Enriquez et al
1995, Irving and Smith 1981, Krikelis et al. 1985). In addition, because of their clinical
importance and high occurrence in environmental waters, different molecular assays for
detection and quantification of these viruses have been developed. By detection and
quantification of adenoviruses in water, the exposure to viruses from contaminated

surface water may be assessed and evaluated.

Transport and Survival of Viruses in Surface Water
Information regarding the transport and survival of viruses in aquatic
environments is critical for assessment of exposure and alternately risk from waterborne
transmission, especially as a result of recreational water contact. The transport and
survival of viruses depend largely on their genetic makeup (i.e. either DNA or RNA) and
structures (i.e. shape, enveloped or nonenveloped). Viruses are protected by their protein

coat or lipid envelope; thus, they are extremely sensitive to factors that can cause protein

14

degradation. The environmental factors controlling the survival of viruses in aquatic
environments depend largely on the type and physical-chemical characteristics of the
water (i.e. surface water, groundwater, marine etc).

The three most important factors affecting virus survival in surface water are
temperature, microbial activity, and solar radiation (Hurst et al. 1980; Yates et al. 1987;
Schijven and Hassanizadeh 2000; Song et al. 2005). Other factors that may affect the
virus inactivation in surface water include agitation, pH, salinity, organic matter, and
suspended solids and, turbidity (Babich and Stotzky 1980). The combinations of some of
the factors above have shown antagonistic effects on virus survival.

Temperature is one of the most important factors affects viral survival (Hurst et
al. 1980; Yates ct al. 1985; Blanc and Nasser 1996). Both human and animal viruses have
been shown to persist longer and occur more frequently at lower temperatures in natural
environments (Pesaro et al. 1995; Lipp et al. 2001).

High temperatures can damage the virus capsid and nucleic acids, may cause
protein denaturation and prevent adsorption of the virus to its host, as well as inactivate
enzymes required for replication (Bitton 1980). Under subtropical climate, Lipp et a1.
(2001) detected enteroviruses from an estuary in southwest Florida only in winter (when
water temperature was below 23 °C). In an in vitro study, enhanced poliovirus survival
and detection was observed at 22°C as compared to 30 °C (Wetz et al. 2004). Viruses
were detected by RT-PCR for at least 60 days at 22 °C, compared to only 30 days at
30 °C in artificial and filtered seawater. Similarly, Tsai et al. (1993) reported that
poliovirus virions could not be detected by RT-PCR after seven days of incubation at

25 °C, compared to after 21 days when incubated at 4 0C.

15

In addition to water temperature, microbial activity is another important factor
controlling survival of viruses in surface water. Wetz et al. (2004) showed that poliovirus
survival in unfiltered natural seawater was much shorter than survival in filtered seawater
or artificial seawater regardless of incubation temperatures (i.e. 22 °C and 30 °C). In
cases where seasonal water temperature fluctuation is apparent (e.g. subtropical climate),
elevated water temperature and microbial activity has a synergistic effect on virus
survival (Gordon and Toze 2003). As temperature increases, bacterial and protozoan
metabolic processes accelerate, predation increases and degradative enzymes (i.e.
extracellular proteases and nucleases) produced by plants, flagellates or bacteria can
degrade viral capsids and damage viral DNA or RNA (Tsai et a1. 1995; Noble and
Fuhrman 1999). Wait and Sobsey (2001) showed that the decreased survival of poliovirus
incubated in the laboratory at 6 °C in water collected during summer was significantly
compared to survival in water collected during other seasons. When comparing in vitro
and in situ survival in seawater, Wait and Sobsey (2001) reported viruses survived
significantly longer at lower temperatures in laboratory conditions but there was no
significant difference in virus survival between seasons in natural seawater (in situ).

Next to temperature and microbial activity, ultraviolet (UV) radiation may also be
a significant factor affecting virus inactivation. UV irradiation inactivates viruses by
causing cross-linking among virus nucleotides (Gerba 2007). Photosensitivity is virus
type-dependent and related to the structure of viruses (i.e. nucleic acid type-
double—stranded DNA or double-stranded RNA, guanine and cytosine content,
enveloped), and virion size as well as the molecular structure of the specific virus (Meng

and Gerba. 1996). In general, non-enveloped viruses (poliovirus, adenoviruses) are more

16

resistant to UV light than enveloped viruses (vaccinia, herpes simplex, influenza) (Jensen
1964; Meng and Gerba 1996). Virus nucleic acid type may be an important determinant
of its UV resilience; dsDNA viruses such as adenoviruses are extremely stable when
exposed to UV because their undamaged DNA strand may serve as a template for repair
by host enzymes (Gerba et al. 2002; Thurston-Enriquez et a1. 2003). Meng and Gerba
(1996) concluded that adenovirus 40 was 1.2, 1.8, 4.1 and 5.6 times more resistant that
adenovirus 41, bacteriophage MS-2, bacteriophage P22. and poliovirus 1, respectively.
Rotaviruses, which consist of double-stranded RNA have been shown to be more
resistant to UV inactivation than hepatitis A virus, coxsackievirus B5 and poliovirus

type 1 (Battigelli et a1. 1993; Wilson et al. 1993). Sinton et al. (2002) found that
bacteriophage inactivation rates in sunlight are ten times higher than their inactivation
rates in the dark. This is consistent with the findings by Johnson et a1. (1997), who
observed 1 loglo inactivation of polioviruses in marine water after 24 h incubation in dark
compare to 3 logm inactivation of polioviruses incubated under the same condition but
exposed to sunlight.

Attachment of viruses to solids, particularly to sediment with high clay and
organic matter content is generally believed to increase their persistence and stability in
natural environments by offering protection from protein degrading enzymes, other
degrading factors and UV inactivation (Babich and Stotzky 1980; Davis et al. 2006;
Gerba and Schaiberger 1975; Green and Lewis 1999; Lipson and Stotzky 1986; Straub et
a1. 1992). Sediment has been suggested as a reservoir and a sink for viruses, Green and
Lewis (1999) reported that enteroviruses and hepatitis A could be detected throughout the

year in sediment in the immediate vicinity of a sewage outfall even though enterovirus

 

* bacteriophage P22 was misidentified as bacteriophage PRDl in this study

17

concentrations peaked in the wastewater during winter months. Likewise, Ferguson et al.
(1996) reported that during wet season, enteric viruses were isolated fi'om both water and
sediment samples, but during the dry season, viruses were isolated from the sediment
only. On the other hand, some researchers that studied the effect of attachment on virus
survival suggested that the outcome of virus attachment depends on the degree of binding
(Schijven and Hassanizadeh 2000). Schijven et al. (1999) found that the inactivation rate
of attached viruses exceeded the inactivation rate of viruses in the solution.

However, soil organic matters, such as humic acid (HA) have been shown to have
some protective effects on viruses (Foppen et al. 2006). Foppen et al. (2006) reported that
in conditions without HA and in the absence of sand, the inactivation rate of PRDl at
518°C was 0.014 day'l, which was 15.5 times higher than in the presence of HA (0.0009
day]. Although humic acids have been shown to be protective of viruses, Babich and
Stotzky (1980) observed that the survival of bacteriophage (bl 1M15 was not affected by
particulate humic acids without the presence of clay minerals such as attapulgite,
vermiculite, and kaolinite and there was no significant difference in the survival of the
bacteriophage between different types of water amended with heat-killed bacterial cells
(Babich and Sotozky 1980). Thus, it was suggested that the presence of clay minerals
helped the adsorption of viruses to soil organic matters, which has a protective effect on
viruses.

In addition, Foppen et a1. (2006) observed that in the presence of dissolved
organic matter (DOM), viruses could be transported for long distances in surface water
and survive longer because DOM usually out-compete viruses in occupying favorable

attachment sites on solid particles; at the same time, anionic surfactants in DOM has

18

found to be protective of viral particles (Blanford et a1. 2005; Lefler and Kott 1974; Ryan
et al. 2002; Zhuang and J in 2003).

Most-viruses are stable between pH 5 and 9. In general, most non-enveloped
enteric viruses are stable at pH levels as low as 3 and as high as 10 to 10.5; pH affects the
survival of viruses in water by changing the aggregation status of viruses and adsorption
of viruses to the surrounding strata. Higher ionic concentrations (low pH) increase the
virus aggregation as well as adsorption of viruses to particles (Langlet et al. 2007). Yates
et al. (1985) found that inactivation of MS2, poliovirus 1 and echovirus was not
significantly affected by pH ranged between 6.0 and 8.2 (Yates and Gerba 1985). Van
Elsen and Boyce (1966) suggested that only pH higher than 10.5 may cause structural
change in viruses (Van Elsen and Boyce 1966). In addition to pH, the presence of other
chemicals may have an effect on the survival of viruses. Van Elsen and Boyce (1966)
observed that under alkaline condition (pH between 8.5 and 9.5), the presence of free
ammonia shows synergistic effects on the survival of poliovirus (van Elsen and Boyce
1966). Yates et al. (1985) found that ammonia, hardness of water (calcium, magnesium
and total hardness), nitrate, total dissolved solids and turbidity did not significantly affect
the inactivation of MS2, poliovirus 1 and echovirus 1. While it was clear that, globally,
waterborne viruses remain a risk, the evidence for the waters in the Great Lakes needed
to be summarized and assessed.

The Economic and Human Health Effects Caused by Deteriorating Water Quality
in the Great Lakes and in Michigan

The States bordering the Great Lakes have the nation’s longest coastline (5,500

miles) which includes over 1000 beaches for recreational activities (Dorfrnan and Stoner

19

2007). At the same time, the Great Lakes also provide drinking water to over 40 million
US. and Canadian citizens and thus, water quality is an important concern for the region
(Government of Canada and US. EPA 1995). The Great Lakes is one of the largest
sources of fresh surface water, holding 90 % of the fresh surface water in United States
(U .8. EPA 2007). The deteriorating water quality in the Great Lakes caused by
waterborne pathogens, associated with both recreational and drinking water exposures
has been identified as one major threat to public health (Dorfinan and Stoner 2007). The
latest reports on the Great Lakes ecosystem demonstrate that although chemical
contamination has been greatly reduced, microbial contaminants remain a threat and
requires more attention (U .S. EPA and Environment Canada 2007). Contamination of
near-shore waters and recreational waters with microorganisms is increasing, and water
quality is deteriorating with associated beach closures and waterborne outbreaks (U .S.
EPA and Environment Canada 2007).

In 1987, the United States and Canada identified 43 areas of concern (AOC) with
the greatest pollution in the Great Lakes basin, which would require an effort to clean up
and restore. Michigan has 11/26 AOCs within the United States (White, Deer, Torch and
Muskegon Lakes, Saginaw River and Saginaw Bay, and the Kalamazoo, Manistiquc, St.
Mary’s, Raisin, Rouge, and Clinton Rivers) and two shared AOCs with Canada (Detroit
and St. Clair rivers). The AOCs are severely degraded geographic areas within the Great
Lakes Basin that fail to meet the general or specific objectives of the Great Lakes Water
Quality Agreement (1987). Beach closings are one of the beneficial use impairments at
nine of these AOCs. These beach closings signify that the areas have high fecal bacterial

indicator loading and are unfit for recreational use.

20

In Michigan, a total of 33,856 perennial river miles have been assessed for
recreational use impairment, and are graded from pristine to degraded or impaired (Edly
and Wuycheck 2006). The impairment status includes poor fish habitats as well as poor
water quality for recreational use. A water quality assessment by the Michigan
Department of Environmental Quality (MDEQ) reported that between 1999 and 2004,
597 of Michigan’s 33,856 perennial river miles were not supporting the total body
contact recreation designated use, and 19 of those 597 perennial river miles were also not
supporting the partial body contact recreation designated use (Edly and Wuycheck 2006).
The primary sources of microbial contaminants to these non-attaining water bodies
included combined sewer overflows (CSOs), sanitary sewer overflows (SSOs), urban
runoff, and waste water effluent (Edly and Wuycheck 2006).

The surface water in the Great Lakes and Michigan, are an important resource for
recreational activities that involve full body contact with water, such as swimming,
water-skiing, and sail-boarding. Recreational beaches in the Great Lakes are an important
income source for businesses in the coastal region and economic losses caused by beach
closures can have huge impacts on local economy. Across the USA, beach closures have
caused billions of dollars of economic loss. In 2003, there were more than 18,000 days of
closings and advisories on coastal beaches. In the Great Lakes region in 2005, 19% of
monitored beaches (of the total of 1085) were posted or closed more than 10% of
swimming season mainly because of high levels of fecal indicator organisms (U .S. EPA
and Environment Canada 2007). In Michigan alone in 2005, water quality standards were
exceeded from 77 of 406 monitored public beaches, resulting in 80 beach closures (474

beach closing days). Based on the benefit transfer policy analysis by Rabinovici et al.

21

(2004), each closure would cost would-be swimmers between $1274 and $37030 per day,
and the net economic loss in Michigan in 2005 may total anywhere between six and
seventeen million dollars (Rabinovici et al. 2004).

Conditions that may contribute to microbiological contamination of surface water
and groundwater in the Great Lakes region are from both point and non-point sources.
Some examples of non-point sources for waterborne pathogens are combined sewer
overflows (CSO), urban runoff, septic tanks, boat dumping, and concentrated animal
feeding operations (CAFO) (Jiang et a1. 2001; Sharma et al. 2003). In cities with
combined sewer systems (CSSs), CSOs are major concerns for surface water quality.

Combined sewer overflows are a problem nationwide. Combined sewer systems
are sewer systems that are designed to collect sanitary sewage and other waste water
(snowmelt, storm water and urban runoff, and industrial wastewater) in the same pipe to
be transported to and treated by wastewater treatment plants (WWTPs). Currently, there
are approximately 772 communities in the United States with combined sewer systems
(MDEQ 2008). CSSs are mainly located in older cities in the Northeast and the Great
Lakes regions. In Michigan, CSOs were identified among the major sources for beach
closings and other water quality impairments. Combined sewer overflows after heavy
precipitation result in discharges of raw or partially treated sewage from sewer systems
that are designed to carry both domestic sewage and storm water to wastewater treatment
facilities (Whitman et al. 1995). The action to regulate and phase out combined sewer
systems in Michigan began in 1988, as a result of a citizen and public interest group
outcry following a large CSO event in Grand Rapids that impaired the water quality as

downstream as Grand Haven (MDEQ 2008). In 1994, the federal government developed

22

a nationwide CSO Control Policy that required CSO communities to implement interim
measures to improve the quality of combined sewer discharges by January 1, 1997, and to
develop CSO Long-Term Control Plans (LTCPs). As of 2007, 75% of 613 untreated CSO
outfalls that existed in 1988 were eliminated and the remaining 25% are scheduled for
correction under LTCPs. In addition, though the total CSO volume depends largely on
the annual precipitation volume, the portion of partially treated discharge increased.

In the United States, including the Great Lakes, bacterial indicators (total
coliform, fecal coliform, E. coli, enterococcus) have been used for judging microbial
water quality to measure fecal contamination and public health risks. Although bacteria
indicator testing is relatively simple and inexpensive, there are several drawbacks
associated with its application. Bacterial indicators are useful as first level indicators of
fecal contamination in watersheds. However, for wastewater effluents, the indicators do
not reflect the human virus load because of different survivability through waste water
treatments. They lack the ability to predict human health risk and the ability to
differentiate between human and animal contamination for natural waters because they
are not host specific (Colford et al. 2007). The source of fecal coliforrn bacteria and other
traditional indicators (e.g., enterococci) are not limited to humans, they are excreted by
other warm blooded animals and have been reported to occur naturally in soils
(Byappanahalli and Fujioka. 1998; Solo-Gabriele et al. 2000). A study has shown that E.
coli and enterococci accumulate on sand and algae mats at Great Lakes beaches during
the summer and are able to survive for over six months on sun-dried algae mats stored at
4 °C (Whitman et al. 2003). In tropical climates, they have been shown to regrow in the

environment afier excretion by their hosts (Springthorpe et al. 1993). Complicating

23

matters, studies have shown that bacterial indicators do not correlate with waterborne
viral and protozoal pathogens in water, traditional bacterial indicators generally die off
quickly in coastal and marine waters when compared to viruses and protozoa (Bordalo
2002; Solic 1992).

In order to improve swimming conditions and reduce beach closures, the United
States, in the BEACH Act of 2000, has set a goal for 90% of monitored high-priority
beaches around the Great Lakes to meet bacterial indicator standards (E. coli and fecal
coliform) for more than 95 % of the swimming season by 2010 (US. EPA 2002).
However, the actual benefit of beach closure decision based on bacterial indicators is not
clear because bacterial indicators do not reflect the risk fi'om many important waterborne
pathogens, such as viruses, stressed pathogenic bacteria (viable but non-culturable), and
protozoa (Borrego et al. 1987). Infectious enteric viruses have been isolated from aquatic
environments as mentioned previously that are in compliance with bacterial indicator

standards in recreational waters (F ong et al. 2005, J iang et al. 2001).

The Occurrence of Viruses in Waters of the Great Lakes
Pathogenic viruses have consistently been isolated from tap water, surface water,
groundwater, and wastewater globally (Deetz et al. 1984; Shieh 1997; Vantarakis and
Papapetropoulou 1998; J iang et al. 2001; Borchardt et al. 2003; Hararnoto et al. 2004).
Viruses have not been monitored extensively in the Great Lakes area but a few studies
have shown that viruses are present in groundwater as well as surface water in the region
(Borchardt et al. 2003; Borchardt et al. 2004; Jenkins et al. 2005; Xagoraraki et al. 2007).

In 2003, a microbiological water quality survey of nine major rivers in the Lower

24

Peninsula of Michigan found that three of five sites that exceeded the E. coli standard for
recreational waters for the State of Michigan also tested positive for viable enteric viruses
(Jenkins et al. 2005). A water quality survey of two beaches in Indiana and Michigan
showed that human adenoviruses were present in the range of 7(:I: 2) X 100 and

3.8(:l: 0.3) X 103 viruses/L (Xagoraraki et al. 2007). The presence of enteric viruses and
elevated levels of microbial indicators suggests that these. waters may pose a threat to the
health of individuals using them for recreational activities. Threats to drinking water are
generally found during outbreaks. An investigation of a waterborne outbreak at Put-In-
Bay, Ohio, including the assessment of virus concentrations and transport behavior in

surface water in the Great Lakes region, was undertaken.

Case Study: Investigation of a Waterbome Outbreak and Virus Occurrence in the
Great Lakes Region

A possible viral outbreak on an island, South Bass Island, in Lake Erie, was
investigated in 2004 (Fong et al. 2007). This allowed for the development of methods and
tools for virus testing and highlighted the concern for sewage contamination and viral
loading to drinking and recreational waters. The investigation of the Put-In Bay outbreak
also showed the importance of a multi-disciplinary approach in linking contamination
sources to transport and exposure. D

A groundwater-associated outbreak of gastrointestinal disease that affected
approximately 1450 people was reported on South Bass Island, OH, between July and
September 2004. In support of the Ohio Department of Environment, water samples were

collected on the island as a follow-up to the outbreak approximately one month after the

25

peak of the cases. Wastewater was suspected to be the primary source of contamination
and viruses were suspected as one of the major etiologic agents of the outbreak.

Between September 15 and September 21, 2004, 16 groundwater wells (PBl-
PB19) that provide potable water on South Bass island, OH in Lake Erie, were tested for
fecal indicator bacteria (total coliform, E. coli, C. perfringens and enterococci), viruses
(coliphages, human enteroviruses (HEV), adenoviruses (HAdV) and norovirus (NV )) and

parasites (Cryptosporidium and Giardia) (Figure 1.1).

26

     
 

Legend
1 Sampling Locations

Figure 1.1 Wells sampled on South Bass Island, Lake Erie, Ohio (labeled PB-l through -

19 from 9/15-9/21/2004 (map reproduced from Ohio EPA 2005).
Eleven and eight wells were positive for total coliforms (range: 2.2 CFU/ 100 ml,
90 CFU/ 100 ml; mean: 12.55 CFU/ 100 ml) and E. coli (range: 0.1 CFU/ 100 ml and 4

CPU/100 ml; mean: 0.64 CFU/ 100 ml), respectively, by the membrane filtration method.

27

All wells were positive for both total coliforrn and E. coli by the Colilert®
Presence/Absence test kit (which has been known to recover chlorine injured bacteria).
Seven wells tested positive for enterococci, with counts ranging from 0.1 CFU/ 100 ml to
6.6 CFU/ 100 ml (mean: 1.38 CFU/ 100 ml). C. perfringens was not detected in any well.
Overall, five wells (PB-3, PB-9, PB-6B, PB-7A, and PB-12) were positive for three
bacterial indicators and PB-3 had the highest counts for all three indicators (total
coliform, E. coli and enterococci). Three wells were positive for both somatic and F‘-
specific coliphages with E. coli C3000 host and four wells were positive for F+-specific
coliphage with E. coli F amp host. Arcobacter spp., an emerging bacterial pathogen, was
present in seven wells tested. None of the wells were positive for the protozoa. Overall,
three wells were positive for all three bacterial indicators, coliphages, and Arcobacter
spp.

For the detection of viruses from groundwater, large volumes of water (1 000
liters) were filtered through 1 MDS filter. Samples were then concentrated to
approximately 30 ml by organic flocculation as described by the US. EPA ICR Microbial
Laboratory Manual (Fout et al. 1996). Water samples were tested for the presence of viable
viruses by inoculating on Buffalo Green Monkey (BGM) cells. These samples were also
tested for the presence of human adenovirus DNA, and human enterovirus and norovirus
RNA through nested polymerase chain reaction (PCR).

Primers in PCR assays were selected from highly conserved regions of the HAdV,
HEV and NV genomes, which allowed for detection of multiple members from each
group of viruses. Primers used for the detection of viral DNAs/RNAs are shown in Table

1.4. For HAdV, the nested-primer set designed by Allard et a1. (Allard et al. 1992) was

28

used to amplify HAdV in this study. The primers are able to identify 47 HAdV serotypes,
including the more common enteric HAdV types 2, 40, and 41 (Allard et al. 1992; Puig et
al. 1994). For HEV detection, a reverse transcriptase (RT)-nested PCR was used with the
pan-enterovirus primer set (ENT-up-2 and ENT-down-l) by De Leon et al. (De Leon
1990) and HEV primer set from Fong et a1. (2005). HEV primer sets amplified 5’
untranslated region of HEV genomes and were able to pick up at least 25 different HEV;
Echovirus 22 is not detected with these primers (De Leon 1990; Donaldson et al. 2002).

No viable virus was detected in any of the 16 samples cultured on BGM cells. By
using PCR, adenoviral DNA was detected in two of three most contaminated wells (PB-9
and PB-12) by bacterial indicator standard. Sampled groundwater wells, in the order of
contamination (from the highest to the lowest bacteria indicator and virus counts) are
listed in Table 1.5. All samples were negative for C. perfringen. Cryptospon’dium spp.,

Giardia spp., noroviruses and human enteroviruses.

29

.mom .3 388.. 3:83 05 com 38: Eofitm 3

 

 

 

 

<o<<ee<ooofiw<<<§< 3.352
8% .33 333:0 3 Mam <2ou<oe<ue<oe€o§<6E$o< 2 :52 >2

88 .3 3 3:03 5 80532285533. .~-§8-.Em
82 .3 3 83 on Obieeoouooeooeou 3.3.35
82 .3 3 83 5 9mm uefiugoefioioofieg Teaozzm.

88 .33 333 0536383353050 33-95 E:
8% .3 3 33:3. 9; 058883500085960505 «8-3.
08853850330858 Lm-><
$3 .3 3 222 com 5305208808596 2-3.

85502339885588 _<->< >33

33 @5on

805.63% 38:933. on 8 .mv ooeoswem E085 25>

 

5m: 8m A mN.m 93 A mNN 52309 Ea avg com A 3A use A nmA amazon .>Z you A ovd 35

A wmé 502509 30mg 29.53 some 8m BSA—«SW .8233 .3 0823 353.5 “5.33:3 BE. douoouee 95> new 38 SEE 3A 2an

30

v v Ad V m + 56 + Em <h-mm

- v V _.o v 3 + _.o v + WN— 21mm
- + _v + 3o v + 3c v + 3 [Erma
+ _v + _.o v + _.o + ow ”Velma
- v v + 06 + v + ca Wmm
- v + + . N + _.o v + in 9mm
- + + + _.o + vd + mm mo-mm
+ v + + ed + md + >6 NTmm
+ + _v + 2 + 2 + We 3mm

 

.. >23 83.3%.; 35 333233 388325 €238 an): 93:38 Es e vansm
.3... .m .38 .m 8.328 50:30
33 3E.

 

333$ 3 33:33:38 as 8: EB 3338
053m mommy—EEO

 

.3560 25> Ba .8839: «£203 A826— 2: 9 83$: 05 805 acreage—So me .838 05 E woman—.8 can moafimm mA 033.

31

”doggone? 58:: ”>33 *

 

~V

~V

~V

—V

#V

~V

—V

~V

~V

~V

~V

—V

—V

_V

fiov

_.ov

_.ov

fiov

_.cv

_.ov

_.ov

fiov

_.ov

fiov

_.ov

fiov

fiov

—.ov

fiov

fiov

fiov

_.ov

fiov

9N

ad

+ fiov

+ _.ov

+ fiov

+ fiov

+ _.ov

meE

7mm

vimm

NAE

3A:

w-mm

:Imm

P58 3 2%...

32

The isolation of these pathogens fiom the wells showed that groundwater on the
island was contaminated by sewage. Moreover, the isolation of human adenovirus DNA
showed that groundwater on South Bass Island was susceptible to virus contamination.
Failure to detect viable enteric viruses may have been due to chlorination of the wells
prior to sampling and inefficient virus concentration and detection method. The possible
sources of microbial contamination included sewage disposal to lake, onsite septic
systems and land application of septage on the island. Waterbome enteric viruses,
because of their small sizes (generally range between 20-200 nm) were able to infiltrate
easily through the limestone-based aquifer on the island. The presence of adenoviruses
confirmed the susceptibility of groundwater on the island to fecal contamination.

In order to link fecal contamination of groundwater to its sources as well as
surface water-ground water interaction, the hydrogeology and hydrodynamics of the
island were examined. Because the bedrock on South Bass Island consists of dolomite
that is between 30 to 65 feet deep across the island (ODH 2005), the porous limestone
aquifer on the island provided little to no natural filtration of microorganisms from
sewage that might normally occur during water movement through the soil into the
groundwater. Transport of bacteria and viruses throughout the subsurface through
fractures in the limestone aquifer is also highly possible. Viruses have been shown to
persist for extended period (for months) in groundwater (Yahya et al. 1993). In addition,
because groundwater on the island is recharged mainly by precipitation and surface
water, heavy precipitation in May and June 2004 (200% and 120% of 50-year average,
respectively), might have contributed to increase transportation of the pathogens from

surface water into ground water.

33

Daily averaged, vertically integrated modeled currents in the South Bass Island

area were plotted for eight days prior, during and after the outbreak, were examined for a
relationship between current speed and direction, and outbreak cases (Figure 1.2). On
July 24, 2004, directly prior to the beginning of the largest peak in cases during the
outbreak, the current pattern shows clockwise circulation around the island with current
speeds exceeding 20 cm/sec on the south shore. Water movement was almost stagnant on
August 22, 2004, around the time of the sudden decrease in cases of the disease (the boil
order was not initiated until August 26, 2004). The current pattern on September 9, 2004,
prior to the virus sampling showed the strongest (more than 20 cm/sec) currents that can
occur during fall storms (Fong et al. 2006).

In conclusion, we were able to show that ground water contamination on South
Bass Island was the cause of the outbreak and sewage was the source of the
contamination. The isolation of adenovirus DNA indicated that the ground water on
South Bass Island is susceptible to virus contamination, and other enteric viruses could
have been presence in the ground water during the time of the outbreak. In addition, the
result of this study also showed that adenoviruses may be more persistent in water than
other enteric viruses, such as norovirus (which was identified in several patient’s samples
but not in the water), thus making them good candidates as index viruses. Also, while
some of the cases documented could have been caused by adenoviruses, adenoviruses
were not adequately addressed as one of the etiological agents, suggesting that a more
sensitive method for the detection and quantification of adenoviruses (i.e. quantitative

PCR) should be applied for future outbreak and water quality evaluations

34

N‘ N fl . . .

. H U A V # a“ . M. H NA womwo m. .
—.

3am—

 

 

 

» ‘ .v C. i *1: l. L til
,_ .e: _ “1.. “AL 2.... o
. . . ... . 5.x. ..a a»
. .. o . 4 . _ __ .
. _ t _
...... ._ __ ii...
... . u ." ............ .l or
_._ . . ._ 3...... ..
m . . Z .. .3...
. . ... . ...; .-..
m u .. _.. ._
fi h w.... n. I". ...
m . . .... .. .... 0... on N
.l ... __ ... _
.m... . _ ..n. ._ ... ._ n
a ., __ . ... _. L ._. m.
n N. u. as” . .. fl. _. ...-9. a
v. .288 zom _ M .. _ _ .. M
H ~ .. _ u“ ... ._ I.
m. . . . __ _. ..- m
e mm. . ER "n a __ on a
ma W 395 E055 _.n s _ s
w . 2.5%.. “ .... 8
A f. . .
. 6
:on
an a as. 8.5385 .....8 use... - -r - .53 Eva. .82
830%..0952'] .
. on

35

Quantitative Microbial Risk Assessment (QMRA)

Exposure to varying levels of viruses through drinking water and recreational
water is likely occurring in the population in the Great Lakes region. The level of disease
that may be associated with this exposure can not be measured with current
epidemiological and health surveillance methods. In order to accurately estimate health
risk related to direct contact with contaminated surface water in the Great Lakes, it is
necessary to measure the concentration of viruses and study the survival and transport
behavior of viruses in an actual water system, and then incorporate these data into a
quantitative risk assessment model. With information on enteric virus occurrence and
concentration in aquatic environment, health risk related to exposure to these viruses in
the environment can be estimated using a risk assessment framework.

Risk assessment is the process of integrating scientific and assessment data
regarding an environmental hazard into a framework to address the risk of exposure and
the potential health impacts (Rose and Grimes 2001). Risk assessment frameworks have
been used extensively to examine human health risks associated with exposure to toxic
chemicals in the environment (e.g. residential area, workplace, heavy-metal contaminated
site), health care costs and benefits of developing better management policies and
prevention controls. Risk assessment is the first step of risk analysis, results fiom risk
assessment can be used to direct risk management and risk communication. For example,
policy makers may refer to risk assessment results to decide if additional treatment of a
contaminated source is necessary to significantly reduce effect on public health.
Quantitative microbial risk assessment (QMRA) has been used to assess health risks

associated with waterborne pathogens in drinking water and treatment requirements

36

under the Surface Water Treatment Rule (Gerba et al. 1996; Haas et al. 1996; Rose et al.
1991)

Risk assessment models of several microbial pathogens in drinking water and
recreational water have been developed (Haas et al. 1993; Haas et al. 1999; Regli et al.
1991; Rose et al. 1991; van Heerden et al. 2005). Data gaps in previous studies included
the lack of actual virus monitoring and survival data. Most virus risk assessments use
virus concentration data from the literature and rarely include actual virus monitoring
data. A basic QMRA fimnework suggested by the National Research Council (NRC)
includes four steps:

1. Hazard identification (identify types of pathogens and description of illnesses
caused by the pathogens, hospitalization, and mortality rate associated with diseases
caused).

2. Dose-response assessment (quantitative relationship between dose and outcome
described as a probability).

3. Exposure assessment (prevalence, concentrations, distribution of a particular
pathogen in time and space in contaminated water or food consumed).

4. Risk characterization (the quantitative likelihood of potential adverse health
outcome based on the above).

Dose-response of viral pathogens in human subjects has not been widely studied
because of high cost associated with these experiments, ethical issues as well as difficulty
in culturing and administering desired dose (most enteric viruses are non cultivable).
Viruses with a published dose response relation in either human or animals include

adenovirus, coxsackievirus, echovirus, poliovirus, and rotavirus (Gerba et al. 1996;

37

 

Crabtree et al. 1997; Haas et al. 1999; Mena et al. 2003). Dose-response for rotavirus is
the most widely-used model for assessing health risk through the ingestion route because
of the high quality human data set and its high potency. Dose-response for adenovirus is
the only model for virus exposure through inhalation of contaminated aerosol. These
dose-response models have been used to calculate probability of infection via various

exposure pathways including recreational exposure (Mena 2002).

Hypotheses and Objectives

Adenoviruses may be one of the best viral pathogens for monitoring fecal
contamination of water and evaluating public health risk because of their high
concentrations and survival in recreational water mentioned previously. Adenoviruses
have been detected in Great Lakes waters and an adenovirus probability of infection
model has been developed (Crabtree et al. 1997 ; Pong et al. 2007; Xagoraraki et al.
2007). To date, however, no systematic examination of adenovirus and the public health
risks from polluted waters have been undertaken in freshwater systems. Thus there is a
need for development of sensitive and specific methods for detecting adenoviruses in
water, and addressing the levels of adenovirus contamination in Great Lakes waters for
the interpretation of the risk of waterborne diseases.

The overall objective of this research was to develop an understanding of the
human health risks posed to people in waterways and beaches in Michigan associated
with the presence of viral pathogens, particularly adenoviruses, from sewage inputs.
Through field-based experiments and transport models, a risk assessment was developed

for recreating at Great Lake beaches. In this study, dose-response probability of infection

38

model for rotavirus (representing high risk) and dose-response for echovirus 12
(representing low risk) coupled with actual virus monitoring data from sewage, CSO and
surface water were used to compare health risks associated with contamination sources
for the Great Lakes beaches.

The main hypotheses were: 1) Human adenoviruses as important waterborne
pathogens, are present consistently in sewage and rivers receiving sewage effluents in
Michigan; 2) Wastewater treatment does not efficiently remove adenoviruses (or their
DNA); 3) virus transport and attenuation in a river can be assessed by running a tracer
study using a bacteriophage (i.e. P22) as a surrogate for waterborne viruses; 4) the risks
. for adenovirus exposure at Great Lakes beaches are elevated to above an acceptable risk
for recreational waters during the swimming season.

The four specific objectives of this research were 1) to evaluate wastewater
treatment efficiency for removing adenoviruses; 2) to study the prevalence of adenovirus
species in wastewater and fecally-contaminated surface water; 3) to use bacteriophage
P22 as a biological tracer and viral surrogate, and to study the transport of this viral
surrogate in surface water; 4) to develop a risk assessment model for recreational use of
the lower Grand River and beaches at Lake Michigan based on the virus prevalence data
and results fiom the tracer study.

To meet these specific goals, in this research:

a) adenovirus distribution and loading in sewage contaminated water (i.e. influent and
effluent of a wastewater treatment plant, river and beaches influenced by wastewater

discharge) using real-time PCR was determined [Chapter 2].

39

b) the transport behavior of viral contaminants in the lower Grand River in Grand

Rapids, M1 to Lake Michigan by using a viral tracer (bacteriophage P22) compared to
a fluorescent tracer was evaluated [Chapter 3].

adenovirus loading data, virus inactivation and attenuation data from the tracer study
was used in a risk assessment model to estimate public health risks associated with
recreational activities, such as wading, fishing, boating and swimming in the lower

Grand River and beaches on Lake Michigan [Chapter 4].

4O

References

Albinana-Gimenez, N., P. Clemente-Casares, S. Bofill-Mas, A. Hundesa, F. Ribas and R.
Girones (2006). "Distribution of human polyomaviruses, adenoviruses, and

hepatitis E virus in the environment and in a drinking-water treatment plan ."
Environ. Sci. Techno]. 40(23): 7416-7422.

Allard, A., B. Albinsson and G. Wadell (1992). "Detection of adenoviruses in stools from
healthy persons and patients with diarrhea by two-step polymerase chain
reaction." J Med. Virol. 37(2): 149-157.

Babich, H. and G. Stotzky (1980). "Reductions in inactivation rates of bacteriophages by
clay minerals in lake water." Water Res. 14(2): 185-187.

Battigelli, D. A., M. D. Sobsey and D. C. Lobe. (1993). "The inactivation of hepatitis A
virus and other model viruses by UV irradiation." Water Sci. Technol. 27: 339-
342.

Bitton, G. (1980). Introduction to Environmental Virology. New York, Wiley-
Interscience.

Blanc, R. and A. Nasser (1996). "Effect of effluent quality and temperature on the
persistence of viruses in soil." Mter Sci. Technol. 33(10-11): 237-242.

Blanford, W. J ., M. L. Brusseau, T. C. Jim Yeh, C. P. Gerba and R. Harvey (2005).
"Influence of water chemistry and travel distance on bacteriophage PRD-l
transport in a sandy aquifer." Water Res. 39(11): 2345-2357.

Bofill-Mas, S., N. Albinana-Gimenez, P. Clemente-Casares, A. Hundesa, J. Rodriguez-
Manzano, A. Allard, M. Calvo and R. Girones (2006). "Quantification and

stability of human adenoviruses and polyomavirus JCPyV in wastewater
matrices." Appl. Environ. Microbiol. 72(12): 7894-7896.

Borchardt, M. A., P. D. Bertz, S. K. Spencer and D. A. Battigelli (2003). "Incidence of
enteric viruses in groundwater from household wells in Wisconsin." A l.
Environ. Microbiol. 69(2): 1 172-1 180.

Borchardt, M. A., N. L. Haas and R. J. Hunt (2004). "Vulnerability of drinking-water
wells in La Crosse, Wisconsin, to enteric virus contamination from surface water
contributions." Appl. Environ. Microbiol. 70(10): 5937-5946.

41

Bordalo, A. A., R. Onrassami, C. Dechsakulwatana (2002). "Survival of faecal indicator
bacteria in tropical estuarine waters (Bangpakong River, Thailand)" J Appl.
Microbiol. 93(5): 864-871.

Borrego, J. J ., M. A. Morinigo, A. de Vicente, R. Comax and P. Romero (1987).
"Coliphages as an indicator of faecal pollution in water. Its relationship with
indicator and pathogenic microorganisms." Water Res. 21(12): 1473-1480.

Bosch, A. (1998). "Human enteric viruses in the water environment: a minireview." Int,
Microbiol. 1: 191-196.

Brown, M. (1990). "Laboratory identification of adenoviruses associated with
gastroenteritis in Canada fi‘om 1983 to 1986." J Clin. Microbiol. 28(7): 1525-
l 529.

Bitchen-Osmond, C. (2002). "52.0.1. Enterovirus." ICTVdB virus description.
http://www.ncbi.n1m.nih.gov/ICTde/ICTVdB/52010000.htm. 2/1/2004.

Byappanahalli, M. N. and R. S. F uj ioka. (1998). "Evidence that tropical soil can support
the growth of Escherichia coli." Water Sci. Technol. 38: 171-174.

Castignolles, N., F. Petit, I. Mendel, L. Simon, L. Cattolico and C. Buffet-Janvresse
(1998). "Detection of adenovirus in the waters of the Seine River estuary by
nested-PCR." Mol. Cell. Probes 12(3): 175-180.

Chapron, C. D., N. A. Ballester, J. H. Fontaine, C. N. Frades and A. B. Margolin (2000).
"Detection of astroviruses, enteroviruses, and adenovirus types 40 and 41 in
surface waters collected and evaluated by the information collection rule and an
integrated cell culture-nested PCR procedure." Appl. Environ. Microbiol. 66(6):
2520-2525.

Cho, H. B., Seung-Hoon Lee, J ang-Cheon Cho and SJ. Kim (2000). "Detection of
adenoviruses and enteroviruses in tap water and river water by reverse
transcription multiplex PCR." Q_an. J Mic_robiol. 46(5): 417-424.

Choi, S. and S. C. J iang (2005). "Real-time PCR quantification of human adenoviruses in
urban rivers indicates genome prevalence but low infectivity." Appl. Environ.
Microbiol. 71(11): 7426-7433.

42

Colford, J. M. J ., Timothy J. Wade, Kenneth C. Schiff, Catherine C. Wright, John F.
Griffith, Sukhminder K. Sandhu, Susan Burns, Mark Sobsey, Gerg Lovelace and
S. B. Weisberg (2007). "Water quality indicators and the risk of illness at beaches
with nonpoint sources of fecal contamination." Epidemiology. 18(1): 27-35.

Crabtree, K. D., C. P. Gerba, J. B. Rose and C. N. Haas (1997). "Waterbome adenovirus:
a risk assessment." Water Sci. Technol. 35(11-12): 1-6.

Curriero, F. C., J. A. Patz, J. B. Rose and S. Lele. (2001). "The association between
extreme precipitation and waterborne disease outbreaks in the United States,
1948-1994." Am. J Public Hlth. 91(8): 1194-1199.

Davis, J. A., S. R. Farrah and A. C. Wilkie. (2006). "Adsorption of viruses to soil: impact
of anaerobic treatment." Elater Sci. Technol. 54(3): 161-167.

De Leon, R., C. Shieh, R. S. Baric, & M. D. Sobsey (1990). Detection of enterovirufl
and hepatitis A viru_s in environmentJal_§_amples by gene probes and polymerase
chain reaction. Proceedings of the 1990 Water Quality Technology Conference,
Denver, CO, American Water Works Association.

Deetz, T. R., E. M. Smith, S. M. Goya], C. P. Gerba, I. Vollet, John J ., L. Tsai, H. L.
DuPont and B. H. Keswick (1984). "Occurrence of rota- and enteroviruses in
drinking and environmental water in a developing nation." Water Res. 18(5): 567-
571.

Donaldson, K. A., D. W. Griffin and J. H. Paul (2002). "Detection, quantitation and
identification of enteroviruses from surface waters and sponge tissue from the
Florida Keys using real-time RT-PCR." Water Res. 36(10): 2505-2514.

Dorfinan, M and N. Stoner. (2007). Testing the waters 2007: A guide to water quality at
vacation beaches. http://www.nrdc.org/water/oceans/ttw/ttw2007.pdf

Dziuban, E. J ., J. L. Liang, G. F. Craun, V. Hill, P. A. Yu, J. Painter, M. R. Moore, R. L.
Calderon, S. L. Roy, M. J. Beach and Centers for Disease Control and Prevention
(2006). "Surveillance for waterborne disease and outbreaks associated with
recreational water--United States, 2003-2004." MMWR Surveill. Summ.
55(s312): 1-30.

43

Edly, K. and J. Wuycheck. (2006). Water Quality and Pollution Control in Michigan:
2006 Sections 303(d) and 305(b) Integrated Report. MDEQ Report No.
MI/DEQ/WB-06/Ol9.

Embrey, M. A., R. T. Parkin and J. M. Balbus (2002). H_andboci of CCL microbes in
drinking water. Denver, CO, American Water Works Association.

Enriquez, C. E., C. J. Hurst and C. P. Gerba (1995). "Survival of the enteric adenoviruses
40 and 41 in tap, sea, and waste water." Water Res. 29(11): 2548-2553.

Fong, T.-T., D. W. Griffin and E. K. Lipp (2005). "Molecular assays for targeting human
and bovine enteric viruses in coastal waters and their application for library-
independent source tracking." Appl. Environ. Microbiol. 71 (4): 2070-2078.

Fong, T., L. S. Mansfield, D. L. Wilson, D. J. Schwab, S. L. Molloy and J. B. Rose.
(2007). "Massive microbiological groundwater contamination associated with a
waterborne outbreak in Lake Erie, South Bass Island, OH." Environ Health
Peggect. 115(6): 856-864.

Foppen, J. W. A., S. Okletey and J. F. Schijven (2006). "Effect of goethite coating and
humic acid on the transport of bacteriophage PRDl in columns of saturated san ."
J Contain. Hvdrol. 85(3-4): 287-301.

Gerba, C. P., D. M. Gramos and N. Nwachuku (2002). "Comparative inactivation of
enteroviruses and adenovirus 2 by UV light." Appl. Environ. Microbiol. 68(10):
5167-5 169.

Gerba, C. P., J. B. Rose, C. N. Haas and K. D. Crabtree (1996). "Waterbome rotavirus: a
risk assessment." Water Res. 30(12): 2929-2940.

Gerba, C. P. and G. E. Schaiberger (1975). "Effect of particulates on virus survival in
seawater." J Water Pollut. Control Fed. 47(1): 93-103.

Girones, R., M. Puig, A. Allard, F. Lucena, G. Wadell and J. Jofre (1995). "Detection of
adenovirus and enterovirus by PCR amplification in polluted waters." Water Sci.
Technol. 31(5-6): 351-357.

Glass, R. 1., J. Bresee, B. J iang, J. Gentsch, T. Ando, R. Fankhauser, J. Noel, U. Parashar,
B. Rosen and S. S. Monroe (2001). Gastroenteritis viruses: an overview.

Symposium on Gastroenteritis viruses, Novartis Foundation Symposia. 238: 5-
19/19-25.

 

Government of Canada and US. EPA (1995). The Grth Lakes: An Environmental Atlas
and Resource Bogghttp://www.epa.gov/glnpo/atlas/index.html

Green, D. H. and G. D. Lewis (1999). "Comparative detection of enteric viruses in
wastewaters, sediments and oysters by reverse transcription PCR and cell
culture." Water Res. 33(5): 1 195-1200.

Griffith, J. H., S. B. Weisberg and C. D. McGee (2003). "Evaluation of microbial source
tracking methods using mixed fecal sources in aqueous test samples." J Water
Hlth. 1(4): 141-151.

On, 2., S. W. Belzer, C. S. Gibson, M. J. Bankowski and R. T. Hayden (2003).
"Multiplexed, real-time PCR for quantitative detection of human adenovirus." J
Clin. Microbiol. 41(10): 4636-4641.

Haas, C., C. S. Crockett, J. B. Rose, C. P. Gerba and A. M. Fazil. (1996). "Assessing the
risk posed by oocysts in drinking water." J AWWA. 88(9): 131-136.

Haas, C., J. Rose and C. Gerba (1999). Quantitative microbimisk assessment. New
York, NY, John Wiley and Sons.

Haas, C. N., J. B. Rose, C. P. Gerba and R. Regli (1993). "Risk assessment of viruses in
drinking water." Risk Anal. 13: 545-552.

Haramoto, E., Hiroyuki Katayama, Kumiko Oguma, Hiromasa Yamashita, Eiichiro
Nakaj irna and S. Ohgaki (2005). "One-year monthly monitoring of Torque teno
virus (TTV) in wastewater treatment plants in J apan." Water Res. 39: 2008-2013.

Haramoto, E., Hiroyuki Katayama, Kumiko Oguma and S. Ohgaki (2007). "Quantitative
analysis of human enteric adenoviruses in aquatic environments." J A l.
Microbiol. 103(6): 2153-2159.

Haramoto, E., H. Katayama and S. Ohgaki (2004). "Detection of noroviruses in tap water
in Japan by means of a new method for concentrating enteric viruses in large
volumes of freshwater." Appl. Environ. Microbiol. 70(4): 21 54-2160.

45

He, J .-W. and S. J iang (2005). "Quantification of enterococci and human adenoviruses in
environmental samples by real-time PCR." Appl. Environ. Microbiol. 71(5): ‘
2250-2255.

Hoebe, C. J. P. A., H. Vennema, A. M. de Roda Husman and Y. T. H. P. van Duynhoven
(2004). "Norovirus outbreak among primary schoolchildren who had played in a
recreational water fountain." J Infect. Dis. 189(4): 699-705.

Hrudey, S., P Payment, PM Huck, RW Gillham and E. Hrudey. (2003). "A fatal
waterborne disease epidemic in Walkerton, Ontario: comparison with other
waterborne outbreaks in the developed world." _Vfluer Sci. Technol. 47(3): 7-14.

Hundesa, A., C. Maluquer de Motes, S. Bofill-Mas, N. Albinana-Gimenez and R.
Girones (2006). "Identification of human and animal adenoviruses and
polyomaviruses for determination of sources of fecal contamination in the
environment." Appl. Environ. Microbiol. 72( 12): 7886-7893.

Hurst, C. J ., C. P. Gerba and I. Cech (1980). "Effects of environmental variables and soil
characteristics on virus survival in soil." Appl. Environ. Microbiol. 40(6): 1067-
1079.

Jenkins, T. M., T. M. Scott, M. R. Morgan and J. B. Rose (2005). "Occurrence of
alternative fecal indicators and enteric viruses in Michigan rivers." J Great Lakes
33$, 31: 22-31.

Jensen, M. (1964). "Inactivation of airborne viruses by ultraviolet irradiation." A l.
Microbiol. 12(5): 418-420.

J iang, S., R. Noble and W. P. Chui (2001). "Human adenoviruses and coliphages in urban
runoff-impacted coastal waters of Southern California." Appl. Environ.
Microbiol. 67(1): 179-184.

J iang, S. C. (2006). "Human adenoviruses in water: occurrence and health implications: a
critical review." Environ. Sci. Technol. 40(23): 7132-7140.

Jones, M. S., H, B. Harrach, R. D. Ganac, M. M. A. Gozum, W. P. dela Cruz, B. Riedel,
C. Pan, E. L. Delwart and D. P. Schnurr (2007). "New adenovirus species found
in patient presenting with gastroenteritis." J Virol. 81(11): 5978-5984.

 

46

KOCW;

Kocw

LeGu

Lee,S

Mar

1632‘

U. L.

Kocwa-Haluch, R. (2001). "Waterbome enteroviruses as a hazard for human health."
Polfih J Environ. Studies 10(6): 485-487.

Kocwa-Haluch, R. and B. Zalewska (2002). "Presence of Rotavirus hominis in sewage
and water. " Polish J Environ. Studies 11(6): 751-755.

Langlet, J ., F. Gaboriaud and C. Gantzer (2007). "Effects of pH on plaque forming unit
counts and aggregation of M82 bacteriophage." J Appl. Microbiol. 103(5): 1632-
1638.

Le Guyader, F., F. H. Neill, M. K. Estes, S. S. Monroe, T. Ando and R. L. Atrnar (1996).
"Detection and analysis of a small round-structured virus strain in oysters

implicated in an outbreak of acute gastroenteritis." Appl. Environ. Microbiol.
62(1 1): 4268-4272.

Lee, S. H. and S. J. Kim (2002). "Detection of infectious enteroviruses and adenoviruses
in tap water in urban areas in Korea." Water Res. 36(1): 248-256.

Lefler, E. and H. Kott (1974). Virus retention and survival in sand. In J. F. Malina and B.
P. Sagik (Eds), Virus survival in water and wpstewater systems, p. 84-91. Austin,
Texas: Center for Research in Water Resources.

Ley, V., J. Higgins and R. Fayer (2002). "Bovine enteroviruses as indicators of fecal
contamination." Appl. Environ. Microbiol. 68(7): 3455-3461.

Li, L., H. Shimizu, L. T. P. Doan, P. G. Tung, S. Okitsu, O. Nishio, E. Suzuki, J. K. Seo,
K. S. Kim, W. E. G. Muller and H. Ushijirna (2004). "Characterizations of
adenovirus type 41 isolates fi'om children with acute gastroenteritis in Japan,
Vietnam, and Korea." J Clin. Microbiol. 42(9): 4032-4039.

Liang, J. L., E. J. Dziuban, G. F. Craun, V. Hill, M. R. Moore, R. J. Gelting, R. L.
Calderon, M. J. Beach, S. L. Roy and Centers for Disease Control and Prevention
(2006). "Surveillance for waterborne disease and outbreaks associated with
drinking water and water not intended for drinking-United States, 2003-2004."
MMWR Surveill. Surnm. 55(12): 31-65.

Lipp, E. K., R. Kurz, R. Vincent, C. Rodriguez-Palacios, S. R. Farrah and J. B. Rose
(2001). "The effects of seasonal variability and weather on microbial fecal

pollution and enteric pathogens in a subtropical estuary." Estuaries 24(2): 266-
276.

47

Lipson, S. M. and G. Stotzky (1986). "Effect of kaolinite on the specific infectivity of
reovirus." Fem_s. Microbiol. Letters 37(1): 83-88.

Lorenzo-Morales, J ., N. Coronado-Alvarez, E. Martinez-Carretero, S. K. Maciver and B.
Valladares (2007). "Detection of four adenovirus serotypes within water-isolated
strains of Acanthamoeba in the Canary Islands, Spain." Am J Trop. Med. Hyg,
77(4): 753-756.

Mac Kenzie, W. R., N. J. Hoxie, M. E. Proctor, M. S. Gradus, K. A. Blair, D. E.
Peterson, J. J. Kazmierczak, D. G. Addiss, K. R. Fox, J. B. Rose and J. P. Davis
(1994). "A massive outbreak in Milwaukee of Cryptosporidium infection
transmitted through the public water supply." N Engg Med. 331(3): 161-167.

Maluquer de Motes, C., P. Clemente-Casares, A. Hundesa, M. Martin and R. Girones
(2004). "Detection of bovine and porcine adenoviruses for tracing the source of
fecal contamination." Appl. Environ. Microbiol. 70(3): 1448-1454.

Maunula L, Ilkka T. Miettinen and C.-H. v. Bonsdorff. (2005). "Norovirus outbreaks
from drinking water." Emerg. Infect. Dis. 1 1(1 1): 1716-1721.

McQuaig, S. M., T. M. Scott, V. J. Harwood, S. R. Farrah and J. O. Lukasik (2006).
"Detection of human-derived fecal pollution in environmental waters by use of a

PCR-based human polyomavirus assay." Appl. Environ. Microbiol. 72(12): 7567-
7574.

MDEQ (2008). Combined Sewer Overflow (CSO) & Sanitary Sewer Overflow (S80)
2007 Annual Report (January 1, 2007 — December 31, 2007).
http://www.michigan.gov/documents/deq/deq-wb-csossoreport07_243725_7.pdf

Mena, K. D. (2002). Environmental exposure to viruses: a human health risk assessment
methodology. In G. Bitton (Ed), The Encyclopedia of Environmenta_]
Microbiology, p. 2741-2752. New York, John Wiley & Sons, Inc.

Mena, K. D., C. P. Gerba, C. N. Haas and J. B. Rose (2003). "Risk assessment of
waterborne coxsackievirus." J AWWA 95(7): 122-131.

Meng, Q. S. and C. P. Gerba. (1996). "Comparative inactivation of enteric adenovirus,
poliovirus and coliphages by ultraviolet irradiation." Water Res. 30: 2665-2668.

48

Moe, C., J. Gentsch, T. Ando, G. Grohmann, S. Monroe, X. Jiang, J. Wang, M. Estes, Y.
Seto and C. Humphrey (1994). "Application of PCR to detect Norwalk virus in
fecal specimens from outbreaks of gastroenteritis." J Clin. Microbiol. 32(3): 642-
648.

Murata, T., N. Katsushima, K. Mizuta, Y. Muraki, S. Hongo and Y. Matsuzaki. (2007).
"Prolonged norovirus shedding in infants <or= 6 months of age with
gastroenteritis." Pedgtr. Infect. Dis. J 26(1): 46-49.

Muscillo, M., M. Pourshaban, M. Iaconelli, S. Fontana, A. Di Grazia, S. Manzara, G.
Fadda, R. Santangelo and G. La Rosa (2008). "Detection and Quantification of
Human Adenoviruses in Surface Waters by Nested PCR, TaqMan Real-Time
PCR and Cell Culture Assays." Water Air & Soil Pollut. 191(1): 83-93.

 

Myrmel, M., E. M. M. Berg, E. Rimstad and B. Grinde (2004). "Detection of enteric
viruses in shellfish fi'om the Norwegian coast." Appl. Environ. Microbiol. 70(5):
2678-2684.

NIH (2006). "Viral gastroenteritis."
http://digestive.niddk.nih.gov/ddiseases/pubs/viralgastroenteritis/index.htrn

Noble, R. T. and J. A. Fuhrman (1999). "Breakdown and microbial uptake of marine
viruses and other lysis products." Aquat. Microb. Ecol. 20(1): 1-11.

Nwachcuku, N. and C. P. Gerba (2004). "Emerging waterborne pathogens: can we kill
them all?" Curr. Opin. Biotech. 15(3): 175-180.

O'Reilly, C. E., Anna B. Bowen, Nytzia E. Perez, John P. Sarisky, Craig A. Shepherd,
Mark D. Miller, Brian C. Hubbard, Michael Herring, Sharunda D. Buchanan,
Collette C. Fitzgerald, Vincent Hill, Michael J. Arrowood, Lihua X. Xiao, R.
Michael Hoekstra, Eric D. Mintz, Michael F. Lynch and Outbreak Working
Group (2007). "Waterbome outbreak of gastroenteritis with multiple etiologies
among resort island visitors and residents: Ohio, 2004." Clin. Infect. Dis. 44: 506-
5 12.

ODH (2005). Investigation of the ground water quality of South Bass Island, Ottawa county,
Ohio. Columbus, OH: Ohio Department of Health.

OH EPA, 2005. South Bass Island, Ottawa county gastrointestinal illness summer 2004.
Columbus, OH: Ohio Environmental Protection Agency. Amended Feb 16, 2006.

49

Okamoto, H., Yoshihiro Akahane, Masato Ukita, Masako Fukuda, Fumio Tsuda, Yuzo
Miyakawa and M. Mayumi (1998). "Fecal excretion of a nonenveloped DNA
virus (TTV) associated with posttransfusion non-A-G hepatitis." J Med. Virol.
56(2): 128-132.

Parshionikar, S. U., S. Willian-T rue, G. S. Fout, D. E. Robbins, S. A. Seys, J. D. Cassady
and R. Harris (2003). "Waterbome outbreak of gastroenteritis associated with a
norovirus." Appl. Environ. Microbiol. 69(9): 5263-5268.

Pesaro, F ., I. Sorg and A. Metzler (1995). "In-situ inactivation of animal viruses and a
coliphage in nonaerated liquid and semiliquid animal wastes." Appl. Environ.
Microbiol. 61(1): 92-97.

Pina, S., M. Puig, F. Lucena, J. Jofre and R. Girones (1998). "Viral pollution in the
environment and in shellfish: human adenovirus detection by PCR as an index of
human viruses." Appl. Environ. Microbiol. 64(9): 3376-3382.

Puig, M., J. Jofre, F. Lucena, A. Allard, G. Wadell and R. Girones (1994). "Detection of
adenoviruses and enteroviruses in polluted waters by nested-PCR amplification."

Appl. Environ. Microbiol. 60(8): 2963-2970.

Rabinovici, S. J. M., R. L. Bernknopf, D. L. Coursey, A. M. Wein and R. L. Whitman
(2004). "Economic and health risk trade-offs of swim closures at a Lake Michigan
beach." Environ. Sci. Technol. 38(10): 2737-2745.

Rajal, V. B., B.S. McSwain, D.E. Thompsona, C.M. Leuteneggerb and S. Wuertz (2007).
"Molecular quantitative analysis of human viruses in California stormwater."
Water Res. 41: 4287-4298.

Rao, V. C., J. L. Mehiick (1986). Environmental biology. In J. A. Cole, C.J. Knowles, D.
Schlessinger (Eds), Asgcts of Microbiology. Washington, D. C., American
Society of Microbiology.

Regli, S., J. B. Rose, C. N. Haa and C. P. Gerba. (1991). "Modeling the risk from Giardia
and viruses in drinking water." J AWWA. 83: 76-84.

Roessler, P. F. and B. F. Severin (1996). Ultraviolet light disinfection of water and
wastewater. In C. J. Hurst (Ed), Modeling disease transmission and its prevention
by disinfection. p. 313-368, Cambridge, UK, Cambridge University Press.

50

Rose, J ., C. Haas and S. Regli (1991). "Risk assessment and control of waterborne
giardiasis." Am J Public Hlth Assoc. 81(6): 709-713.

Rose, J. B. and D. J. Grimes (2001). Reevaluation of microbial water quality: powerful
new tools for detection and risk assessment. Washington, DC, USA: American
Academy of Microbiology.

Ryan, J. N., R. W. Harvey, D. Metge, M. Elimelech, T. Navigato and A. P. Pieper (2002).
"Field and laboratory investigations of inactivation of viruses (PRDl and MS2)
attached to iron oxide-coated quartz sand." Environ Sci Technol 36(11): 2403-
2413.

Schijven, J. F. and M. Hassanizadeh (2000). "Removal of viruses by soil passage:
overview of modeling, processes, and parameters." Crit. Rev. E_nviron. Sci.
Technol. 30(1): 49-127.

Shanna, S., P. Sachdeva and J. S. Virdi (2003). "Emerging water-bome pathogens."
App]. Microbiol. Biotech. 61(5-6): 424-428.

Shieh, Y.-S. C., R. S. Baric, and M. D. Sobsey (1997). "Detection of low levels of enteric
viruses in metropolitan and airplane sewage." Appl. Environ. Microbiol. 63:
4401-4407.

Soares, C. C., Eduardo M. Volotao, Maria Carolina M. Albuquerque, Fabiano M. da
Silva, Tatiane R. B. de Carvalho, Carlos M. Nozawa, Rosa E. Linhares and N.

Santos (2002). "Prevalence of enteric adenoviruses among children with diarrhea
in four Brazilian cities." J Clin. Virol. 23(3): 171-177.

Solic, M., and N. Krstulovic (1992). "Separate and combined effects of solar radiation,
temperature, salinity, and pH on the survival of faecal coliforrns in seawater."
l\_4_ar. Pol]. Bull. 24: 411-416.

Solo-Gabriele, H. M., M. A. Wolfert, T. R. Desmarais and C. J. Palmer (2000). "Sources
of Escherichia coli in a Coastal Subtropical Environment." Appl. Environ.
Microbiol. 66(1): 230-237.

Song, I., C. Y. Choi, S. O'Shauglmessy and C. P. Gerba (2005). "Effects of temperature
and moisture on coliphage PRD-l survival in soil." J Food Protect. 68(10): 2118-
2122.

51

Springthorpe, V. S., C. L. Loh, W. J. Robertson, S. A. Sattar (1993). "In situ survival of
indicator bacteria, MS-2 phages and human pathogenic viruses in river water. "
Met Sci. Technol. 27(3-4): 413-420.

Straub, T. M., Ian L. Pepper and C. P. Gerba (1992). "Persistence of viruses in desert
soils amended with anaerobically digested sewage sludge." Appl. Environ.
Microbiol. 58(2): 636-641.

Theron, J. and T. E. Cloete (2002). "Emerging waterborne infections: contributing
factors, agents, and detection tools." Crit. Rev. Microbiol. 28(1): 1-26.

Thurston-Enriquez, J. A., C. N. Haas, J. J acangelo, K. Riley and C. P. Gerba (2003).
"Inactivation of feline calicivirus and adenovirus type 40 by UV radiation." A l.
Environ. Microbiol. 69(1): 577-582.

Tsai, Y. L., B. Tran and C. J. Pahner (1995). "Analysis of viral RNA persistence in
seawater by reverse transcriptase-PCR." Appl. Environ. Microbiol. 61(1): 363-
366.

Turcios, R. M., M.-A. Widdowson, A. C. Sulka, P. S. Mead and R. 1. Glass (2006).
"Reevaluation of epidemiological criteria for identifying outbreaks of acute
gastroenteritis due to norovirus: United States, 1998-2000." Clin Infect Dis 42(7):
964-969.

US. EPA. (1998). "Drinking water contaminant candidate list." Notice. Fed. Regplat.:
63:10274—10287.

US. EPA (2002). Great Lakes strategy 2002: a plan for the new millennium.
http://www.epa.gov/glnpo/gls/index.html

US. EPA and Environment Canada (2007). State of the Great Lakes 2007-highlights.
http://binational.net/so]ec/English/sog12007highlights_en.pdf

Vaidya, S. R., S. D. Chitambar and V. A. Arankalle (2002). "Polymerase chain reaction-
based prevalence of hepatitis A, hepatitis E and TT viruses in sewage from an
endemic area." J Hepatol. 37(1): 131-136.

52

van der Poel, W. H. M., Jan Vinje, Reina van der Heide, Marion P. G. Koopmans, Maria-
Inmaculada Herrera and A. Vivo. (2000). "Norwalk-like calicivirus genes in farm
animals." Emerg. Infect. Dis. 6(1): 3641.

van Elsen, A. and A. Boyce (1966). "Disruption of type 1 poliovirus under alkaline
conditions: role of pH, temperature and sodium dodecyl sulfate (SDS)." Virol.
28(3): 481-483.

van Heerden, J ., M. M. Ehlers and W. O. K. Grabow (2005). "Detection and risk
assessment of adenoviruses in swimming pool water." J Appl. Microbiol. 99(5):
1256-1264.

van Heerden, J ., M. M. Ehlers, J. C. Vivier and W. O. K. Grabow (2005). "Risk
assessment of adenoviruses detected in treated drinking water and recreational
water." J Appl. Microbiol. 99(4): 926-933.

Vantarakis, A. C. and M. Papapetropoulou (1998). "Detection of enteroviruses and
adenoviruses in coastal waters of SW Greece by nested polymerase chain
reaction." Water Res. 32(8): 2365-2372.

Wadell, G. A., A. Allard, M. Johansson, L. Svensson and I. Uhnoo. (1987). Enteric
adenoviruses. In Enteric aydenoviruses, p.73-91, Chichester, United Kingdom:
John Wiley & Son, Inc.

Walls, T ., A. G. Shankar and D. Shingadia (2003). "Adenovirus: an increasingly
important pathogen in pediatric bone marrow transplant patients." Lancet Infect.
Dis. 3: 79-86.

Wetz, J. J ., E. K. Lipp, D. W. Griffin, J. Lukasik, D. Wait, M. D. Sobsey, T. M. Scott and
J. B. Rose (2004). "Presence, infectivity, and stability of enteric viruses in
seawater: relationship to marine water quality in the Florida Keys." Mar. Poll.
_Bpl_l, 48(7-8): 698-704.

WHO (2004). World Health Report 2004 - Annex Table 3 Burden of disease in DALY's
by cause, sex and mortality stratum in WHO regions, estimates for 2002.
http://www.who.int/whr/2004/annex/topic/en/annex_3_en.pdf

WHO (2005). World Health Report 2005 - Annex Table 3 Annual number of deaths by
cause for children under five years of age in WHO regions, estimates for 2000-
2003. http://www.who.int/whr/ZOOS/annex/annex3-4.xls

53

Whitman, R. L., D. A. Shively, H. Pawlik, M. B. Nevers and M. N. Byappanahalli
(2003). "Occurrence of Escherichia coli and Enterococci in Cladophora
(Chlorophyta) in Nearshore Water and Beach Sand of Lake Michigan." A l.
Environ. Microbiol. 69(8): 4714-4719.

Widdowson, M.-A., E. H. Crarner, L. Hadley, J. S. Bresee, R. S. Beard, S. N. Bulens, M.
Charles, W. Chege, E. Isakbaeva, J. G. Wright, E. Mintz, D. Fomey, J. Massey,
R. 1. Glass, and S. S. Monroe (2004). "Outbreaks of acute gastroenteritis on cruise
ships and on land: identification of a predominant circulating strain of norovirus--
United States, 2002." J Infect. Dis. 190: 27-36.

Wigand, R., H. G. Baumeister, G. Maass, J. Kuhn and H. J. Hammer. (1983). "Isolation
and identification of enteric adenoviruses." J Med. Virol. 11: 233—240.

Wilson, B., P. Roessler, E. V. Dellen, M. Abbaszadegan and C. P. Gerba (1993).
Coliphage MS-2 M1 UV water disinfection efficacy test surrogate for bacteria]
and viral pathogens. Proceedings 1992 Water Quality Technology Conference,
Toronto, CN, American Water Works Assoc, Denver, Co.

Xagoraraki, I., D. H. W. Kuo, K. Wong, M. Wong and J. B. Rose (2007). "Occurrence of
human adenoviruses in two Great Lake recreational beaches." 133p]. Environ.
Microbiol.73 :7874-788 1 .

Yahya, M. T., L. Galsomies, C. P. Gerba and R. C. Bales (1993). "Survival of
bacteriophages MS-2 And PRD-l in ground-water." Mter Sci. Technol. 27(3-4):
409-412.

Yates, M. V. and C. P. Gerba (1985). "Factors controlling the survival of viruses in
groundwater." Water Sci. Technol. 17: 681-687.

Yates, M. V., C. P. Gerba and L. M. Kelley (1985). "Virus persistence in groundwater."
Appl. Environ. Microbiol. 49(4): 778-781.

Yates, M. V., S. R. Yates, J. Wagner and C. P. Gerba (1987). "Modeling virus survival
and transport in the subsurface." J Contiarn. Hydro]. 1(3): 329-345.

Zhuang, J. and Y. J in (2003). "Virus retention and transport as influenced by different
forms of soil organic matter." J Environ. Oua1_. 32(3): 816-823.

54

CHAPTER 2
MONITORING OF HUMAN VIRUSES IN SURFACE WATER AND WASTEWATER

IN MICHIGAN

Abstract

Enteric viruses are important pathogens of surface water and have frequently been
detected from waters of the Great Lakes. Human adenoviruses have been suggested as
index viruses because of their high prevalence and persistence in aquatic environments.
In this study, we aim to quantify adenoviruses in different aquatic environments: waste
water, surface water and combined sewer overflow discharges. Raw sewage (n=13),
primary-treated effluent (n=13), secondary-treated effluent (n=10) and chlorinated
effluent (n=10) were collected from an East Lansing waste water treatment plant between
August 2005 and August 2006. Surface water samples (n=26) were collected from seven
sampling sites at the lower Grand River in Grand Rapids, MI. Combined sewer overflow
samples (n=6) were collected from CSO retention basin in Grand Rapids, MI.
Conventional PCR and real-time PCR assay were used for virus detection and
quantification. Adenoviruses were detected in 100% of waste water and CS0 discharge
samples. Average adenovirus DNA concentration in raw sewage and primary sewage
were 1.15 x 106 viruses /L and 1.12 x 10‘5 viruses /L, respectively. Analysis of variance
(ANOVA) shows that virus concentrations in CS0 samples (average: 5.35 x105 viruses
/L) were not significantly different from virus concentration in raw sewage and primary-
treated effluent (p-value: 0.39). In addition, adenovirus removal was less than 2 logo

(99%) at the East Lansing wastewater treatment plant. Adenovirus type 41 (60%), type

55

12 (29%), type 40 (3%), type 2 (3%) and type 3 (3%) were predominant in raw sewage
and primary effluents (n=28). Multiple adenovirus serotypes were detected in six
samples. Six of twenty surface water samples analyzed by real-time PCR and showed
virus concentration above detection limit (average: 7.8 x103 viruses /L). This research
demonstrated that wastewater effluents and surface water in Michigan contain high levels
of viruses and may not be suitable for full-body recreational activities. High
concentration of adenoviruses in these waters may be due to inefficient removal during

wastewater treatment and high persistence of viruses in the environment.

Introduction

Enteric viruses are important waterborne pathogens. They are fi'equently isolated
from fecal-contaminated water and have been linked to numerous waterborne outbreaks
(Craun 1991; Tani 1995; J iang et al. 2001; Lee and Kim 2002). This group of pathogens
includes adenoviruses, rotavirus, hepatitis A virus, noroviruses and enteroviruses. In the
Great Lakes region, enteric viruses were isolated fi'om recreational beaches and ground
water for municipal usage, indicating elevated public health risk when consuming or
coming into contact with these waters (F ong et al. 2007; Xagoraraki et al. 2007)
Although recent developments in molecular detection assays substantially increases
detection of viruses from waters, from a management standpoint, it is impractical to test
for all viruses when determining microbial quality of water. Here, we proposed using
adenoviruses as an index virus for monitoring water quality.

Adenovirus, which has a high prevalence in water, has been suggested as a

preferred candidate as an index organism for viral pathogens because they fit some of the

56

criteria of an ideal indicator (Pina et a]. 1998; Griffin et al. 2001). Human adenoviruses
(HAst) are present in higher concentrations in sewage than other enteric virus; and it is
estimated that more than 90% of the population are seropositive for adenovirus. An
infected patient may excrete up to 10ll viral particles per gram of feces (D'Ambrosio et
al. 1982; Haramoto et al. 2005; Jiang et al. 2005; Pina et al. 1998; Wadell et al. 1987).

Adenoviruses were first isolated from humans and identified as the causative
agent of epidemic febrile respiratory disease among military recruits in the 19505
(Hilleman 1954; Rowe 1953). Human adenoviruses are the second most important viral
pathogen of infantile gastroenteritis after rotavirus (Basu et al. 2003; Cruz et al. 1990;
Topkaya et al. 2006; Uhnoo et al. 1984). Serotypes of adenoviruses have different tissue
tropisms (i.e. cells and tissues of a host which support growth of a particular virus) and
have been found to cause symptomatic infections in several organ systems, including the
respiratory system (pharyngitis, acute respiratory disease and pneumonia), eye
(conjunctivitis), gastrointestinal tract (gastroenteritis), central nervous system
(meningoencephalitis) and genitalia (urethritis and cervicitis) (Crabtree et a]. 1997;
Kapikian 1992). Human adenovirus types 40 and 41 have been associated with
gastroenteritis in children, while human adenovirus type 4 is linked to persistent
epidemics of acute respiratory disease in the United States (Cruz et al. 1990; McNeil
1999). It was once estimated that 2-7% of all lower respiratory tract illnesses in children
may be caused by adenoviruses (Brandt et a1. 1969; F oy et a]. 1973).

Transmission includes the fecal-oral route and inhalation of aerosols.
Adenoviruses have been associated with outbreaks in different settings, including

military camps (Kolavica-Gray et al. 2002; Chmielewicz et al. 2005; Kajon et al. 2007),

57

hospitals (Chabemy et al. 2003; Hatherill et al. 2004; J alal et al. 2005), and daycare
centers (Akihara et al. 2005; Shimizua et a]. 2007) and schools (Harley et al. 2001).
Adenoviruses waterborne outbreaks have involved swimming in swimming pools (Turner
et al. 1987; Papapetropoulou et al. 1998).

The aim of this part of the research was to evaluate the presence of adenoviruses,
from sewage inputs. Raw sewage, wastewater effluent, CSO discharges and surface water
in the lower Grand River, Michigan were surveyed for the occurrence and concentration
of human adenoviruses. Real-time PCR was used for quantification of adenoviruses.
Predominant adenovirus genotypes in sewage were determined in order to evaluate the
role of surface water contamination as a possible vehicle for the transmission of
adenovirus, and the relevance of adenoviruses as an additional tool in water quality

assessment.

Materials and Methods

Sample collection and concentration. Grab water samples were collected from
the East Lansing wastewater treatment plant approximately once a month between
August 2005 and August 2006. Wastewater treatment processes in the East Lansing
wastewater treatment plant consist of primary treatment (sedimentation), secondary
treatment (aeration, activated sludge and secondary sedimentation) and tertiary treatment
(chlorination, rapid gravity sand filters, dechlorination and post-filtration aeration).

Surface water samples were collected from the lower Grand River in Kent and
Ottawa counties during a one-week intensive study in June 2005 and intermittently
between June 2005 and August 2007. A total of 16 surface water samples were collected

from different sites at the lower Grand River during the intensive study (Figure 2.1).

58

Following the intensive study, several sites were chosen for continued monitoring
because of high bacterial indicator concentrations (data not shown) and detection of
HAdV DNA from two of those sites (Deer Creek and Sixth Street Park) (Figure 2.1). Our
sample collection locations included three beach sites (Rosy Mound, North Shore and
North Beach Park) and four park sites (Riverside Park, Deer Creek Park, Grand River
Park and Sixth Street Park).

Between February and June 2008, six combined sewer samples were collected

from CSS retention basin in Grand Rapids, MI.

   
  
 

 

. Positive site
0 Negative site

Muskegon . '
A Monitored srte

 

 

 

  

Coopersville

 

 

 

Figure 2.1 Sampling sites along the lower Grand River during the intensive study in June

2005, sites that was positive for adenoviruses (0), sites that were negative for adenovirus

(O) and sites selected for continued monitoring between 2005 and 2007 (A).

59

The procedures of Haramoto et a]. (2004) were used for water filtration and
concentration for viruses. Briefly, between 500 ml and two liters of sewage or surface
water were filtered through a 90-mm, type HA, negatively charged membrane (Millipore,
Billerica, Mass.) with 0.45-um pore size. A volume of 100 ml of 0.5 mM HzSO4 was then
passed through the membrane, and viral particles were eluted with 10 ml of 1 mM NaOH.
Eluates were stored in a tube containing 0.1 ml of 50 mM HzSO4 and 0.1 ml of 100x Tris-
EDTA (TE) buffer for neutralization before firrther concentration. All IO-ml eluates were
stored at - 20°C.

For further purification and concentration, eluates were thawed and centrifuged in
Amicon Ultra 100K concentrator columns (Millipore). The final volume of concentrated
eluate recovered for each sample was approximately 200 u]. Concentrates were stored at -
80 °C. Concentrated samples were extracted for viral nucleic acids and purified through
commercial spin columns from Qiagen (QIAGEN, Valencia, CA) following the
manufacturer’s protocol. Purified viral DNA was eluted in 60 u] of RNase-free water.

Detection and quantification of human adenoviruses. A real-time TaqMan
PCR assay was performed for the quantification of HAdV DNA in water samples
following the protocol by Xagoraraki et al. (2007). The forward primer (J TVXF ), reverse
primer (J TVXR) and TaqMan probe (J TVXP) designed by J othilcumar et al. (2005) were
used (Table 2.1) and targeted the hexon gene of HAdV. PCR reaction mixtures contained
10 pM forward primer, 10 11M reverse primer, 10 uM probe, 5 [.11 DNA template, and
PCR-grade water for a total volume of 20 u]. The PCR conditions were as follows: hot-
start denaturation step at 95 °C for 15 min, followed by 50 cycles with a 95 °C

denaturation for 10 s, 55 °C annealing for 30 s and 72 °C elongation for 15 3. All

60

amplification reactions were carried out in duplicate. PCR products were selected
randomly for visualization by gel electrophoresis on a 2 % average strength Omnipur
agarose gel (EM Science, Darmstadt, Germany). The gel was stained with gelStar nucleic
acid stain and viewed under UV light. Both real-time PCR assays were performed in a
Roche LightCycler® 2.0 Instrument (Roche Applied Sciences, Indianapolis, IN). The
samples (i.e., viral DNA extracts) and standards were each run at least in triplicate. All
PCR runs included a negative control reaction (PCR-grade H20 without template) and a
positive control reaction. The crossing point (Cp) of each PCR reaction was
automatically determined by the LightCycler® Software 4.0 and used to calculate the
hexon gene concentration. The concentrations of HAst in the river water samples from
the studied recreational sites was normalized from hexon gene copies per liter to viruses
per liter using a ratio of one (i.e., each HAdV viral particle consists of one copy of hexon
gene as previously reported). The detection limit of this real-time PCR assay was
determined to be 10 copies/PCR reaction through serial dilution of cloned PCR arnplicon
(Xagoraraki et al. 2007).

Nested PCR and adenovirus DNA sequencing. Adenovirus species present in
water samples were identified by amplifying and sequencing of the hypervariable region
1-6 of adenovirus hexon gene. Samples that were positive by the real-time PCR assay
were amplified using a conventional PCR assay prior to sequencing. Primers used and
PCR conditions were as described by Lu and Erdman 2006 (Table 2.1); primers AdhexF 1
and Adhele yield amplicons ranging in size from 764 to 896 bp. Ifinsufficient DNA
was amplified fi'om the first PCR, 3 second PCR was performed using internal primers

AdhexF 2 and AdhexR2, which yields amplicons between 688 and 821 bp (Lu and

61

Erdman 2006). The sensitivity of this conventional nested-PCR assay was determined by
limited dilution experiments of pure adenovirus stock in cell culture lysates (~108 viral
particles ml '1). Virus stocks were extracted for DNA, serially diluted to 10 0 virus DNA
copies/reaction (as quantified by real-time PCR) and assayed by conventional PCR. The
detection limit of this conventional PCR was between 10 and 100 virus DNA
copies/reaction.

Amplicons were visualized on an agarose gel and purified using a QiaQuick DNA
purification kit (Qiagen, Valencia, CA, USA) prior to sequencing. All HAdV positive
samples were identified by sequencing of at least two independent PCR products in both
directions, i.e., each nucleotide was determined at least four times. Sequencing was
performed by the Research Technology Support Facility at Michigan State University.
Sequences were determined on an ABI PRISM® 3100 Genetic Analyzer (Applied
Biosystems).

Sequence analysis. Sequences were aligned with hexon protein sequences of
selected adenovirus prototype strains available from Genbank using web-based ClustalW
(http://www.ebi.ac.uk/clustalw/) with the following default settings: gap opening (10),
gap extension (0.2) and hydrophilic residue gap penalties enabled. For analysis of
serotypes, approximately 600 bp of sequences of the PCR products were used. Amino
acid alignment of the sequences was performed in ClustalW. A phylogenetic tree was
then constructed using the neighbor-joining method. The branching confidence was
estimated by bootstrapping with 1000 resarnplings in MEGA 4.1 (Beta). The GenBank
accession numbers of adenovirus prototypes used in alignment and phylogenetic analysis

are: species A: human adenovirus type 12 (AB330093), type 18 (DQ149610), and type

62

31 (AB330112); species B]: human adenovirus type 3 (EF494650), and type 7
(AC000018); species B2: adenovirus type 35 (AC000019); species C: adenovinrs type 1
(AC000017), and type 2 (EU867481); species D: adenovirus type 8 (AB361058), type 19

(P13330133), and type 51 (AB330132); species B: adenovirus type 4 (A3330085); and

species F: adenovirus type 40 (AB330121) and type 41 (EF429128).

63

 

~Nw-www

ocom
589m— 28 n..— @343

32..

am
2.
mm

3
mm
me

we

<00<55<0<0<000555500 $9085....
05<500m<<E>0§00>00 8.35 $8850... 58.8 .0382
5505550053055 3 55:3.
5.. 88:
<05F05050P<0<05E seem $588.... 585 >3:
55500085050505 5 E 2.8..
50508055000055... 3 E 5..
523%

0<050<5<0005000<00 8.3m 3 59.: 8&5 >3...

 

.3 so 388.32. mm
. a... 3..

oocouowom 533:2 69 ab

8598 3833
Am- 3 coco—50m 0:00 0:82 operate—bin >82

 

.8388 03m? 5 <75 “0.33053 mo 5:085 05 5 com: 8305 98 505.5 mm 2an

64

Virus recovery efficiency. The virus recovery efficiency of the virus adsorption-
elution method by Haramoto et a1. (2005) was evaluated in two seeded studies. In the first
experiment, the efficiency of the method in recovering naturally occurring adenovirus in
sewage was tested. Two different volumes (3.8 liters and 700 ml) of MilliQ water were
inoculated with raw sewage (collected from East Lansing wastewater treatment plant) in
the ratio of 100,000:1, l0,000:1, 1,000:l, 100:1 and 10:1. Inoculated water was mixed at
room temperature for at least an hour before processing. Seeded samples were
concentrated, extracted and detected by real-time PCR following protocols described
previously. For each concentration, three to five replicates were tested. For the sewage-
seeded experiment, viruses were eluted directly from the membrane with lmM NaOH.

In the second experiment, the efficiency of the method in recovering
bacteriophage P22 from environmental water was evaluated. Bacteriophage P22 was
inoculated into one-liter of surface water samples and CSO discharge samples in the ratio

range of 10:1 and 5000:1. Both eluate and filtrate were assayed for infectious phage using

phage assay (Adams 1959). Retention ratio, R filmtion , by the filter was calculated as

follow:

(T om, seeded - T at”! filtered )
T 0t“, seeded

 

Rfiltration =

where Total seeded and Total filtered are the total number of virus seeded and total

number of virus after filtration, respectively.
Bacteriophage P22 was eluted fiom the membrane using two methods: 1) 20-m1
of lmM NaOH was dispensed onto the membrane and the eluent was collected in a 50ml-

tube with neutralizing buffer; 2) membrane was transferred into a 50-ml tube containing

65

20ml lmM NaOH, the 50—ml tube was then pulse-vortexed for 10 s to elute the viruses,
membrane was removed after vortexing and buffers (SOmM H2804 and 10X TE buffer)
were added to neutralize the NaOH. Percent recovery was calculated using the formula

below:

number of bacteriophage P22 eluted
number of bacteriophage P22 seeded

 

% recovery = * 100 %

Results and Discussion

Waste water. Between August 2005 and August 2006, 46 waste water samples
(13 raw sewage, 13 primary-treated, 10 secondary-treated and 10 tertiary-treated
samples) were collected. Wastewater samples were assayed by real-time PCR for
determining the concentration of adenoviral DNA present (Figure 2.2). Adenoviruses
were consistently isolated in significantly higher concentrations from raw sewage and
primary effluent compare to secondary and tertiary effluents (p-value: <0.001). The
concentrations of HAst ranged between 2.63 x 105 and 2.82 x106 viruses/L (mean: 1.15
x106 viruses/L) in raw sewage, 53.7 and 4094 viruses/ml (mean: 1.12 x 106 viruses/ L) in
primary effluent, 1.05 x 103 and 4.42 x 104 viruses/ L (mean: 2.0 x 104 viruses/L) in
secondary effluent and 1.35 x 104 and 4.28 x 105 viruses/ L (mean: 8.3 x 104 viruses/ L)
in tertiary effluent. A seasonal trend was not observed for the concentration of
adenoviruses DNA in wastewater samples. Figure 2.3 and Table 2.2 shows average
adenovirus concentration and logm removal ratio of human adenovirus at the East
Lansing wastewater treatment plant. The ratio removal was calculated by using the
average virus concentration in a particular treatment process over the virus concentration

in the raw sewage. The overall adenovirus logio removal from raw sewage to tertiary

66

effluent was approximately 1.14 logic. Tukey's honestly significant difference test
showed that virus concentrations were not significantly different between raw sewage
and primary effluent (p-value: 0.74), and between secondary effluent and tertiary effluent

(p-value: 0.1 1).

67

.88 2&3. e5 88 new? e858 .3222 .maaaq

“mam E :53 Eng—mob 633883 30538 5 89a 038:8 .833 233 E 38833 Ewe: >0; mo 3388850 NN gamma

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5.3:
8.82 8.22. 8.9.2. 8%: 8.22 8.52 858 8.5.. 8.80 8.32 8.50 8.8m 8.92
n .. F . .- L . . L , _ , p 0 . I N
W.
L. 1 i I u . 1- m
H
v
m
.1 I .l V )
a L F. m
r r m
e 0
L L ..A...
.L I.
r - m m
9
N
r h
- 8
83:30 88ng ensue. sem-

 

 

68

Table 2.2 Arithmetic mean and range of adenoviruses detected from each treatment step

and average logm removal compared to virus concentration in raw sewage.

 

 

Sample Number of Mean Range Log Removal
Type samples (n) (103 viruses /L) (103 viruses/L)
Raw 13 1152* 263-2817
Primary 13 1123‘ 53.7-4094 0.01
Secondary 10 20’” 1.05-44.2 1.77
Tertiary 10 83’” 13.5-428 1.14

 

* virus concentrations were not significantly different between raw sewage and primary
effluent (p-value: 0.74).
** virus concentrations were not significantly different between secondary effluent and

tertiary effluent (p-value: 0.1 1).

 

 

 

 

 

 

71
s 6]
3e; 5.
E.”
> 3

2 .

Raw Prirmuy Secondary Tertiary
Sanpletype

Figure 2.3 Mean adenovirus concentrations (loglo virueses/L) in raw sewage, primary

effluent, secondary-treated effluent and tertiary-treated effluent.

69

A total of 28 wastewater samples (resulted in 35 isolates) were amplified by
nested PCR for genotyping (Table 2.3). Samples that produced a high enough
concentration of amplified adenoviral DNA were sequenced. Sequences were cropped to
approximately 600 bp in length and aligned in ClustalX 2.0.10, with reference sequences
from GenBank. Blast analysis of all amplicon sequences identified 92 to 100 % sequence
identity with reference prototypes listed below, number of insertion/deletion between
amplicons and closest prototypes ranged between 0 and 6 (Table 2.3). Five adenovirus
prototypes (adenovirus type 2, 3, 12, 40 and 41) were dominant in raw sewage and
primary effluent fi'om wastewater in East Lansing, MI. Adenovirus type 41 (21/35; 60%)
was the most frequently isolated adenovirus, followed by adenovirus 12 (1 1/35; 28.5%),
adenovirus 40 (1/35; 2.9%); adenovirus 2 (1/35; 2.9%) and adenovirus 3 (1/35; 2.9%)
(Fig. 2.4). A neighbor-joining tree was generated from alignment of the nucleotide
sequence of amplicons obtained from this study and hexon gene sequences on GenBank
corresponding to adenovirus prototype species A to F (Fig. 2.5). This finding is in
agreement with other studies. that found that enteric serotypes (adenovirus type 40 and
41) are predominant among all adenovirus serotypes in aquatic environments, especially

in sewage (Santos et a1. 2004; Jiang et al. 2005; Haramoto et al. 2007).

70

Table 2.3 Comparison between human adenovirus isolates detected in raw sewage and
primary effluent of East Lansing wastewater treatment plant and their closest prototypes
on GenBank. Adenovirus prototypes used in the comparison are: species A: human
adenovirus type 12 (AB330093), type 18 (DQ149610), type 31 (AB330112); species Bl:
human adenovirus type 3 (EF494650), type 7 (AC000018); species B2: adenovirus type
35 (AC000019); species C: adenovirus type 1 (AC000017), type 2 (EU867481); species
D: adenovirus type 8 (AB361058), type 19 (AB330133), type 51 (AB330132); species B:

adenovirus type 4 (AB330085); and species F: adenovirus type 40 (AB330121) and type

 

 

41 (EF429128).
Date Sample Highest score prototype No. of
Type type % nt identity insertion/deletion

Aug-05 raw 41" 100 0
Sep-OS raw 12* 93 3
i 41“ 100 0
Oct-05 raw 12* 93 3
41" 100 o
Nov-05 raw 41 98 O
Dec-05 raw 1 2 99 1
2 93 6
Jan-06 raw 41" 100 0
Feb-06 raw 3 100 0
Mar-06 raw 41" 100 0
Apr-06 raw 12* 93 3
41“ 100 0
May-06 raw 12* 93 3
June-06 raw 12* 93 3
July-06 raw 41" 100 0
Aug-06 raw 1 2 92 4
4 1 It * t 99 O
June-07 raw 4] * * 100 O
July-07 raw 12 93 0
Aug-05 primary 41 :1: II: 100 O

71

Table 2.3 cont’d

Sep-OS primary 41'” 100 0
Oct-05 primary 41*" 99 0
Nov-05 primary 12* 93 3

41 99 1
Dec-05 primary 41 ** 100 0
J an-06 primary 41 ** 100 O
Feb-06 primary 41" 100 0
Mar-06 primary 40 99 0
Apr-06 primary 12 93 3
May-O6 primary 41" 100 0
June-06 primary 41" 100 0
Aug-06 primary 41" 100 0
July-07 jrimary 12 96 0

 

* These isolates were 100% homologous to each other and were most closely related to
prototype HAdV 12.

** These isolates were 100% homologous to each other and prototype HAdV 41.

*** These two isolates were 100 % homologous to each other and were most closely

related to prototype HAdV 41.

72

HAdV2 ‘7 HAdV3
HAdV“ (2.9%) / (2.9%)
(2.9%) 5 [\-

   
 

(‘ HAdV 41
(60%)

Figure 2.4 Adenovirus serotypes isolated fiom raw sewage and primary effluent samples

collected fiom East Lansing waste water treatment plant (n=3 5).

73

100

 

 

74 Raw0707

Raw0905b
Pri0406
Raw0806

Pri0707
Raw1205

97

1Ad40 F

 

 

 

 

100'- Pri0306
— Ad31 A
—— Ad18 A

 

 

 

 

100

Adl C
Raw12056

 

 

65
100

 

 

Ll Raw1105

E Raw0806d

 

100 — Adz C
Raw0805

L Pril 105b
Ad41 F

 

 

Raw1205d

 

Ad4 E
Adl9 D

 

 

 

 

 

AdSl D

 

 

 

 

Ad8 D

 

Ad35 B2

 

 

 

63

 

'— Ad7 B1

 

100.

l———-l
0.05

Ad3 B1
100 Raw0206

 

Figure 2.5 A neighbor-joining tree of adenovirus hexon amplicons detected in waste

water. The scale indicates nucleotide substitutions per position. Numbers above the

branches indicate boostrap percentage (>50) based on 1000 replicates.

Reference strains of human adenovirus were selected from GenBank under the accession

number indicated in the text.

74

In this study, adenovirus 41 was identified as the predominant serotype in sewage,
followed by adenovirus type 12, 40, 2 and 3. This trend is agreeable with results from
other environmental studies (Santos et a1. 2004; van Heerden et al. 2005). Santos et al.
(2004) isolated adenoviruses 40 and 41 from 62 out of 69 sewage and surface water
samples collected in San Paulo city, Brazil, over a three-year period. In South Africa,
adenovirus types 2, 40, 41 and species D HAst were isolated from treated drinking
water and river water (van Heerden et al. 2005). However, though enteric adenovirus
(serotype 40 and 41) were most frequently identified in river water, treated drinking
water was predominated by species D human adenovirus. In addition, Gray et al. (2007)
reported the most prevalent serotypes among civilian (n=1608) were types 3 (34.6%),

2 (24.3%), 1 (17.7%) and 5 (5.3%). Among specimen collected from gastrointestinal
tract, only 26 of 234 (11.1%) were adenovirus species F (i.e. adenovirus 41). This
suggests that adenovirus 41 may have greater persistence in natural environments than
other adenovirus serotypes.

Adenoviruses species F (types 40 and 41) have been identified as one of the most
prevalent viruses in the etiology of childhood gastroenteritis in developed countries
(Topkaya et al. 2006). Several researchers claimed that adenovirus type 41, the most
prevalent serotype isolated in the current study, had gradually replaced serotype 40 as the
predominant serotype isolated from gastroenteritis patients globally starting in the 19808
and was currently identified as the most prevalent serotype in wastewater (Grimwood et
al. 1995; Logan et al. 2006; Shinozaki et a1. 1991; Yamashita et al. 1995). These findings
were consistent with the fact that in the current study adenovirus 40 was detected in only

one out of 35 isolates. Adenovirus type 12, the second most prevalent adenovirus isolated

75

in this study (1 1/35), was often associated with meningoencephalitis and no waterborne
outbreak related to it has been identified to date. However, it was recently identified as
the etiologic agent of a diarrhea outbreak in a hematology ward in London (J alal et al.
2005)

Adenoviruses type 2 and 3 were both isolated from one waste water sample. They
are generally associated with pneumonia and childhood respiratory diseases (Horwitz
2001). Adenovirus type 3 has also been identified as the etiologic agent for
pharyngoconjunctival fever and conjunctivitis in several outbreaks associated with
contaminated recreational water (Foy et a1. 1968; Martone et al. 1980; McMillan et a1.
1992). Gray et al (2007) identified adenovirus type 3 and type 2 as the two most
prevalent adenovirus types in civilian specimens (34.6% and 24.3%, respectively).
Adenovirus type 3 is also the second most prevalent (2.6%) in specimens collected from
military trainees (Gray et al. 2007).

The slightly higher concentration of adenovirus in tertiary effluent (after tertiary
treatment) may be the artifact of real-time PCR caused by the presence of a higher level
of inhibitory substances in samples after secondary treatment. This is in concert with the
findings by He and Jiang (2005), in which higher concentrations of adenovirus DNA was
detected in secondary effluent.

Surface Water. Surface water samples were collected from the lower Grand
River during a one-week intensive study in June 2005 and periodically over a two-year
period between 2005 and 2007. During the intensive study, a total of 16 surface water
samples were collected fi'om different sites at the lower Grand River and human

adenovirus was detected by conventional PCR in four samples (16%) (Table 2.4).

76

Following the intensive study, several sites were chosen for continued monitoring
because of high bacterial indicators concentration (data not shown) and detection of
HAdV DNA from two of 16 intensive sites (Deer Creek and Sixth Street Park). Over the
two-year period, 26 samples were collected among the seven sampling sites (North Beach
Park, North Shore, Riverside Park, Rosy Mound, Deer Creek park, Grand River Park and
Sixth Street Park) (Table 2.5). Adenovirus DNA was detected in four of six samples
collected in 2005 through conventional PCR. For the 20 remaining samples collected in
2006 and 2007, adenovirus concentration ranged between < 0.01 and 66 viruses /ml
(mean: 7.76 viruses/ml) was quantified in these samples by real-time PCR. Six of these

twenty samples with adenovirus DNA concentration were above the detection limit.

77

Table 2.4 Surface water samples positive for adenovirus DNA by nested PCR collected
during one-week intensive sampling of the lower Grand River, Michigan, June 19-June
24, 2005. Samples were arranged in the order of sampling sites, from Lake Michigan to
the most inland site on Grand River, Riverside Park, Grand Rapids, MI. Refer to Figure

2.1 for location of sampling sites.

 

Site Nested PCR
results

 

Lake Michigan 1 -
Lake Michigan 2 -
Rosy Mound -
N. Beach Park -
Petty's Bayou +
Spring Lake -
Deer Creek Park +
Lamont -
Crockery Creek -
Johnson Park 1 -
Johnson Park 2 -
Kent county 1 +
Kent county 2 -
Kent county 3 -
Sixth Street Park +
Riverside Park -

 

78

 

seem

 

 

 

8.: bus" 003% 0:05
modv wham 22% £52
modv vodv modv - Mam 0283mm
scam
Nodv Nodv 86v 3.: :80m 5.52
34 36v _odv mcdv vodv + + 030m #0000 000G
Nodv mafia vodv Nodv + + fig 0005 55m
- 9:52 30%
389% 8693 co: :3 25:8 09 _Qmo 35:8 234:2 mQSE y 3N9:
$30 mam—953 0:8: 86
3:52 mom 0068053230280
Soak Ea

mo\m\: :00309 42 £0.43 0:80 0032 05 Sch $388 0083 000003 Som— £E\ 8255 >305 .5 5300020 mom 00602 Wm 033.

79

CSO samples. Six CSO samples were collected from a wastewater retention
basin in Grand Rapids, between February and June 2008. Adenovirus DNA was detected
from 100 % of CSO samples with concentrations ranged between 6.02 x104 viruses/L and
1.32 x106 viruses/L (average: 5.35 x105 viruses/L; standard deviation: 500.81) (Fig. 2.6).
The concentration of adenovirus DNA in the CSO discharge was not significantly
different from adenovirus DNA concentration in raw sewage and primary-treated samples
(p-value: 0.39). This suggests that in the case where virus concentration in CSO samples
is unknown, virus concentrations in raw sewage or primary-treated samples may be used

as substitutes for estimating total virus discharge during a CSO event.

HAdV conc (Iog1o viruses/L)

 

 

 

2/17 4/9 4/1 1 6/5 6/9
Sampling date

Figure 2.6 Human adenovirus concentrations (logic viruses /L) in CSO discharge samples

collected from Market St retention basin in Grand Rapids, Michigan.

80

Virus recovery efficiency. The virus recovery efficiency of the virus adsorption-
elution method by Haramoto et al. (2005) was evaluated in two seeded studies, a sewage
seeded study and a bacteriophage P22 seeded study. For the sewage seeded study,
detection limit is also affected by sample volumes. Conventional nested-PCR was able to
detect adenovirus DNA in 1110000 dilution in the larger volume tested (0.3 8ml of sewage
seeded in 3.8 L) but only up to 1:1000 dilution in the smaller volume tested (0.7ml of
sewage in 700 ml MilliQ water).

In the second experiment, bacteriophage P22 was seeded into river water.
Bacteriophage P22 was inoculated into one-liter of surface water samples in the virus to
volume ratio range of 10:1 and 5000:]. The P22 recovery was determined by plaque
assay. Preliminary results by plaque assay showed that elution of viruses by transferring
the membrane into elution buffer (N aOH) followed by pulse-vortexing is one log more
efficient than eluting directly from the membrane. The recovery efficiency of eluting
directly from the membrane range between 0.001 and 0.11% while pulse-vortexing the
membrane in elution buffer gives recovery range between 0.11 and 19.94%.

The low recovery from seeded experiments may be due to loss of viruses during
elution step, viruses may not be completely eluted fi‘om membrane as the membrane was
folded into quarters before vortexing or inactivated during elution. Bacteriophage may
not be representative of animal virus recovery because of the sensitivity of low pH during

the elution step (McAlister et al. 2004).

81

Conclusions

In this part of the research there were two general hypotheses I was addressing:

1) human adenoviruses as important waterborne pathogens, are present consistently in
sewage and were occasionally detected in rivers receiving sewage effluents in Michigan;
2) wastewater treatment does not efficiently remove adenoviruses (or at least their DNA).
The real-time PCR assay used in the current study was not able to evaluate viability of
viruses detected after disinfection step, but virus concentration in tertiary treated effluent
showed little physical removal.

In general, there is not a consistent seasonal trend for the presence adenoviruses in
wastewater and surface water in Michigan but their isolation from surface water can be
used to indicate fecal contamination and lack of efficient treatment. In general, waste
water treatment reduces level of adenoviruses by 95 % from untreated to treated sewage.
Real-time PCR is quantitative and more sensitive than conventional PCR. Quantitative
detection of adenoviruses in surface water could give a better understanding of the risk
associated with recreation in contaminated water.

The presence of adenoviruses in surface waters does represent to some extent a
public health risk. It has been shown that there is a connection between environmental
and clinical virus isolates in a given year within specific geographic areas (Sedmak et al.
2003). Sedmak et al. (2003) compared clinical and sewage isolates of enteroviruses fi'om
Milwaukee, Wisconsin collected between 1994 and 2002, and found that the predominant
clinical serotype was most often the predominant sewage serotype for that year. In this
study, typing of isolates from sewage and environmental water may help to alert the

community health professionals to predominant and new serotypes that are in circulation

82

in the populations, providing new insights to vaccine development and other prevention
strategies.

Concentrations of human adenoviruses obtained from this study using real-time
PCR were the basis of the exposure and were then incorporated into the quantitative risk
assessment model for evaluating human health risk from recreation in the lower Grand

River after a CSO event (Chapter 4).

83

References

Akihara, S., T. G. Phan, T. A. Nguyen, G. Hansman, S. Okitsu and H. Ushijirna. (2005).
"Existence of multiple outbreaks of viral gastroenteritis among infants in a day
care center in Japan." Arch. Virol 150(10): 2061-2075.

Basu, G., J. Rossouw, T. K. Sebunya, B. A. Gashe, M. De Beer, J. B. Dewar and A. D.
Steele (2003). "Prevalence of rotavirus, adenovirus and astrovirus infection in
young children with gastroenteritis in Gaborone, Botswana." East African Med J
80(12): 652-655.

Brandt, C., H. Kim, B. J. A. Vargosko, J. Arrobio, B. Rindge, R. Parrott and R. Chanock
(1969). " Infections in 18000 infants and children in a controlled study of

respiratory tract disease. I. Adenovirus pathogenicity in relation to serologic type
and illness syndrome." Am J Epidemiol 90(6): 484-500.

Chabemy, I. F ., P. Schnitzler, H. K. Geiss and C. Wendt (2003). "An outbreak of
epidemic keratoconjunctivitis in a pediatric unit due to adenovirus type 8." Infect.

Cont. Hosp. E12. 24(7): 514-519.

Chmielewicz, B., J. Benzler, G. Pauli, G. Krause, F. Bergrnann and B. Schweiger (2005).
"Respiratory disease caused by a species BZ adenovirus in a military camp in
Turkey." J Med. Virol 77(2): 232-237.

Crabtree, K. D., C. P. Gerba, J. B. Rose and C. N. Haas (1997). "Waterbome adenovirus:
a risk assessment." Water Sci. Technol. 35(11-12): 1-6.

Craun, G. F. (1991). "Causes of waterborne outbreaks in the United States." Water Sci.
Technol. 24: 17-20.

Cruz, J. R., P. Caceres, F. Cano, J. Flores, A. Bartlett and B. Ton’m. (1990).
"Adenovirus types 40 and 41 and rotaviruses associated with diarrhea in children
from Guatemala." J Clin. Microbiol. 28(8): 1780-1784.

D'Ambrosio, E., N. Del Grosso, A. Chicca and M. Midulla (1982). "Neutralizing
antibodies against 33 human adenoviruses in normal children in Rome." J H .
89: 155-161.

84

Fcng, T., L. S. Mansfield, D. L. Wilson, D. J. Schwab, S. L. Molloy and J. B. Rose.
(2007). "Massive microbiological groundwater contamination associated with a
waterborne outbreak in Lake Erie, South Bass Island, OH." Environ Health
Peggect. 115(6): 856-864.

Fcy, H. M., M. K. Cooney and J. B. Hatlen (1968). "Adenovirus type 3 epidemic
associated with intermittent chlorination of a swimming pool." Arch. Environ.
Health 17: 795-802.

Fcy, H. M., M. K. Cooney, R. McMahan and J. T. Graystcn (1973). "Viral and
mycoplasmal pneumonia in a prepaid medical care group during an eight-year
period." Am. J. Epidemic]. 97(2): 93-102.

Gray, G. C., T. McCarthy, M. G. Lebeck, D. P. Schnurr, K. L. Russell, A. E. Kajon, M.
L. Landry, D. S. Leland, G. A. Stcrch, C. C. Ginccchio, C. C. Robinson, G. J.
Demmler, M. A. Saubclle, S. C. Kehl, R. Selvarangan, M. B. Miller, J. D.
Chappell, D. M. Zerr, D. L. Kiska, D. C. Halstead, A. W. Capuanc, S. F.
Setterquist, M. L. Chorazy, J. D. Dawson and D. D. Erdman (2007). "Genotype
prevalence and risk factors for severe clinical adenovirus infection, United States
2004-2006." Clin. Infect. Dis. 45(9): 1120-1131.

Griffin, D. W., E. K. Lipp, M. R. McLaughlin and J. B. Rose (2001). "Marine recreation
and public health microbiology: Quest for the ideal indicator." BicScience 51(10):
817-825.

Grimwccd, K., R. Carzinc, G. L. Names and R. F. Bishop (1995). "Patients with enteric
adenovirus gastroenteritis admitted to an Australian pediatric teaching hospital
fi'cm 1981 to 1992." J Clin. Microbiol. 33: 131-136.

 

Haramoto, E., H. Katayama and S. Ohgaki (2004). "Detection of noroviruses in tap water
in Japan by means of a new method for concentrating enteric viruses in large
volumes of freshwater." Appl. Environ. Microbiol. 70(4): 2154-2160.

Haramoto, E., H. Katayama, K. Oguma and S. Ohgaki (2005). "Application of cation-
ccated filter method to detection of noroviruses, enteroviruses, adenoviruses, and

torque tenc viruses in the Tamagawa river in Japan." Appl. Environ. Microbiol.
71(5): 2403-2411.

85

Haramoto, E., H. Katayama, K. Oguma and S. Ohgaki (2007). "Quantitative analysis of
human enteric adenoviruses in aquatic environments." J Appl. Microbiol. 103(6):
2153-21 59.

Harley, D., B. Harrower, M. Lyon and A. Dick. (2001). "A primary school outbreak of
pharyngoconjunctival fever caused by adenovirus type 3." Commun Dis Intell.
25(1): 9-12.

Hatherill, M., M. Levin, J. Lawrenson, N. Y. Hsiac, L. Reynolds and A. Argent (2004).
"Evolution of an adenovirus outbreak in a multidisciplinary children's hospital." J
Paediatr. Child. Health 40(8): 449-454.

He, J.-W. and S. J iang (2005). "Quantification of enterococci and human adenoviruses in
environmental samples by real-time PCR." Appl. Environ. Microbiol. 71(5):
2250-2255.

Hilleman, M. R., and J. H. Werner (1954). "Recovery of new agents from patients with
acute respiratory illness." Proc. Soc. Exm Biol. Med. 85: 183-188.

Horwitz, M. S. (2001). Adenoviruses. In B. N. Fields, D. M. K. B. Knipe, P. M. Howley
and D. E. Griffin (Eds), Fields Virology, p. 2301-2326. Philadelphia, PA,
Lippincott Williams & Wilkins.

J alal, H., D. F. Bibby, J. W. Tang, J. Bennett, C. Kyriakou, K. Peggs, D. Cubitt, N. S.
Brink, K. N. Ward and R. S. Tedder (2005). "First reported outbreak of diarrhea

due to adenovirus infection in a hematology unit for adults." J Clin. Microbiol.
43(6): 2575-2580.

J iang, S., H. Dezfulian and W. Chu (2005). "Real-time quantitative PCR for enteric
adenovirus serotype 40 in environmental waters." _(Lan. J Microbiol. 51(5): 393-
398.

J iang, S., R. Noble and W. P. Chui (2001). "Human adenoviruses and coliphages in urban
runoff-impacted coastal waters of Southern Califomia." App]. E_nviron.
Microbiol. 67(1): 179-184.

Jcthikumar, N., T. L. Cromeans, V. R. Hill, X. Lu, M. D. Sobsey and D. D. Erdman
(2005). "Quantitative real-time PCR assays for detection of human adenoviruses
and identification of serotypes 40 and 4]." Appl. Environ. Microbiol. 71(6): 3131-
3 136.

86

Kajon, A., J. Moseley, D. Metzgar, H. S. Huong, A. Wadleigh, M. K. Ryan and K.
Russell (2007). "Molecular epidemiology of adenovirus type 4 infections in US
military recruits in the postvaccination era (1997a€“2003)." J Infect. Dis. 196(1):
67-75.

Kapikian, A. Z., R. G. Wyatt (1992). Viral gastrointestinal infections. In Textbook of
pediatric infectiousg'seases, p. 667-676, Philadelphia, PA: W. B. Saunders Co.

Kolavic-Gray, S., L. Binn, J. Sanchez, S. Cersovsky, C. Polyak, F. Mitchell-Raymundo,
L. Asher, D. Vaughn, B. Feighner and B. Innis (2002). "Large epidemic of
adenovirus type 4 infection among military trainees: epidemiological, clinical,
and laboratory studies." Clin. Infect. Dis, 35(7): 808-818.

Lee, S. H. and S. J. Kim (2002). "Detection of infectious enteroviruses and adenoviruses
in tap water in urban areas in Korea." Water Res. 36(1): 248-256.

Logan, C., J. J. O'Leary and N. O'Sullivan (2006). "Real-tirne reverse transcription-PCR
for detection of rotavirus and adenovirus as causative agents of acute viral
gastroenteritis in children." J Clin. Microbiol. 44(9): 3189-3195.

Lu, X. and D. D. Erdman (2006). "Molecular typing of human adenoviruses by PCR and
sequencing of a partial region of the hexon gene." Arch. Virol. 151(8): 1587.

Martone, W. J ., J. C. Hierholzer, R. A. Keenlyside, D. W. Fraser, L. J. D' Angelo and W.
G. Winkler (1980). "An outbreak of adenovirus type 3 disease at a private
recreation center swimming pool." Am J Epidemiol 111: 229-237.

McAlister, M., H. Aranha and R. Larson. (2004). Use of bacteriophages as surrogates for
mammalian viruses. PDA/EMEA European Virus SJafetv Forum. J. Robertson, S.
Karger. 118: 89-98.

McMillan, N. S., S. A. Martin, M. D. Sobsey, D. A. Wait, R A. Meriwether and J. N.
MacCorrnack. (1992). "Outbreak of pharyngoconjunctival fever at a summer
camp: North Carolina." MMWR. 41: 342-347.

McNeil, K. M., R. M. Hendrix, J. L. Lindner, R. R. Benton, S. C. Monteith, M. A.
Tuchscherer, G. C. Gray, and J. C. Gaydos. (1999). "Large, persistent epidemic of
adenovirus type 4-associated acute respiratory disease in US. army trainees."
Emerg. Infect. Dis. 5: 798-801.

87

Papapetropoulou, M., A. C. Vantarakis. and J. I. Jan;36(1):101-3. (1998). "Detection of
adenovirus outbreak at a municipal swimming pool by nested PCR amplification."
J Infect 36(1): 101-103.

Pina, S., M. Puig, F. Lucena, J. Jofre and R. Girones (1998). "Viral pollution in the
environment and in shellfish: human adenovirus detection by PCR as an index of
human viruses." Appl. Environ. Microbiol. 64(9): 3376-3382.

Rowe, W. P., R. J. Huebner, L. K. Gilmore, R. H. Parrot, and T. G. Ward (1953).
"Isolation of a cytopathic agent fi'om human adenoids undergoing spontaneous
degradation in tissue culture." Proc. Soc. Exp. Biol. Med. 84: 570-573.

Santos, F. M., M. J. Vieira, P. Garrafa, T. A. Monezi, V. H. Pellizari, C. M. Harsi and D.
U. Mehnert (2004). "Discrimination of adenovirus types circulating in urban
sewage and surface polluted waters in Sao Paulo city, Brazil." 1V_ater Sci. Tech:

Water Supply 4(2): 79-85.

Sedmak, G., D. Bina and J. MacDonald (2003). "Assessment of an enterovirus sewage
surveillance system by comparison of clinical isolates with sewage isolates fi'om
Milwaukee, Wisconsin, collected August 1994 to December 2002." AppL
Epviron. Microbiol. 69(12): 7181-7187.

Shimizua, H., Tung Gia Phana, Shuichi Nishimuraa, Shoko Okitsua, Niwat Maneekamb
and H. Ushijima (2007). "An outbreak of adenovirus serotype 4] infection in
infants and children with acute gastroenteritis in Maizuru City, Japan." Infect.
Genet. Evol. 7(2): 279-284.

Shinczaki, T., K. Araki, Y. Fujita, M. Kobayashi, T. Tajima and T. Abe. (1991).
"Epidemiology of enteric adenoviruses 40 and 41 in acute gastroenteritis in
infants and young children in the Tokyo area." S_cand. J Infect. @236): 543-
547.

Tani, N., Y. Dchi, N. Jurumatani, K. chemasu. (1995). "Seasonal distribution of
adenoviruses, enteroviruses and reoviruses in urban river water. " Microbiol.
Immunol. 39: 577-580.

Topkaya, A. E., Benli Aksungar, F atrna Qzakkafl and N. Capan (2006). "Examination of
rotavirus and enteric adenovirus in children with acute gastroenteritis." Tfirk
Mikrobiyol Cem Derg 36(4): 210-213.

88

Turner, M., G. R. Istre, H. Beauchamp, M. Baum and S. Arnold. (1987). "Community
outbreak of adenovirus type 7a infections associated with a swimming pool."
South Med. J 80(6): 712-715.

US. EPA (2001). Method 1602: male-specific (F +) and somatic coliphage in water by
single agar layer (SAL) procedure. Washington DC: US. Environmental
Protection Agency.

Uhnoo, I., G. Wadell, L. Svensson and M. E. Johansson (1984). "Irnpcrtance of enteric
adenoviruses 40 and 41 in acute gastroenteritis in infants and young children." J
Clin. Microbiol. 20(3): 365-372.

van Heerden, J ., M. M. Ehlers, A. Heim and W. O. K. Grabow (2005). "Prevalence,
quantification and typing of adenoviruses detected in river and treated drinking
water in South Africa." J Appl Microbiol 99(2): 234-242.

Wadell, G. A., A. Allard, M. Johansson, L. Svensson and I. Uhnoo. (1987). Enteric
adenoviruses. In Enteric apenoviruses, p. 73-91. Chichester, United Kingdom:
John Wiley & Son, Inc.

Xagoraraki, I., D. H. W. Kuo, K. Wong, M. Wong and J. B. Rose (2007). "Occurrence of
human adenoviruses in two Great Lakes recreational beaches." Appl. Environ.
Microbiol. 73: 7874-7881.

Yamashita, Y., M. Hattori, M. Oseto, M. Mori, H. Inouye, K. Takagi, Y. Ishirnaru and S.
Nakano. (1995). "Epidemiological studies on enteric adenovirus gastroenteritis in
children." Kansenshogaku Zasshi. 69(4): 377-382.

89

CHAPTER 3
THE USE OF BACTERIOPHAGE P22 AS A TRACER FOR STUDYING VIRAL
TRANSPORT AND SURVIVAL IN THE LOWER GRAND RIVER, MICHIGAN,

USA

Abstract

More than 50% of water-related acute gastroenteritis cases may be attributable to
viruses. Viruses are highly persistent in aquatic environments and their transport in
surface water is not firlly elucidated. There has been a strong interest in using
bacteriophages in tracer studies because they highly resemble pathogenic viruses and
more adequately describe colloidal virus transport than dissolved chemical tracers. The
objective of this research was to characterize factors affecting virus transport in surface
water. In the first part of the study, the transport behavior of bacteriophage P22, an
extensively-studied index virus, was evaluated in a dual-tracer study performed on a
40km reach of the lower Grand River in Grand Rapids, Michigan. We examined the
performance of bacteriophage P22 in relative to Rhodamine WT (RWT), a conventional
chemical tracer. Our analysis indicated that bacteriophage P22 can be successfully used
as a tracer in complex surface water environments such as the lower Grand River.
Bacteriophage P22 reacts somewhat differently than RWT to surficial geology and
transient storage zones indicating different processes contributed to virus attenuation
within the same “reach”. Estimated bacteriophage P22 attenuation rates were found to be
in the range of 0.27 to 0.57 per day. The highest attenuation rate was found in a “reach”
with high suspended solids concentration, relatively low dissolved organic carbon content

and sediment with high clay content. Virus attenuation estimates, along with other river

9O

parameters (i.e. river discharge, river width, CSO discharge etc.) related to virus transport
were then incorporated into an analytical hydrology model for predicting virus arrival

time and concentration at a location downstream after a CSO event.

Introduction

Enteric viruses are important etiologic agents for waterborne acute gastroenteritis.
More than 100 types of viruses have been isolated from sewage contaminated water and
virus related waterborne outbreaks have been reported in marine water, chlorinated-
recreational waters, pond, groundwater, and rivers (Bosch 1998; Griffin et al. 2003).
Factors that influence their transport in complex surface water systems are not fully
elucidated; the majority of studies examining virus transport either were smaller scale
studies or were performed in laboratories. Because viruses are present in water as
colloidal particles, they may be transported more conservatively than dissolved chemical
tracers, such as lithium, bromine, flucrescein dye, and Rhodamine WT (RWT) (Rossi et

al. 1998).

Bacteriophages are viruses that infect and replicate in host bacterial cells and have
no adverse effect on humans or animals. Bacteriophages have been suggested and used as
an index virus for indicating human virus fecal contamination because of their
morphological similarity to pathogenic viruses, such as enteroviruses, adenoviruses and
noroviruses. In addition, bacteriophages are very host-specific, they only infect bacteria
with compatible receptors. Laboratory-strain bacteriophages are increasingly being used
as tracers in hydrology studies because they are not present in natural waters, they have a

low detection limit, as well as have inexpensive culture and enumeration methods.

9]

Bacteriophages were first used in studying hydrology of aquifers in the early
1970s (Martin and Thomas 1974). Bacteriophages offer several advantages over colored
dyes as tracers for water movement: they can easily be obtained in high titer in a
relatively small volume (e. g. 1014 pfu in 1-10 liters versus 30-40 liters for Rhodamine
WT dye) and are less costly than chemical tracers. The phage assay employed examines
viable viruses and this allows for examination of both virus transport and virus
inactivation in water. Bacteriophages are not visible to the eye, even in high
concentrations and this may allow for their application in testing water movements in
inhabited urban areas. In addition, because of the host specificity of the bacteriophages,
different bacteriophages can be injected at the same time or at different time points in the

same water system (Rossi et al. 1998).

Large varieties of bacteriophages have been recovered from diverse geographic
areas, such as coliphages (viruses that infect E. coli), and bacteriophages P22, PRD],
H/l40, and T2 (Ackermann and DuBow 1987). Bacteriophages range in sizes from ten to
several hundred nanometers and differ morphologically (Rossi et al. 1998). In general,
bacteriophages have a capsid that encapsulates the genetic material, a neck that connects
its capsid to its tail, a tail and tail fibers (for attachment and adsorption into host cells).
Bacteriophages are categorized into six families based on their morphology and nucleic
acid type (i.e. DNA versus RNA, enveloped capsid versus nonenveloped capsid, tailed
versus tailless and contractibility of the tail). Bacteriophages, such as bacteriophage P22
(previously confused with bacteriophage PRDl in some studies), bacteriophage <DX174
and bacteriophage MS2 have been used as biological tracers to study water movements

and transports of microbiological water pollutants in surface water, groundwater, aquifers

92

and soil (Dowd et al. 1998; Rossi et al. 1998; John and Rose 2005; Collins et al. 2006).
Factors affecting attenuation of bacteriophages/viruses in surface water are complex, and
may depend upon interactions of several factors. In general, temperature and sunlight
irradiation are considered two most influential factors for virus survival in surface water

(Ferguson et al. 2003).

Recovery rate of bacteriophage from water may also be dependent upon water
composition. Bacteriophages are readily adsorbed to suspended particulate matter and
settle to the bottom of the water column through aggregation. Some previous tracer
studies reported inconsistent virus recovery and this may have deterred hydrologists
from using phage to monitor wastewater and surface movement. In one of the earliest
tracer studies, reported recoveries of phage tracers fi'om waste water were 25% and
0.023% for bacteriophage F52 and bacteriophage T7, respectively (Kinnunen 1978).
However, in a more recent surface water tracer study, recovery of bacteriophage type
H40/1 was higher, 61%, compared to fluorescein dye recovery of 49% (Rossi et al.

1998)

In Michigan, 33% of nine major rivers tested positive for the presence of viable
enteric viruses and they are suspected to be the chief cause of swimming-associated
diseases in recreational waters (Jenkins et al. 2005). In older cities in Michigan, such as
Detroit, Lansing and Grand Rapids, discharge of untreated/partially treated waste water
from the combined sewer systems during heavy precipitation events is one of the main
sources of surface water contamination. Though chemical tracers have been used to study

contributing factors to flow path of chemical contaminants in surface water, the reliability

93

of these tracers in predicting microorganism flow paths has not been evaluated. In this
study, the performance of a biological tracer, bacteriophage P22 in a complex surface
water system (the lower Grand River) was evaluated relative to RWT and was used to
determine the relative importance of environmental factors that influence the attenuation
of viruses. Site-specific parameters and inactivation rates were obtained from this tracer
study to estimate virus arrival times and concentrations at recreational sites downstream
from CSO discharge outlets at Grand Rapids area to beaches at Lake Michigan. In
addition, analysis based on a transient storage (TS) model by Shen et al. (2008) was

compared with parameters obtained fi'cm analytical solutions used in this study.

Material and Methods

Phage Identification and Isolation

Bacteriophage P22, which was once confused with bacteriophage PRDl, has been
extensively used in groundwater studies as a biological tracer for tracing subsurface flow
and shown as a reliable tool for studying the transport of contaminants (Enriquez et al.
2003; Harvey and Ryan 2004; John and Rose 2005). Recently, in an attempt to develop a
molecular detection assay for the bacteriophage tracer “PRDl” in our laboratory
(Michigan State University (MSU) Water Quality and Health Laboratory), it was shown
that the bacteriophage genetically characterized as bacteriophage P22 (MSU strain P22 is
available on Genbank with accession number AB362338). The same isolate has been
used in many hydrology studies and was misclassified as bacteriophage PRD]. Some of
the affected studies are bacteriophage PRD] publications ficm laboratories of Dr C. P.

Gerba, Dr J. B. Rose (John and Rose 2005; Nicosia et al. 2001; Paul et al. 1995; Paul et

94

al. 1997) and Dr S. Farrah (Lukasik et al. 2000; Lukasik et al. 2003; Scott et a1. 2002), Dr
R. Harvey (Abudalo et al. 2005; Bales et a1. 1995; Blanford et al. 2005; Harvey 1997;
Harvey and Ryan 2004; Loveland et a1. 1996; Pieper et al. 1997; Ryan et al. 1999; Ryan
et al. 2002; Van Cuyk et al. 2004) and Dr M. Abbazadegan (Drees et al. 2003; Gerba et
al. 2003). The bacteriophage research involving bacteriophage P22 are shown in Table

3.1.

95

Table 3.] Identity of bacteriophage “PRDl” stocks received from Dr Charles Gerba,

University of Arizona, between 1993 and 2002.

 

Dates of
Publication

Direct /Indirect

Evidence

Investigator /
Laboratory

Identity of
Stock

References

 

1995-2005

Direct sequencing

Rose/ USF

P22

Paul et al. 1995; Paul et
al. 1997; Nicosia et al.
2001; John & Rose 2005

 

1995-2005

Direct sequencing

Harvey/ USGS

P22

Bales et al. 1995;
Loveland et al. 1996;
Harvey 1997; Pieper et al.
1997; Ryan et al. 1999;
Ryan et al. 2002; Harvey
and Ryan 2004; Van
Cuyk et al. 2004;
Abudalo et al. 2005;
Blanford et al. 2005

 

1999-2002

Direct Sequencing

Schijven/NL

Schijven et al. 1999;
Schijven and
Hassanizadeh 2000;
Schijven et al. 2000;
Schijven et al. 2002a;
Schijven et al. 2002b;
Schijven and Simunek
2002

 

2000-2003

Direct sequencing

Farrah/ UFL

P22

Lukasik et al. 2000; Scott
et al. 2002; Lukasik et al.
2003

 

2003

Direct sequencing

Abbaszadegan/
ASU/U A

P22

Drees et al. 2003; Gerba
et al. 2003

 

 

2004

 

Direct Sequencing

 

Brion/UKY

 

PRDl

 

Brion et al. 2004

 

96

 

Bacteriophage P22 was isolated by Zinder and Lederberg in the 19508 and has
since been used as a generalized transducing agent for S. enterica serovar typhimurium
LT2 (Zinder and Lederberg 1952). It was the first generalized transducing phage to be
discovered. Bacteriophage P22 contains double stranded DNA (dsDNA) that is
approximately 43,400 bp and has a very short tail. Bacteriophage P22 infects smooth
strain of Salmonella typhimurium (those that carry O-antigen surface polysaccharide)
(Steinbacher et al. 1997). Bacteriophage P22 is relatively more stable than other
bacteriophages without dsDNA, such as M82 (single stranded RNA (ssRNA)) and
(DX174 (ssDNA) because it is able to repair damages to its DNA using host cell repair
mechanism. Both bacteriophages PRD] and P22 have isoelectric point of less than pH
4.0 and attach poorly to soil particles under field conditions (usually pH 6 to 8) (Ryan et
al. 1999; Schijven and Hassanizadeh 2000). Bacteriophages P22/PRD1 are favorable
surrogates for ground water studies because they can tolerate a large range of temperature
and have low inactivation rates of 0.02-0.05 logm per day between 10 and 23 °C (Harden
et al. 2003; Harvey and Ryan 2004; Paul et al. 1995). In addition, both bacteriophages
are structurally and morphologically similar to human adenovirus, an important
waterborne pathogen; thus bacteriophages P22/PRD1 are good model viruses for
understanding the behavior of relevant pathogenic viruses, e. g. adenoviruses (Belnap and
Steven. 2000). The comparisons between bacteriophage PRDl, bacteriophage P22 and

human adenoviruses are shown in Table 3.2.

97

 

3:500:3wa 0% 095% 3:333:08”:

 

 

 

 

 

 

bougame 3308 £53000» Eat—083mb .8508 28003 3 a. m .2
:00 58:: “gamma 00.2030 SBSEEM. 53> £00003 3:-580 are!
an atom .38: a: s. .535 a: 3 08::
<75 Bun—30-0330 .<ZQ 002.050-0358 r<75 00053003530 082.09
228683 was
«850 _.o.& 083m 55 .00.?— vE: 032E893
230000—09: .3238”: 0:8 team .3605083 00:5 .Rhcfimea 0.5335
02 oz 02 0ae_0>=m
E: 8-2. E: 8-0m 8: me 0am
03380300200033590. 00203025 083380H mine...—
u0m==>e=0c< ...—:55 «N.— 253 25>

 

.5353 5255 acetone: :0 £35803 58:: use mam 0w0fiot0aoan Jam omega—cross: mo shamans—OD Nd 030B

98

Tracer Study

Study Site. This dual tracer study was conducted on May 8, 2006 on a 40-km
stretch of the lower Grand River, from the city of Grand Rapids to Coopersville,
Michigan. The Grand River is the longest river in the lower peninsula of Michigan,
running 420 km, originating from the city of Jackson, through Lansing, Grand Rapids and
discharges at Grand Haven, Lake Michigan. Currently, there are eight CSO outlets that
discharge approximately 6100 m3 partially and untreated sewage into the lower Grand
River in 2007. The Ann Street Bridge, which located near CSO outlets at downtown
Grand Rapids, was selected as the injection point. Sample collection was carried out at
three downstream bridges (site 1: Wealthy St. Bridge; site 2: 28th Street and site 3: Lake
Michigan Drive) based on feasibility for sample collection (Figure 3.1). Distances of site
1, 2 and 3 from the injection point are 4.56, 13.69 and 28.38 km, respectively. Distances

of each site from injection point are shown in Table 3.3.

Because this river is significantly wider in comparison with river width reported

in many previous tracer studies, sampling was done at three locations (left, right and

, center) at site 1 and site 2 and two locations at site 3 (left, right) to account for lateral
variability (variability in flow rate across the river). A United States Geological Survey
(U SGS) river discharge gaging station (#04119000) is located 1.05 km upstream from
the first sampling site. The river discharge on the study dates (7 am on May 8th to
7:30 am May 9th) gradually declined from 3230 to 3010 cubic feet per second (cfs)
(114,066 to 106297 m3/ 5). In addition, an RD Instrument 1200 kHz Rio Grande
Acoustic Doppler Current Profiler (ADCP) was used to measure discharges at all the

sampling locations.

99

 

 

 

 

 

 

 

 

F‘ l ”v ‘ fl 2 -.
< / \P ,. N :okkford
\ .1 \ fr
)J'J .\(;:oopersville‘~~.,k K \
)3 i R Ark \ J/'
ame \ Cqmstock P
68t St. Bridge “8' \V/ (i
2 /,~ . i Walks; m Ann St. Bridge
<5J Allendale Lakeéllichigan Drive x ‘
2:3 .5 Wealthy St. Bridg- Grand Rapids
\\_m_¥__ / . ‘ East Grand Rapidsl
- 8th St. Bnd 'e
ELegend Jenis- KmfidxvmeWyommg “LN—r, mgr/f-
‘ uses Flow Gage nggfigne E \\ Kemwmg
Grand River / ‘1 __
ADCP Transects § r) cut'e'vm" /
Jamestown r"
I Tracer Study suedl 0_ 2_ 4 fig" W12 1%ometers

 

 

 

Figure 3.1 Map of the Grand River showing the spatial distribution of land use in the

watershed and the three sampling locations.

Rodamine WT and bacteriophage P22 preparation. Rodamine WT
20% (w/w) solution was used in the study. Rhodamine WT tracer was prepared based on
a peak concentration limit of 10 pg L"1 at the last sampling station, which is 37.43 km
downstream from the injection site (only RWT samples were collected from this site).
The bacteriophage P22 stock used in this study was obtained from Samuel Farrah,
University of Florida, Gainsville, FL, and was maintained on the host Salmonella
typhimurium LT-2 (ATCC 19585). Bacteriophage P22 stock were grown by inoculating

lOO-ml log-phase S. typhimurium host with one milliliter of bacteriophage P22 stock

100

(~ 1011 pfu/ml) and incubated at 37 °C for approximately 3-5 hours. After incubation,
0.01 g of lysozyme and three milliliters of 0.2 M sterile BDTA were added to the flask
and mixed well. The culture was then centrifuged at 4000 rpm for 10-15 min and the
supernatant was filter sterilized through a 0.45 pm membrane. Bacteriophage P22 stock
was stored at 4 'C until used.

Tracers release and sampling conditions. Rhodamine WT and bacteriophage
P22 solutions were released into the river (slug release) from the Ann Street Bridge on
May 8, 2006 at 7:00 am. A total of 8770 g of RWT and 16 liters of bacteriophage P22 (4
x 1011 pfu/ml) were released. At each station, 100-ml grab samples were collected
manually from just below water surface. Two samples were taken at the same time. One
was stored in a dark cooler for RWT analysis, and 5X trypticase soy broth (TSB) was
added to the other sample to stabilize the bacteriophage for bacteriophage P22 analysis.
All samples were kept on ice and transported to the laboratory within 12 hours afler
collection and analyzed within 48 hours. Meanwhile, water temperature, pH, suspended
solids and weather data (i.e., ambient temperature, rainfall, wind, etc.) were noted during
sampling. A Turner Designs 10-AU field fluorometer (Turner Designs, Inc, Sunnyvale,
California), was used to initially detect the dye at the first three sites. In order to capture a
complete breakthrough curve, the sampling interval for both tracers was decreased after
receiving a RWT signal. Rhodamine WT samples were analyzed in the laboratory using
the same 10-AU unit. Total suspended solids concentration was determined according to
Standard Method 2540-D-Total Suspended Solids Dried at 103-105 °C (Greenberg et al.

1992)

101

Bacteriophage P22 detection and enumeration. Water samples were assayed
for bacteriophage P22 following the double layer agar plate assay originally described by
Adams (1959). Samples from site 1 were diluted to 10'3 concentration, and between 1 ml
and 2 ml of each sample in at least duplicate were assayed for the phage presence on
tryptic soy agar (TSA). The plates were incubated for 24 h at 37°C. The detection limit of
this method is less than one plaque-forming unit per ml of river water sample.

Tracer data analysis. Estimation of the average velocity (m/s) and average
recovery rate (%) of bacteriophage P22 at each sampling station were computed by fitting
a sine curve to the observed bacteriophage P22 data. Velocity and recovery rates were
compared with calculations computed by Shen et al (2008). Initial virus concentration

was estimated using the discharge data obtained from the USGS gauging station. The

initial concentration of bacteriophage P22 in the river after tracer release, C0 , was
estimated using the formula below:

Ctracer ' Qtracer
= 1
O (Qtracer + QR) ( )

 

where Ctmce, denotes average concentration of bacteriophage P22 in the discharge

(pfu/ml); grace, is the flow rate of tracer (m3/s); Q, is the discharge of the river (m3/s).

Average velocity (m/s), Vave , at each site was calculated by dividing time of

peak concentration on bacteriophage P22 breakthrough curves, t peak , by distance

 

traveled, x.
t peak
Vave = x (2)

102

Predicting Virus Concentration after a CSO Event

For prediction of virus concentration at Lake Michigan beaches after a CSO

event, an analytical virus transport model was developed from the formula below

 

 

 

 

 

 

 

 

(O’Loughlin and Bowmer 1975):
f Ux _ 1 W
——(1-F) _ _ _
e2E erfr{x Utrj-erf x U(t 1')I“ +
C0 2.097 2 E(t—r)
C(x,t)=—2—< U h - >
x P T
—(1+F) _
e25 erfc[x+Utr)—erf x+U(t t)l"
NE 2 E(t-r)
where F: 1+417 and 77:E (3)
U2

in which, C(x.t) denotes tracer concentration as a function of distance traveled, x (m),

and elapsed time since tracer release, I (s); t represents spill duration (5); C0 is the initial

tracer concentration at discharge point (virus/ml'l); U is average velocity (ms'l);

E is longitudinal dispersion coefficient (mzs'l); k is first order decay rate constant (8").

For uniformity in presentation, virus inactivation estimates were obtained from
analysis performed by Shen et a1. (2008). Shen et al. (2008) used a watershed hydrologic
model based on the Soil and Water Assessment Tool (SWAT) to verify the estimated
lateral inflow and includes contributions from the lateral inflow, and solute exchange
between the main channels and storage zones in solute transportation estimation. Solute

transport in the longitudinal direction was then described using the transient storage (TS)

103

formulation (Runkel 1998). Briefly, the model solves two separate equations — one for
solute concentration in the main channel and another in the storage zones. Solute
exchange between the bulk flow and the storage zones was described using a first-order

exchange coefficient a . The governing equations appear as shown below (Runkel 1998):

 

ac QaC 16 6C qL
——=————+—— AD— +— C - + C —C -kC 4
6t Aax A6x( 6x) A("C)a(s ) 0
ac, A

= —C—C —kC 5
at aAs( S) S ()

in which Q denotes the discharge, A and As the cross-sectional areas, C and Cs the
concentrations in the main channel and the storage zones respectively, D the dispersion
coefficient, q L the lateral inflow, a the mass exchange coefficient between the main
channel and the storage zones and k the first order inactivation/decay rate (for
bacteriophage P22). Concentration associated with lateral inflow, CL, was assumed zero
in this case.
Equations (4) and (5) were applied on a reach basis and parameters were estimated
separately for RWT and bacteriophage P22.

For evaluating the relative performance of the two tracers, Shen et al. (2008) used
equation 6, which estimates the fractional recovery of tracer mass by integrating the

tracer breakthrough data assuming complete mixing:
1
=—— C x,t x,t dt 6
M f ( )Q( ) ( )

Here f is the fractional recovery, C(x, t) is the tracer concentration, Q(x, t) is the discharge,
and M0 is the mass released. Because the actual river discharge was not available at all

spatial locations (x), river discharge from the USGS gauging station was used.

104

 

Sensitivity Analysis

Site specific data were collected and analyzed for seasonal and monthly trends.
Historical river discharge data (between 1930-2008) for the lower Grand River at Grand
Rapids (gage no.: USGS 04119000) were obtained from the USGS website
(http://waterdata.usgs.gov/nwis/dv?referred_module=sw&site_no=04119000). CSO
discharge data in Grand Rapids were obtained fiom CSO and $80 annual reports
published by the Michigan Department of Environmental Quality
(www.deq.state.mi.us/csosso). Distributions of these parameters were evaluated and
fitted using EasyF it sofiware (MathWave Technologies). Seasonal and monthly means
for river discharge, CSO discharge volume, CSO discharge duration and frequency were
then computed using SAS software (SAS, Cary, NC). Analysis of variance (ANOVA)
was conducted to evaluate differences in seasonal and monthly means of these
parameters. For parameter that showed significant seasonal/monthly variations in
AN OVA, Tukey's honestly significant difference procedure was conducted to determine

groups that were significantly different from each other.

Output distributions of viral concentrations were obtained using Monte Carlo
simulation techniques that randomly sample each parameter according to its distribution.
A minimum of 10,000 iterations was performed for each simulation. All distribution _
functions were assumed to be statistically independent. Monte Carlo simulation was
carried out with Risk solver (Frontline Systems Inc., Incline Village, NV), an add-on to

Microsoft Excel software.

105

 

Univariate sensitivity analyses were performed to determine the influence of input i
parameters on the shape of virus breakthrough curve and virus concentration downstream

from discharge point.

Results and Discussion

Recovery estimation of bacteriophage P22 using sine curve fitting method was
very close to recovery estimation by Shen et al. (2008) using eq. (6) (Table 3.3). By using
the sine curve fitting method, recovery of bacteriophage P22 at site 1, site 2 and site 3
was 32.93%, 28.56% and 16.52%, respectively, compare to 31.04%, 27.20% and 17.38%
computed by Shen et al (2008). Because breakthrough was incomplete at site 3 (tracer
concentration in the river water did not drop to baseline level when sampling ended),
virus recovery at the site may be underestimated.

Recoveries of bacteriophage P22 were lower than the recoveries of RWT at all
three sites (113.95%, 108.01% and 64.27%, respectively) (Shen et al. 2008). The low
recovery of bacteriophage P22 may be attributable to the turbulence generated by a dam
in downtown Grand Rapids (before site 1) that inactivated the bacteriophage as well as to
potential losses associated with aerosolization of bacteriophage P22 (during release from
the bridge) resulting in overestimation of the initial mass released. These levels of
recovery were also reported in previous literature and may reflect the actual microbial
pathogen survival rate after discharge into a water body (Hodgson et a1. 2004).
Bacteriophage P22 concentrations at each sampling site are shown in Figure 3.2.
Bacteriophage P22 concentrations recovered from all sample sites are shown in Figure

3.3.

106

60508 mafia 258 25 .3 33:58 .0388 .383 095.6 98 .3622» Omani“ :

888 .3 Ho 5mm .3 8:588 .0382 2823 omega“ was .3620.» omen?“ .

 

 

 

 

 

 

 

 

 

e: 5 iv

~92 mm: one mmd 3.3 m 85

cm ewe

owwa 8.3 find 36 $42 N 85
5. 5.8.6

madm 3.3” $5 Ned 03. fl cum

.2. b.2509. . bosom: .. a}: an}: as:
.x. .u.»< e\e .m.»< .3329» .u.»< .5320.» .u.»< 35qu «gm

 

 

 

.wfium 255 25 Ba Awoon .3 Ho eonmv 388 RoEoBHwE a we?» Nam 0953583

Co #6302 0920.6 can .3620.» emanate mo :85988 demon cocoa? 05 Bot 85 mam—mane. 05 3 8ng m.m 053.

107

 

Examination of the reach-averaged velocities for bacteriophage P22 and RWT g
seems to indicate that bacteriophage P22 arrived early compared to RWT though arrival
and peak times for both tracers were fairly close (within 15 minutes of each other) at all
three sites. This may be attributed to the lower detection limit of bacteriophage P22
compare to RWT (this was also noted in Rossi et al. (1998) for comparison between
recoveries of bacteriophage H40/1 and uranine). Other differences in the transport of the
two tracers can be explained by difference in adsorptivity to riparian vegetations, high
organic matter content and suspended solid content as well as the effects of land use
characteristics. Bacteriophage P22 arrived at site 1 approximately 2:38 h after dye
release, at site 2 at 6:47 h and it took 13:45 h to arrive at site 3, which is approximately
28 km from the injection site. Peak bacteriophage concentrations at site 1, site 2 and site

3 were 1477 pfu/ml, 388 pfu/ml and 74 mel, respectively.

108

 

Moog. 838388 05 no 083 39.28 295m o>£moq “Eu 2: .«o 0:5 ”52H.

 

8.x £85 5 anew 5 m: as 8 5 same 5 9c .5 names: 83 .e
was 582 5 seem 5 E 58% 5 58cm 5 3V .5 names: 83 .m

 

$5 58% fi 588 .6 man 58% .K 5:5 .6 B25 am 5% .N
8 see a... as; 5.. E; 38% em 38% am swarm am Ea; ._
canes sesame
deem daze 55 5e 5.29 53.
:2 NE deem waaeam

 

.86 mfiaqam some 3 0:5 x8e 28 33.5 H3 0:535 28 NE owmnaotfioam in 033.

109

 

 

 

 

 

 

 

 

 

 

 

A5 .8383 be. as...
Awe QM
4
o o o
0 Q 0 G 0
4
0
1|“
4“ o
o o «coo 4
o o o
0 ‘ 0 ‘
0
lie. all
«a 00 <4 0
4
.. “aa‘ ‘ w
o
9
o.
0900.“

 

a Na 2:5

ZWMA

oo+m. _

_o+m;

No+m;

mo+m4

vo+m4

llO

([ur/njd) uogauuaauoo ZZd

 

 

 

 

 

 

 

 

 

pan 2&5
A5 8:3? see 2.5.
m_ N. Z c. m m b o
_ . l _ _ _ t S-m._
4
I ll 4 in lr 8+m:
o o u
e Z
0 o m
o o 4 a
‘0 ‘0 O m.
i. 0 0 « ii? 0 ~O+m.— W
o o 4 4 o m
Q C O )
o m 4 4 000 N
o o 4 4 .ml
0 dial-kl «d 8+m._ (
04¢ 00.. e
2330 w“ 00 000
.5280. 444««mooo one
$044 4444

 

- .IL mo+m._

111

 

_|

manna E 003 cougfioeoo omenaotogoan 033 0950.3 .0wEB 02.5 55:32 03 .m 0% 2. A8 3833 mwm 003 count—H0280
0m0aqot083 300m 0920.6 .039. Hm 5mm .N 0:0 24 3V 383m 2.: 00? 5:550:00 0523200803 033 090.82.." .0w25

um 3:00? ._ 0:0. 3. A3 dc 550 _E 0.3 0% 08¢ 0:003 0020 mam—@500 “0 9:38 053055250 Nam 0w0€ot080m Nd 0.35m

 

 

 

 

 

0N.m 0.55m
E 808.5 as: as:
8 2 M: t 2 2 I 5
p Li » _ F p _ ~OIm. —
4
T 8+5 "m
o .0
4 a
0
4 n
3
a
u
1
a it 5+5 u
4 u.
. m.
o o o 4 4 o x)
d
o a M 44 4 m‘« 4 4 4 4 4 4 4 o o 0 MW
0 biggie No+m._ (
£02.33 0 .
604.2.— 4

 

 

- mo+m. _

112

.AEE Ema: 0000—00 00.00 0:3 33> 002m 9:38.00 00b: :0 00 983mm Smoc 8038:5028 Nam 0wmnaot080m m.m 0.5mE

 

r ON
«v g
0 r Wm
0

302 o 02:00 0 03 4 flaw h o m
Wm

3.1: chc 0000—0.— ..3? BE.

NM _.m o.m QN w.m EN 9N Wm 1N Wm mm flu
_ _ _ lr‘ p L Ll _ ‘_ _ 0 WC! 3
a 4 04 8- 0.0 W

o

I . u.
% 0% or 40 mo .m
on“ % 4 O o M o. lo.“ d «M.
‘ 0 . AXv ‘ 1m.— 3
0% Q 4 o W m
0 Z
3
0
u
3
w
m

 

113

 

Shen et a1. (2008) determined the average first-order inactivation rate in reach 2 V
and reach 3 using the formula:

k = k0 + k, 1(1) (7)
where, k; is the rate of inactivation caused by solar radiation (d'l le); 1(t) is the net
shortwave radiation (kW) as a function of time (and geographical location); k0 denotes
virus loss caused by other factors such as temperature and sedimentation. It was reported
that average first-order inactivation rates in reach 2 and reach 3 was 0.27 and 0.57 per
day, respectively, similar solar inactivation values obtained at both reaches suggest
comparable effects of solar irradiation. The inactivation rates of bacteriophage P22 in the
water column in this study are similar to the rates reported in other studies (Schijven et al.
1999). Schijven et al. (1999) reported an inactivation rate of 0.3 per day for
bacteriophage PRDl in a constructed wetland with significant surface flow. The higher
inactivation rate for reach 3 compared to reach 2 may be attributed to a higher total
suspended solid concentration (TSS) in reach 3 and inactivation caused by solar radiation
as bacteriophage P22 traveled downstream and dilutions. Bacteriophages are known to
adsorb to suspended particulate matter, and due to their colloidal nature, can aggregate
into clumps large enough to settle out of the water column although the factors that
influence these processes are complex (Ferguson et al. 2003). This information, coupled
with the fact that reach 3 had the highest TSS, higher percentage clay, lower dissolved
organic carbon (DOC) and the lowest cross-sectional average velocities, suggests that
adsorption and sedimentation was the main cause of bacteriophage P22 attenuation at

reach 3.

114

In order to better characterize virus attenuation, Shen et al. (2008) also examined
the influence of transient storage sizes and land use characteristics (urban versus
agricultural and forested land uses) in their tracer analysis. They concluded that the
estimated transient storage zone sizes correlated negatively with urban land use and
positively with agricultural and forested land uses for both tracers. Because of the
complexity of transient storage size and land use characteristics effects, these factors
were not included in the analytical solution for estimating virus concentration after a

CSO event.

Sensitivity Analysis

Sensitivity analysis was performed to determine the influence of site specific
parameters (i.e. area of channel, river discharge, CSO discharge volume and CSO
discharge duration) and virus inactivation factors (i.e. sunlight intensity, virus
inactivation by factors other than sunlight irradiation) on final virus concentration.
Distributions of some of these parameters, i.e. river discharge, CSO discharge volume
and CSO discharge duration, were first evaluated and fitted using EasyFit sofiware.
Seasonal and monthly means for river discharge, CSO discharge volume, CSO discharge
duration and fi'equency were subsequently computed using SAS software. Analysis of
variance (AN OVA) was conducted to evaluate differences in seasonal and monthly
means of these parameters. For parameter that shows significant seasonal variations in
ANOVA, Tukey's honestly significant difference procedure was conducted to determine
groups that are significantly different fi‘om each other (in order to determine the most

representative set to be used in risk assessment).

115

 

Annual, seasonal and monthly means for these site specific parameters were then
evaluated for selection of sets of values that most represent important climatological
events in the lower Michigan. For river discharge, historical daily discharge data for the
hydrology station on the lower Grand River in Grand Rapids, fiom 1930 to 2008, were
collected from USGS website. Annual, seasonal and monthly means were computed in
SAS for analysis of trend. Figure 3.4 shows that annual means for river discharge
between year 1930 and 2008 were relatively consistent with perhaps a slight trend
upward. Therefore, discharge data from all years were used for computing seasonal and

monthly averages.

Mean annual streamflow (m3/s]
E

0 i—T W'TTV FTTTF" YTTT llTfl‘TT—TTTTY’WT WW’T—l—r—T—T—WTT'T TW' TWT‘T—TTTT.

1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

Year

Figure 3.4 Mean annual river discharge for the lower Grand River at Grand Rapids, MI,

between 1930 and 2008.

CSO monitoring data in Grand Rapids between summer 2000 and 2008, were

obtained from CSO and $80 annual reports published by the Michigan Department of

Environmental Quality. Total CSO discharge volume between year 2000 and 2008 are

116

 

shown in Figure 3.5. Though the number of CSO outlets in Grand Rapids has decreased
fiom 26 (untreated) in 1988 to 8 (7 unheated, 1 partially treated) in 2008, total CSO
volume fluctuated from year to year, influenced heavily by annual precipitation (MDEQ
2008). The highest total CSO discharge volume was observed in 2008 (8.22 x 105 m3),
followed by 2004 (7.45 x 105 m3), and 2001 (2.38 x 105 m3), respectively. Total and
mean CSO discharge followed similar trends except in January, in which an elevated
mean in CSO discharge was observed (Figure 3.6). This was probably caused by an

unusually high single event discharge volume during snowmelt in January.

mm]

2003 2004 2005
Year

cs0 volume (loglo m3)

 

 

 

Figure 3.5 Annual CSO volume discharging into the lower Grand river in Grand Rapids
area, MI, between year 2000 and 2008.
* Partial year data were used for year 2000 because CSO reporting in Grand Rapids

started in July 2000.

117

 

 

 

4304 4 lb 8.E+05
A —-— Mean Discharge A
to -. l ._ to
5 4.304 \ -o—Total Discharge 7.E+05 g
0 3.E+04 -. 4— 6.E+05 0
2" \ E
3 3.13404 ~~ \ ~~ S.E+05 g
2 .1 ‘ L. .g
c: 25404 ~ \ I \ ~ 4.5+05 :
8 2.5404 4 \ / \ /\ 4» 3.E+05 8
U \ I \ /\ l U
: 1.5+04 «~ A \ / \ / \ . 2.E+OS .5
u l / ‘ .
2 5.5403 4L r \ \ \ l.E+05 ,2
0.5400 e 4 4 a 4 4— 4 0.E+00
& ‘6 6 0‘ 0‘ "
y“ 0&6 Vs...» t9 go (at (o @435 o 00(0) 400” a?"
x 3 Month 9‘9 é" 0°“

Figure 3.6 Mean and total CSO discharge volume into the lower Grand river from CSO
outlets in Grand Rapids, MI, by month. Mean and total CSO volume generally followed

the same trend except in January, when mean CSO volume was the highest.

Comparison between monthly means for river discharge and CS0 discharge
volume showed that they did not follow the same trend. Mean river discharge peaked in
March and then declined until it hits the lowest point in August (Figure 3.7). However,
monthly average CSO discharge showed several peaks over a year, suggesting that it had
little correlation with river discharge. The highest peak for CSO discharge occurred in
January, followed by four smaller peaks in May, July, October and March. This
phenomenon can be explained by climatological events such as snow melt during end of
Winter (between January and March) and high precipitation in Summer (between May
and July). Correlation analysis shows that mean river discharge and mean CSO volume

were not significantly related to each other.

118

In addition, monthly average CSO discharge duration did not show any apparent
trend over the year (Figure 3.8). Average CSO discharge durations ranged between 0.66
and 5.17 h, with an overall average of 3.88 h. Analysis of variance did not show
significant difference between monthly average for river discharge, CSO volume and
CS0 duration. This suggested that monthly data fluctuated regularly and were not ideal

for predicting final virus concentration in relative to climatological events.

 

 

 

 

250 ~ 40000
—‘_ Strealllflow [- A
__ n
7'? 200 -_ A \ -—o—CSODischarge - 35000 E
R
5
3 150
O
"E 4
a 100 i
0
‘m’
50 «L
o 4
-\ t
0 ‘6 ‘- 6 0‘ 0‘
5°“ 0’6 $29 v9 445‘ at“ 8* 0"" 03° ~5° é° e5?
5 Month 9° *3

Figure 3.7 Mean river discharge (m3/s) versus mean CSO discharge volume (m3).

Correlation analysis did not show any significant correlation between river discharge and

CSO discharge.

119

CSO duration (b:

1—4

 

 

0 l l T I
d 6’ <\\ «50 V3 \. 6 o" 0" " $
or f 0" vs 0 0“” 5° ,0" ,0 0° s .3” 0“"

\‘i’ <5th
Figure 3.8 Average CSO discharge duration (h) by month, no apparent trend was

observed for CSO discharge duration by month.

Analysis of seasonal means of both river discharge and CS0 discharge volume
gave a more apparent trend. Analysis of variance (AN OVA) showed that there was a
significant difference in seasonal river discharge (p-value:<0.001) (Table 3.5). According
to Tukey's honestly significant difference procedure, river discharge in Spring and Winter
were significantly higher than river discharge in Summer and Fall.

It can be observed from Figure 3.8 that average CSO discharge volume in Spring
was higher than other seasons; however, AN OVA concluded that average CSO discharge
volume was not significantly different seasonally (p-value: 0.53). The lack of significant
difference in average seasonal CSO discharge volume may be highly attributable to large
variances (range of CSO discharge volume within a season). In addition, seasonal CSO
discharge durations had a relatively smaller range (between 3.3 and 4.6 h), and were not

significantly different from each other (p-value: 0.57) (Figure 3.9).

120

 

Table 3.5 Overall and seasonal mean, and standard deviation for river discharge, CSO

volume and CSO discharge duration. p-values for AN OVA results were listed in column

four. Significant differences in means were observed in seasonal river discharge.

 

 

 

 

Variable Mean Standard p-
deviation value

River discharge 106.64 33.50 <0.01
(m3/s)

Spring 180.42 123.19

Summer *66.87 50.31

Fall *67 .93 56.01

Winter 110.39 85.01
cso volume (111’) 9332.14 38042.53 0.53

Spring 15576.51 55716.48

Summer 6991.80 31986.31

Fall 5210.13 21862.10

Winter 12631.63 28564.44
CSO duration (11) 3.88 5.47 0.57

Spring 4.63 6.79

Summer 3.31 4.71

Fall 4.17 5.37

Winter 3.34 4.28

 

' average river discharge for Summer and Fall were significantly different fiom river

discharge for Winter and Spring.

121

200 r T 18000
‘ -Streamflow l

  

 

 

 

 

‘7 4
I; 180i —o—CSODischarge 1 16000
«:1 1601 14000
‘3’ ”01 l 12000
o 120 T i
r: , - 10000
g 100 + l
o l + 8000
8° r I
g 60 l T 6000
9 i t 4000
E 40 J ..... l‘
20 f l 2000
0 1. - ‘1 0
Spring Summr Fall Winter Overall
Season

Figure 3.9 Mean river discharge (m3/s) and mean CSO volume (m3) by season.

. . 5
E l
€41
g l
a 31
3 l
:421
8 1
U 1
H
t l
E 0
Spring Sumner Fall Winter Overall
Season

Figure 3.10 Mean CSO discharge duration (h) by season.

122

Discharge (log

m3)

 

In order to determine the importance of site specific parameters on output
distribution of virus concentrations, parameters such as virus concentration in CSO
discharge, CSO discharge volume, CSO discharge duration, river discharge, river width
etc with their respective mean, standard deviation and distribution type were entered into
eq. (3) (p. 103), and randomly sampled using Monte Carlo simulation. A minimum of
10,000 iterations was performed for each simulation. Monte Carlo simulation showed
that in general, CSO discharge volume, virus concentration in CS0 discharge and river
discharge were the three most influential factors in deciding final virus concentration at a
location. Total uncertainty attributable to each factor is distance, and time related. For
example, at 50 km fi'om CSO discharge points, CSO volume contributes to approximately
20 % of total uncertainty in final virus concentration, while initial virus concentration and
river discharge each contributed to approximately 8 % and 2 % of total uncertainty. Virus
inactivation and sunlight irradiation had minimal effect on final virus concentration

(contribute to less than 1 % of total uncertainty).

Conclusions
In conclusion, this study found that the transport behavior of bacteriophage P22
correlated well with RWT, a conventional tracer and was a suitable tracer in a complex
fresh water system as indicated by the travel times, reach-averaged velocities, percent
mass recovery, and response to other environmental factors (residence times, storage
zone characteristics and dependence on land use patterns were examined by Shen et al.

2008). The fact that bacteriophage P22 has a lower recovery rate than RWT does not

123

affect its candidacy as an ideal tracer for microbial pathogens because its behavior may
be more representative of colloid substances, i.e. microorganisms, in surface water.
Bacteriophage P22 has inactivation rates between 0.27 and 0.57 per day and these
values were similar to the values reported in other studies (Schijven et al. 1999).
Parameters obtained or sorted in this study (i.e. virus inactivation rates, CSO discharge
volume, CSO discharge duration, river discharge etc) were evaluated and incorporated
into a risk assessment model to estimate risk of recreation in the lower Grand River in the

aflermath of a sewage overflow (Chapter 4).

124

 

References

Abudalo, R. A., Y. G. Bogatsu, J. N. Ryan, R. W. Harvey, D. W. Metge and M.
Elimelech (2005). "Effect of ferric oxyhydroxide grain coatings on the transport
of bacteriophage PRDl and Cryptosporidium parvum oocysts in saturated porous
media." Environ. Sci. Technol. 39(17): 6412-6419.

Ackermann, H.-W. and M. S. DuBow (1987). Chapter 2 Bacteriophage Taxonomy. In
Viruses of Prol_<_arvotes. V1 General properties of bacteriophages, 1:15-20. Boca
Raton, Florida, CRC Press, Inc.

Adams, M. H. (1959). Bacteriophages, Interscience Publishers, New York.

Bales, R. C., S. Li, K. M. Maguire, M. T. Yahya, C. P. Gerba and R. W. Harvey (1995).
"Virus and bacteria transport in a sandy aquifer, Cape Cod, MA." Ground Water
33(4): 653-661.

Belnap, D. M. and A. C. Steven. (2000). "Déja vu all over again': the similar structures of
bacteriophage PRDl and adenovirus." Trends Microbiol. 8: 91-93.

Blanford, W. J ., M. L. Brusseau, T. C. Jim Yeh, C. P. Gerba and R. Harvey (2005).
"Influence of water chemistry and travel distance on bacteriophage PRD-l
transport in a sandy aquifer." Water Res. 39(11): 2345-2357.

Bosch, A. (1998). "Human enteric viruses in the water environment: a minireview." fl
Microbiol. 1: 191-196.

Brion, G. M., N. B. O'Banion and G. L. Marchin (2004). "Comparison of bacteriophages
for use in iodine inactivationzbatch and continuous flow studies." J Water Health
2(4): 261-266.

Collins, K. E., A. A. Cronin, J. Rueedi, S. Pedley, E. Joyce, P. J. Humble and J. H.
Tellam (2006). "Fate and transport of bacteriophage in UK aquifers as surrogates
for pathogenic viruses." Eng. Geol. 85(1-2): 33-38.

Dowd, S. E., S. D. Pillai, S. Wang and M. Y. Corapcioglu (1998). "Delineating the
Specific Influence of Virus Isoelectric Point and Size on Virus Adsorption and
Transport through Sandy Soils." Appl. Environ. Microbiol. 64(2): 405-410.

125

 

Drees, K. P., M. Abbaszadegan and R. M. Maier (2003). "Comparative electrochemical
inactivation of bacteria and bacteriophage." Water Res. 37(10): 2291-2300.

Enriquez, C., A. Alum, E. M. Suarez-Rey, C. Y. Choi, G. Oron and C. P. Gerba (2003).
"Bacteriophages M82 and PRDl in turfgrass by subsurface drip irrigation." J
Envir. Engrg. 129(9): 852-857.

Ferguson, C., A. M. de Roda Husman, N. Altavilla, D. Deere and N. Ashbolt (2003).
"Fate and transport of surface water pathogens in watersheds." Crit. Rev. Environ.
Sci. Technol. 33(3): 299-361.

Gerba, C. P., K. R. Riley, N. Nwachuku, H. Ryu and M. Abbaszadegan. (2003).
"Removal of Encephalitozoon intestinalis, calicivirus, and coliphages by
conventional drinking water treatment." J Environ. Sci. Hlth. A38: 1259-1268.

Greenberg, A. E., L. S. Clesceri and A. D. Eaton, Eds. (1992). Standard methods for the
examination of water and wastewater. Washington, DC, American Public Health
Association, American Water Works Association, and Water Environment
Federation.

Griffin, D. W., K. A. Donaldson, J. H. Paul and J. B. Rose (2003). "Pathogenic human
viruses in coastal waters." Clin. Microbiol. Rev. 16(1): 129-143.

Harden, H. S., J. P. Chanton, J. B. Rose, D. B. John and M. E. Hooks (2003).
"Comparison of sulfur hexafluoride, fluorescein and rhodamine dyes and the
bacteriophage PRD-l in tracing subsurface flow." J Hydrol 277(1-2): 100-115.

Harvey, R. W. (1997). "Microorganisms as tracers in groundwater injection and recovery
experiments: a review." FEMS Microbiol. Rev. 20(3-4): 461-472.

Harvey, R. W. and J. N. Ryan (2004). "Use of PRDl bacteriophage in groundwater viral
transport, inactivation, and attachment studies." FEMS Microbiol. Ecol. 49(1): 3-
16.

Harvey, R. W. and J. N. Ryan (2004). "Use of PRDI bacteriophage in groundwater viral
transport, inactivation, and attachment studies." Fem_sLMicrobiol. Ecol. 49(1): 3-
1 6.

126

 

Hodgson, C. J ., J. Perkins and J. C. Labadz (2004). "The use of microbial tracers to
monitor seasonal variations in effluent retention in a constructed wetland." Water
1Q; 38(18): 3833-3844.

Jenkins, T. M., T. M. Scott, M. R. Morgan and J. B. Rose (2005). "Occurrence of
alternative fecal indicators and enteric viruses in Michigan rivers." J. Great Lakes
Res. 31: 22-31.

John, D. E. and J. B. Rose (2005). "Review of factors affecting microbial survival in
groundwater." Environ Sci Technol 39(19): 7345-7356.

Loveland, J. P., J. N. Ryan, G. L. Amy and R. W. Harvey (1996). "The reversibility of
virus attachment to mineral surfaces." Colloids Surf. 107: 205-221.

Lukasik, J ., M. L. Bradley, T. M. Scott, M. Dea, A. Koo, W.-Y. Hsu, J. A. Bartz and S.
R. Farrah (2003). "Reduction of poliovirus 1, bacteriophages, Salmonella
Montevideo, and Escherichia coli 0157:H7 on strawberries by physical and
disinfectant washes." J Food Protect. 66(2): 188-193.

Lukasik, J ., T. M. Scott, D. Andryshak and S. R. Farrah (2000). "Influence of salts on
virus adsorption to microporous filters." Appl. Environ. Microbiol. 66(7): 2914-
2920.

Martin, R. and A. Thomas (1974). "An example of the use of bacteriophage as a
groundwater tracer." J Hydrol. 23(1-2): 73-78.

Nicosia, L. A., J. B. Rose, L. Stark and M. T. Stewart (2001). "A field study of virus
removal in septic tank drainfields." J Environ Qual 30(6): 1933-1939.

O'Loughlin, E. M. and K. H. Bowmer (1975). "Dilution and decay of aquatic herbicides
in flowing channels." J Hydr0126: 217-235.

Paul, J. H., J. B. Rose, J. Brown, E. A. Shinn, S. Miller and S. R. Farrah (1995). "Viral
tracer studies indicate contamination of marine waters by sewage disposal
practices in Key Largo, Florida." Annl. Environ. Microbiol. 61(6): 2230-2234.

Paul, J. H., J. B. Rose, S. C. Jiang, X. Zhou, P. Cochran, C. Kellogg, J. B. Kang, D.
Griffin, S. Farrah and J. Lukasik (1997). "Evidence for groundwater and surface
marine water contamination by waste disposal wells in the Florida keys." Water
1% 31(6): 1448-1454.

127

Pieper, A. P., J. N. Ryan, R. W. Harvey, G. L. Amy, T. H. Illangasekare and D. W.
Metge (1997). "Transport and recovery of bacteriophage PRDl in a sand and

gravel aquifer: effect of sewage-derived organic matter." Environ. Sci. Technol.
31(4): 1163-1170.

Rossi, P., N. Dorfliger, K. Kennedy, 1. Muller and M. Aragno (1998). "Bacteriophages as
surface and ground water tracers." Hvdrol Earth Svstem Sci. 2(1): 101-110.

Runkel, R. L. (1998). One-dimensional transport with inflow and storage (OTIS)—A
solute transport model for streams and rivers, p. 1-73, U.S. Geol. Surv. Water
Resour. Invest. Rep.

Ryan, J. N., R. A. Ard, M. Elimelech, R. W. Harvey and P. R. Johnson (1999).
"Bacteriophage PRDl and silica colloid transport and recovery in an iron oxide-
coated sand aquifer." Environ. Sci. Technol. 33(1): 63-73.

Ryan, J. N., R. W. Harvey, D. Metge, M. Elimelech, T. Navigato and A. P. Pieper (2002).
"Field and laboratory investigations of inactivation of viruses (PRDl and MS2)
attached to iron oxide-coated quartz sand." Environ. Sci. Technol. 36(11): 2403-
2413.

Schijven, J. F. and M. Hassanizadeh (2000). "Removal of viruses by soil passage:
overview of modeling, processes, and parameters." Crit. Rev. Environ. Sci.
Technol. 30(1): 49-127.

Schijven, J. F., S. M. Hassanizadeh and H. A. M. de Bruin (2002). "Column experiments
to study nonlinear removal of bacteriophages by passage through saturated dune
sand." J Contam. Hvdrol. 58(3): 243-259.

Schijven, J. F., S. M. Hassanizadeh and R. H. A. M. de Bruin (2002). "Two-site kinetic
modeling of bacteriophages transport through columns of saturated dune sand." 1
Contam. Hvdrol. 57(3-4): 259-279.

Schijven, J. F., G. Medema, A. J. Vogelaar and S. M. Hassanizadeh (2000). "Removal of
microorganisms by deep well injection." J Contam. Hydro]. 44(3-4): 301-327.

Schijven, J. F. and J. Simunek (2002). "Kinetic modeling of virus transport at the field
scale." J Contain. Hvdrol. 55(1-2): 113-135.

128

 

Schijven, J. F., W. Hoogenboezem, S. M. Hassanizadeh and J. H. Peters. (1999).
"Modeling removal of bacteriophages M82 and PRDl by dune recharge at
Castlicum, Netherlands." Mter Resour. Res. 35: 1101-11 1 1.

Scott, T. M., J. Lukasik and S. R. Farrah (2002). "Improved method for recovery of
bacteriophage from large volumes of water using negatively charged microporous
filters." C_an. J Microbiol. 48(4): 305-310.

Shen, C., M. S. Phanikumar, T. T. Fong, I. Aslarn, S. P. Mthnurry, S. L. Molloy and J.
B. Rose (2008). "Evaluating bacteriophage P22 as a tracer in a complex surface
water system: the Grand River, Michigan." Environ. Sci. Technol. 42(7): 2426-
243 l .

Steinbacher, S., S. Miller, U. Baxa, A. Weintraub and R. Seckler. (1997). "Interaction of
Salmonella phage P22 with its O-antigen receptor studied by X-ray
crystallography." Biol. Chem. 378: 337-343.

Van Cuyk, S., R. L. Siegrist, K. Lowe and R. W. Harvey (2004). "Evaluating microbial
purification during soil treatment of wastewater with multicomponent tracer and
surrogate tests." J Environ. Gull, 33(1): 316-329.

Zinder, N. and J. Lederberg (1952). "Genetic exchange in Salmonella." J Bacteriol. 64:
679-699.

129

 

CHAPTER 4 i
RISK OF GASTROINTESTINAL DISEASE ASSOCIATED WITH EXPOSURE TO
ENTERIC VIRUSES AT RECREATIONAL BEACHES DOWNSTREAM OF CSO

SITES ON THE LOWER GRAND RIVER, GRAND RAPIDS, MICHIGAN

Abstract

In cities with combined sewer systems, intentional discharge of untreated waste
water directly into surface water during heavy precipitation events has been identified as
one of the main culprits for surface water quality degradation. The objective of this study
was to evaluate the risk of viral gastrointestinal disease associated with combined sewer
overflow (CSO) discharges into the lower Grand River in Grand Rapids, Michigan.
Concentrations of human adenoviruses in surface water (n=20), raw sewage (n=13) and
CSO discharge (n=6) of Grand River were measured and used in exposure assessment.
Probability of enteric virus infection fiom swimming and wading in the contaminated
surface water after a CSO event were estimated. Estimation for virus attenuation in the
river was obtained from a virus tracer study. Parameters used in the model included CSO
discharge duration, CSO discharge volume, virus concentration in CS0 discharge, .
hydrogeometry of the river segment (i.e. river discharge, depth, width, cross-sectional
area and slope etc.) and sunlight intensity. Risk estimations were performed on five
exposure scenarios: (i) swimming in a recreational park 25 km downstream fi'om a CSO
outlet, (ii) swimming in Lake Michigan beaches 50 km downstream from a CSO outlet,
(iii) wading/angling in a recreational park 25 km downstream from a CSO outlet, (iv)

wading/angling in Lake Michigan beaches 50 km downstream from a CSO outlet, (v)

130

swimming in the recreational parks or Lake Michigan beaches without a CSO event.
Dose-response relationships for rotavirus and echovirus 12 were used in the study for
comparing risk differences between a highly infectious and a moderately infectious virus.
A point risk estimate was conducted using the means of input variables. The point risk
estimate produced mean probability of infection using the rotavirus model of 2.9x10'l
and 2.2x10'l for swimming and wading, respectively, with a duration of elevated risk
(when illness risk is greater than EPA recreational freshwater level of 8 x 103) that began
at 12h after discharge and ended at 33.5 h after discharge 25 km from discharge point. At
the same location, the echovirus 12 model estimated a risk of 5.4x10’2 for swimming and
1.5x10’2 for wading, with duration of elevated risk between 14.5 h and 27.5 h after
discharge. An interval risk estimate was performed using Monte Carlo techniques to
characterize uncertainty in input variables (i.e. CSO discharge duration, CSO discharge
volume etc). Overall, point risk estimates gave higher mean risk values than interval risk
estimates. Using river discharge and CSO parameters for summer, interval risk estimate
gave durations of elevated risk that lasted up to 200 and 350 hours at 25 km and 50 km
downstream from the discharge point, respectively. Sensitivity analysis showed that both
probability of infection and duration of elevated risk (duration after CSO discharges
when risk of illness > 0.008) at a location were highly dependent upon river discharge.
With the use of this risk analysis model, effects of CS0 discharge on surface water
quality in other watersheds can be estimated as long as site specific parameters are

available. Long-terrn virus survey data are essential for better estimation of seasonal risk.

131

 

 

Introduction

1“ -.

Enteric viruses (i.e. adenoviruses, enteroviruses, rotaviruses) are frequently
detected in surface water contaminated by fecal sources, and are known to cause not only
gastroenteritis, but also a wide spectrum of other diseases, such as conjunctivitis,
respiratory infections, myocarditis, encephalitis, and diabetes, in infected individuals.
The presence of these enteric viruses in surface water designated for recreational purpose
elevates public health risk associated with ingestion of contaminated water during
recreational activities.

In the United States, major sources for enteric viruses in the surface water are
urban nmoff, agricultural runoff, discharges from wastewater treatment plants, and in
cities with older sewer systems, discharges from combined sewer systems (CSSs) and
sanitary sewer systems (SSSs). Currently, there are approximately 770 CSS communities,
mainly concentrated in the Northeast and the Great Lakes regions, serving a total of 40
million people in the country. The Federal Govemment’s effort to control CSOs started in
1994, when the U. S. EPA published CSO Control Policy as the national fiamework. In
Michigan, the first CSO policy was drafted by the Department of Environmental Quality
in 1983. However, the first non-contested permit requiring a long term CSO correction
program was issued to the Grand Rapids waste water treatment plant (WWTP) only in
Fall 1988, following a large CSO event in the city that affected water quality downstream
in Grand Haven (MDEQ 2008). To date, Michigan communities have eliminated 75% of
the 613 untreated CSO outfalls that existed in the year 1988 and the remaining 25% are
scheduled for correction/elimination through implementation of long term control plans

(LTCPs). However, water quality after CSO or any sewage spill remains a public health

132

 

concern to the individuals recreating at recreational parks and beaches downstream of
discharge sites. High concentrations of waterborne pathogens, especially viruses, in
CSOs, may pose adverse risk to human health.

Risk assessment is a valuable tool for estimating adverse effects associated with
certainhazards and has been used to estimate health risk fi'orn microorganisms and
chemicals. It is often an initial step prior to risk management, for example, in the case of
wastewater discharge, providing important information for setting treatment criteria.
Quantitative risk assessment gives a probability and describes the magnitude of a
hazardous event based on exposure patterns and dose-response data. In the field of water
quality, quantitative risk assessment has been used for the interpretation of microbial
water quality data to estimate the public health effects of low levels of waterborne
pathogens in the water (Donovan et a1. 2008; Haas et al. 1993; Soller et a1. 2003; van
Heerden et al. 2005). Previously, the greatest limitation and uncertainty for risk
assessments has been the lack of information about survival, transport, occurrence of and
exposures to key specific infectious agents of concern.

The objective of this analysis was to characterize and quantify virus-related
gastrointestinal disease risk for users of recreational parks and beaches located
downstream of CS0 discharge sites in Grand Rapids following a CSO event. Grand
Rapids, the second largest city in the state of Michigan, is the first city in the state of
Michigan that implemented a long term CSO control plan, partly because of the
recreational activities taking place in the lower Grand River. Grand Rapids is located
approximately 50 km upstream from the mouth of the lower Grand River that discharges

into Lake Michigan. In 2007, the number of CS0 outfalls in the city of Grand Rapids was

133

reduced from 26 to eight (one partially treated and seven untreated), and released an
annual discharge of 6100 m3 partially and untreated sewage into the river (MDEQ 2008).
The landuse directly downstream of the CSSs is mainly urban (52.2%) and slowly
transforms to agricultural (69.2%) when approaching the mouth of the river. In this
analysis, representative concentrations of viruses in the river during baseflow and alter
CSO events were calculated from existing virus monitoring data (Chapter 2). Estimation
for virus attenuation while being transporting down the river were obtained from a tracer
study, in which bacteriophage P22 was used as a biotracer (Chapter 3). Probabilities of
gastrointestinal infection and illness from recreating at several recreational parks and
beaches of the lower Grand river in Grand Rapids and Grand Haven, MI, following CSO
events were estimated based on established dose-response relationships for several
waterborne enteric viruses (i.e. human rotavirus (high risk), human echovirus 12

(moderate risk)) representing different risk levels (Haas et al. 1993).

Materials and Methods
Hazard identification. Human adenoviruses were chosen as the index viruses for

this study because of their frequent isolation and higher persistence compared to other
enteric viruses in aquatic environments. In addition, the use of quantitative PCR assay
allowed easy enumeration of this virus in the environment. Outbreaks of adenoviruses
associated with swimming in contaminated swimming pool water has been reported
(Turner et al. 1987; Papapetropoulou and Vantarakis 1998). Route of transmission for
adenoviruses include inhalation, ingestion and dermal contact and for recreational water,

ingestion of fecal contaminated water is considered a primary route. This is supported by

134

 

 

the monitoring data and literature with enteric adenoviruses, including adenovirus 41 as J
predominant viruses detected (refer to Chapter 2). In this analysis, adenovirus was chosen
as the worst case scenario representative because of their high occurrence and persistence
in environmental water. An end point of gastroenteritis with a fecal-oral transmission
route was used as the health hazard as AGI is the most commonly reported virus
attributable illness associated with recreational water. Literature on waterborne
adenovirus outbreaks report attack rates around 50% (Gray et a1. 2007).
Dose response determination. A number of human and animal studies had been
conducted to determine the infectivity of viruses on their hosts (Couch 1966; Katz and
Plotkin 1967; Supter 1963). The risk of infection based on dose-response models for
several enteric viruses, representing different potencies are listed on Table 4.1. Dose-
response models for these viruses generally were obtained fiom human feeding studies,
with the exception of Coxsackievirus dose response study (Suptel 1963).
The exponential and beta-Poisson models have been found to provide the best fit
to the experimental data. Each model is set up according to several assumptions.
Exponential model assumes that one microorganism is capable of initiating an infection
and that host-microorganism interactions are constant. In exponential model, “r”
represents the proportion of ingested microorganisms that survive to cause an infection.

In Table 4.1, adenovirus 4 and poliovirus I are the two viruses that have exponential dose

response models. Exponential models are represented by the equation below:

135

Pi(N)=1—-exp(—rN) (1)
where:

P, : probability of infection;

N: number of organisms ingested or inhaled

r = proportion of ingested microorganism that survive to cause an infection

A beta-Poisson model, on the other hand, assumes heterogeneity for either the
infectivity of individual microorganisms or host-microorganism interactions. A beta-
Poisson model differs from an exponential model in that it is characterized by two
parameters, [3 and a. As (1 increases, the model becomes closer to an exponential model

(Haas et al. 1993). The equation for a beta-Poisson model is:

N-
P,N=1—1—“ 2
() (+fl) ()

where:

N: number of viruses ingested
a, [3: parameters define the dose-response curve

The most potent virus group, which had the lowest infectious level in Table 4.1, is
an adenovirus. The exponential dose response relationship for adenoviruses was
developed from a study evaluating human inhalation of adenovirus 4 particles by Couch
et a1. (1966) (Haas et al. 1993). However, because the route of exposure for the
adenovirus 4 study was inhalation, it is not ideal model to represent risk through
ingestion and for enteric viruses. The second most infectious virus in Table 4.1 is
poliovirus HI. This dose response study conducted by Katz and Plotkin (1967) is

inappropriate and may over estimate risk in our case, because it was performed by direct

136

 

 

 

delivery of attenuated poliovirus into stomach of premature infants, a very susceptible J
population group. Consumption of one coxsackievirus B4 virus gave a probability of
infection of 8.0 x 103, however, it may not be suitable for studying human health risk as
this dose response relationship for coxsackievirus B4 was extracted fiom a mouse
inhalation study and mortality was used as the end point of analysis (Suptel 1963).
Overall, the rotavirus model seemed to be the most suitable one for estimating risk
caused by a highly infectious enteric virus because it is the most potent virus in which
study was conducted on adult human subjects through a fecal oral route. Echovirus 12,
which is approximately 100 times less potent than rotavirus when a single particle is
consumed, was chosen to represent moderately infectious viruses. Echovirus 12 model
resemble the rotavirus model in which it is also conducted on adult human subjects and
doses were administered orally. Table 4.1 provides the point estimates for parameters in
several virus dose response models and comparison of probability of infection when

consuming or inhaling one viral particle.

 

137

.83: .3 3 33m 3.3 98: .3 ... 38. 88.. 3833.. 33... 33.32 388333. 33.3 383......

 

 

$2 93...... E: u n

n 88.0 .3 3 32.3 .38.... 38-38. .32... u ... 88338.88 . 8.383..
8.2: u a

e .38... 3.: 35% 8.8.... 38.38.. in: u ... 5838.88 2 3.883....

«ES... 5

m :8... 8% 388m 88. 88338.. u . 3.82888 8853.388
$2 .888. 888...

3 good .3 go .852 525.: 38-38.“ H .— 3355qu H Eamon
33 NY: n a

n $2... .3 3 .....B .38.... 38.38. .3"... u a 5338.80: ӣ383.
$2 882.. A888. .33 um

N 53... 33 83. .38.... 3.938.. .83. n ... 83.8-33 B 3.8.6.3..

. :88 88. .388 .38.... 8.3388 N23. n .. 333.825 3 3.8.683.

T .2. 25>
wan—=2 ”ES—Ema as 8:3— 9—32:32— .25.:
«Ey— xma 38:233— .252 83235829 .232 8532.989 25>

 

$2.83 30.35% wcscomoaoc 83 2.835 3.33 .8 2032.: 3:38.. omen 3. 033.

138

Exposure assessment. The third step of risk assessment is exposure assessment,
which was determined by estimating the number of virus ingested during recreational
activities. Exposures were determined by the magnitude of a contamination event (virus
concentration in CSO discharge, CSO discharge volume and duration of discharge),
distance from overflow outlet and travel time, transport and survival properties of viruses
as well as volume consumed during recreational activities. In this study, several exposure
scenarios were set up to elucidate the influence of different factors on the magnitude of
risk: (i) swimming in a recreational park 25 km downstream fiom a CSO outlet, (ii)
swimming in Lake Michigan beaches 50 km downstream from a CSO outlet, (iii)
wading/angling in a recreational park 25 km downstream from a CSO outlet, (iv)
wading/angling in Lake Michigan beaches 50 km downstream from a CSO outlet, (v) risk
of swimming in the recreational parks or Lake Michigan beaches without a CSO event.

Site-specific parameters such as river discharge, CSO discharge volume, CSO
discharge duration, area of channel and slope were either extracted from historical data or
estimated from field measurements. Information regarding CSO discharge volume and
discharge duration from CSO outlets in Grand Rapids between 2000 and 2008 were
obtained from Michigan Department of Environmental Quality
(http://www.deq.state.mi.us/csosso/). The State of Michigan required municipalities to
report CSO events starting in summer 2000. Historical river discharge data from 1930 to
2007 for the USGS gage at Grand Rapids (hydrological station: USGS#04119000) were
obtained from the USGS website
(http://waterdata.usgs.gov/nwis/dv?referred_module=sw&site_no=041 19000).

Inactivation caused by sunlight irradiation on a specific day was determined by

139

 

computing the mean sunlight intensity of the day using SolarCalc 1.0 from USDA-ARS
web page (http://www.ars.usda.gov/services/software/download.htm?softwareid=62).
Mean attenuation caused by factors other than sunlight was estimated using data obtained
from the tracer study (Chapter 3).

Information regarding the occurrence of enteric viruses was essential for risk
assessment. In this study, virus concentration data were obtained from a monitoring study
of the Grand river. Concentration of viruses in waste water, surface water and CSO
discharge of the Grand River were quantified using real-time PCR, detailed protocols and
results are described in Chapter 2. Based on the consideration that DNA in viruses with
damaged viral capsids are reported to degrade within a short period of time, one copy of
virus DNA was used to represent one viable virus in this risk estimate (Wetz et al. 2004).

Distribution of all exposure parameters included river discharge, CSO discharge
volume, CSO discharge duration, virus concentration in CSO and volume consumed
during a recreational event, and were determined using EasyFit software (MathWave
Technologies, Spokane, WA).

Mean, standard deviation and range of virus concentrations in surface water and

CSO discharges are presented in Table 4.2. The initial concentration of viruses in the

river after CSO discharge, C0 , was estimated using the formula below:

C0 = CCSO ' QCSO (3)
(QCSO + QR)

 

where CCSO denotes average concentration of viruses in CSO discharge (pfu/ml); QCSO

is the flow rate of CSO discharge (m3/s); QR is the flow rate of the river (m3/s).

140

The changes in concentration of viruses as they were transported downstream

were computed using an analytical formula developed by O’Loughlin and Bowmer

 

 

 

 

 

 

(1975»
{-U—x-(l-F )- —Ut1“ x-U(t-r)l“ _ ‘
e2E erchL—J—erfCE ] +

CO 25 2430—7)

C(x,t) =T< _ d }
e%(1+r )rerfc[x+UtF)_erfc x+U(t—r)I' ,
l _ 25 2450-1) _ J

where l"=,/1+477 and 7):-[EE—v (4)

where, C(x,t) denotes tracer concentration as a fimction of distance traveled, x (m), and

elapsed time since tracer release, t (s); t represents spill duration (8); C0 is the initial

virus concentration at discharge point (virus/ml'l); U is average velocity (ms'l);
E is longitudinal dispersion coefficient (mzs'l); k is first order decay rate constant (5").

Number of viruses ingested, N, were calculated based on the formula:

N=C(x,t) *V (5)

where, V denotes volume of contaminated water consumed (Table 4.3).

141

 

 

 

 

8333233 OmU
382 3.8 3.83 83mm 33 .3328: ..m 3.332
3% 88v 3% N: 0.08m 832
83..
we... 88v 8... a: 833m 882
V3..
RS 8... .33. S 35. 83.0
83..
3.9 88v 88v mo 0382.3.
83..
a... 88v 93 m: .33.... 3.5
83..
3.8 88v 3. ...: 3.5 :35
63>» 32% 3:30
8.8 58v ”.2 3: one 0335 3:3
...—.33
2.8.2
E on 32 83 S .3338 .38 83.6
mu. 2: a. 3% 32 82.5.".
85?.»
5% SN 2: an: 23. .3338 .5. «82.3 33.8
883369 388=8 29:3» .
8.5582 SEE-:2 9:655 :32 he 523.383 093 .8332

Awaken—rat 55358350 «>53.— ue .02 2955 ufiifium

.mowhfiommv OmU 3:3 .533 .83.. 8833833 E 33238350 3.8308033 mo owns“ 383 count/ow 3.835% .532 N...» 2an

142

Risk characterization. The final step in a risk assessment is to combine
information from the previous three steps: hazard identification, dose response
assessment and exposure assessment to characterize health risk associated with a
specified hazard. In this study, risk estimation was based on two estimation models,
comparing both point and interval estimations. A point estimate model consists of a
single numerical value for each parameter resulting in a value for probability of infection,
usually using the mean, median or on occasion, 95% confident limit. In contrast, an
interval estimate model presents either a confidence interval or a full probability
distribution of risk, taking into account distributions and uncertainties associated with
input parameters.

Point risk estimate

A point risk estimate of exposure to viruses when swimming in contaminated
water 25 km and 50 km downstream of the discharge point was computed based on mean
values for river discharge (summer), CSO discharge volume (summer) and CSO
discharge duration (overall). River discharge and CSO discharge volume values for
summer were used because they were significantly different from those values computed
for winter and spring (Chapter 3) and summer represents a key exposure period for
recreational activities, such as swimming and wading. In addition, summer was identified
as a high risk season considering a higher fi'equency of CSO events. In order to show
influences of virus concentration in CSO discharge on the risk estimate, the highest and
the lowest virus concentrations determined from monitoring CSO discharges were used.
Risk estimates were determined using the beta-Poisson dose response models for

rotavirus and echovirus 12, each representing highly and moderately infectious viruses.

143

 

Consumption volumes of 50 ml and 10 ml were used for swimming and wading/angling,
respectively (see Table 4.3; Dufour et al. 2006; Donovan et al. 2008). Point estimates
were used for infection, Pi, morbidity, Pmorbidity and mortality rate, Pmortality with
formula as described below.

The daily risk of morbidity (Pmorbidity) was calculated using a morbidity rate of

0.5 (Haas et al. 1993):

Pmorbidity = P i * 0'5 (6)
The risk of mortality was calculated using a mortality rate of 0.001 (Bennett et al. 1987):

Pmortality = Pmorbidity "' 0-001 (7)

The annual risk of infection, morbidity and mortality, Pd , were estimated based
on the formula:

Pd =1—(1-r)" (8)
where, r is the single event risk and d is the total number of days exposed.

The annual risk for swimming was estimated based on the number of CSO events
in a regular swim season in Michigan, which normally begins on Memorial day (May 31)
and ends by Labor day (September 1), a maximum of 12 events per swimming season
was used (assuming one swim event per CSO day). The annual risk for wading angler
was based on the number of CSO events during a typical fishing season in the lower
Grand River, which generally starts in mid May and ends by late September, a maximum

of 20 events per angling season was computed based on historical data. Mean sunlight

intensity (I (t) ) of 348.22 (simulated value for 07/15/2008 by SolarCalc 1.0 software)

144

 

was used in simulations. A summary of parameters and their values used in the point

estimate are listed in Table 4.3.

Table 4.3 Parameters and their associated values used in point risk estimation.

 

 

 

 

 

 

 

 

Variable Symbol Unit Parameters Notes
River Discharge QR m3/s 107
CSO discharge Vcso m3 9332
volume
CSO discharge DCSO h 4 ch0= Vcso/Dcso; CSO
duration was assumed to discharge
at a constant rate
Virus C550 viruses/ml low conc: 60; Chapter 2
concentration in high conc: 1322
CSO discharge
Water ingestion rate
Swimming V ml/event 50 Dufour et a1. 2006
Wading V ml/event 10 Donovan et al. 2008
Exposure frequency
Swimmer d Days/yr 12 Based on maximum no. of
CSO event during a swim
season
Wader/angler d Days/yr 20 Based on maximum no. of

CSO event during an
angling season

Mean sunlight 1(t) kW ‘ 348.22 Value for 7/15/2008 was
intensity used for all simulation '

 

Inactivation due k, d'IkW'I 0.4
to sunlight
irradiation

 

Inactivation due kg (1'1 0.3
to all other
factors

 

145

Interval risk estimate

Recently, interval risk estimates have been preferred by researchers because they
provide a probability range of the resulting risk, giving a better sense of precision and
distribution of the estimation.

As a first step in performing an interval risk estimate, the distribution of each
parameter was defined either based on actual data fitting or the best estimates.
Distribution type, mean and standard deviation of input parameters used in this study are
summarized in Table 4.4. Parameters used for characterizing virus transport included
river discharge, CSO discharge volume, CSO duration, and virus concentration in CSO
discharge. For estimating single event risk, accidental consumption volume for adult and
non—adult were used as determined from experimental data from Dufour et al. (2006).
Annual risk for swimming and wading referred to swimming and angling season in the
lower Michigan. Due to the lack of actual data, triangular distribution was used to
describe annual occurrence and frequency of swimming and angling events. Monte Carlo
techniques were used to characterize the uncertainty associated with parameters. Monte
Carlo techniques are computational algorithms that rely on repeated random sampling to
compute results. It has been widely used in risk assessment to generate a distribution of
derived risk caused by uncertainty and variability of data.

The exposure and risk estimation models were developed in both Excel and
Matlab (Mathworks, Inc., Natick, MA). Output distributions of viral concentrations and
risk of infections were obtained using Monte Carlo simulation techniques. A minimum of

10,000 iterations was conducted for each scenario. All distribution fimctions were

146

 

855

assumed to be statistically independent. Monte Carlo simulation was run with Risk solver

 

(Frontline Systems, Inc, Incline Village, NV), an add-on to Microsoft Excel.

147

 

Sugfimxafi

 

 

 

 

 

 

. noeoe «usage: 33935 5.38 a seeim
5:259... osmium
wocw
.a a .588 2 5:5 a 9:35.
Sea
.3 8 Sega 993-55
xv": _muQm .wmugofi annoy—mo..— ~5>0>E S wngm
meow
4m go Boga 9.63
”Qua gun—m .Eufiofi _m—Eecwoq 826:5 S wEEEEm
.3589: .533
owns—.83
OwU 5
namfiufioa grab—50:8
w" .36me .Sufie haggard. _Ehomzh> 080 25>
*«flcflahflfi
Qmunmdmmufioa 0328:.
5805 355.3 a cam 08
£2 mun—m
.Nmmwncaofi ”885w «iv—5:9»
”~48 munm.~§u§oe 99.28:.
£805 Eaton—worm ...E 08; OwU
:43qu Kwéoucmofi ..
”853m mom.mmu..Qm a 032.8%
.vwdoTfiQb ”.365 38.8ch $8 «Q ..95—
882 8:3, 55.552: :5 .253 230335

 

dosafiumo xmt 330%: 5 wow: cosmFow 6.8986 23 SEE dab noun—£36 3368mm £05 95 Eouogm v6 033.

148

 

A>ow.§wEo.§.333\\xEé 3383 309930 630m 359800 use 56mm
330 38§§>=m co Eggnog 5322 see 853% as... 838% see 3% awesome 08 .3 558 38M

.Am8\>ow.mwm:.8m28m3\\u&£ .anmmm cough—OHS 883 3:232 mOmD 69:59 .5??va .mEmam 23.5
a 882 3388 8:5. House} mam: Emcee? ace Basso 22; @3338:er 5% ”mafia b.5— .

Eugene eaegm ”am ..

 

 

 

 

£98.“ .650

=« 3 26

moon .3 3 55m o¢.o-moo.o gotch— Lu 3 sou—gags

:ouflufih

£333

3 one

wcom .3 Ho seam hvdfimd gomED TBS—Lu ax couwfiuoas

cots—25m

=m wow com: 36:85

33 832R “fins;

5 82> ~33 5 at so:
such—Sis:
.m 70on

finch—.552 E_swcath SE35 3 SE:

3:8 3. 2.5

149

Recreational Risk on Days without a CSO Event

Both point and interval risk estimates were performed for swimming and wading
in the area on days without a CSO event in order to examine the risk difference when
distributions of parameters are included in analyses. Dose response models for rotavirus
and echovirus 12 were used for risk comparisons between highly infectious viruses and
moderately infectious viruses.

In point estimate, ingestion rates of 50 ml and 10 ml were used for swimming and
wading, respectively. An average virus concentration of 7.76 viruses /ml was used. The
annual risk for swimming was estimated based on a maximum of 94 days and the annual
risk for wading was based on an annual exposure of 140 days. Partial Monte Carlo
analysis was performed using average ingestion volumes for swimming (50 ml) and
wading (10 ml), in conjunction with lognormal distribution of virus concentration in
surface water (mean: 7.76 viruses/ml, standard deviation: 19.8) . Distributions for
maximum exposure days for swimming and wading activities were also included in the
partial Monte Carlo analysis. The annual risk for swimming was estimated based on a
regular swim season in Michigan, which normally begins on Memorial day (May 31) and
ends by Labor day (September 1), with a triangular distribution, with a maximum of 94
days (min:1, mode:12, maxz94). The annual risk for wading was based on a typical
fishing season in the lower Grand River, which generally starts in mid May and ends by
late September (triangular distribution: minzl, modez27, male40). In the full Monte
Carlo analysis, distributions of water ingestion rate during swimming for adults
(mean: 19 ml, standard deviation: 19 ml) and non-adults (mean: 38 ml; standard

deviation: 31 ml) were added. All parameters used are listed in Table 4.5.

150

Table 4.5 Parameters and their associated distribution type, mean and standard deviation

used in risk estimation for recreation in surface water without a CSO event.

 

 

 

 

Parameters Symbol Unit Distribution Values Notes
Point estimate
Water
ingestion
Swimming V ml/event 50
Wading V ml/event 10 Donovan et al.
2008
Virus C0 viruses/ mean=7.76 n=20; non-detects
concentration ml were substituted
with detection
linrit’0.50
Exposure
frequency
Swimming d Days/yr 94
Wading d Days/yr 140
Partial Monte
Carlo
Virus C0 virus/ml Lognormal mean=7.76, n=20; non-detects
concentration SD=19.8 were substituted
with detection
limit*0.50
Exposure
frequency
Swimmer d Days/yr Triangular Min=1,
mode=12,
max=94
Angler d Days/yr Triangular Min=1,
mode=27,
max=l40
Full Monte
Carlo
Water
ingestion
Swimming V ml/event Lognormal mean=l9, n=12;
(adult) SD=19

Dufour et al. 2006

Swimming V ml/event Lognormal mean=3 8, n=41;
(non-adult) SD=31 Dufom et al. 2006

 

151

 

 

Sensitivity Analysis. The influence of input parameter values (i.e. mean, standard
deviation and distribution) on risk estimation were determined by sensitivity analyses.
Individual parameters were plotted against probability of infection outputs to examine if

an important trend occurred.

Results

Public health risks associated with recreation in viral contaminated surface
water were estimated for days followed a CSO event and on days without such an
event. Results of risk estimation through point risk and interval risk estimation
methods were compared. Sensitivity analyses were carried out to determine the most
influential factors contributing to health risks.
Point risk estimate

Point risk estimate of swimming at recreational sites located 25 km and 50 km
downstream of discharge point were calculated based on the mean summer river
discharge, mean summer CSO discharge volume, mean CSO discharge duration and
two virus concentrations (i.e. 6O viruses/ml and 1322 viruses/ml). Incidental
consumption volume of 50 ml/event was used. Annual risk of infection was calculated
based on a maximum of 12 swim days per year. All parameters used are summarized in
Table 4.3 (p. 143).

In general, risk of rotavirus infection is 0.7 to 1.6 loglo higher than risk of
echovirus 12 infection, based partly on the potencies of the virus models chosen for
moderately and highly infectious viruses. In the scenario in which CSO virus

concentration was 60 viruses/ml, the mean risk for rotavirus infection when swimming

152

 

 

were 1.4 x 10'1 and 8.6 x 10'2 for the 25-km and SO-km locations, respectively; the mean

 

risk of echovirus 12 infection for these two locations were 3.9 x 10'3 and 2.1 x 103. For
the scenario in which CSO virus concentration was 1322 viruses/ml, mean risks of
rotavirus infection were 2.8 x 10'I and 1.9 x 10'], while the mean risks for echovirus 12
infections were 5.4 x 10'2 and 3.2 x 102, for the ZS-km and SO-km locations, respectively.
The peak infection risk happens approximately 20 hours after the CSO event for the 25-
km location and 40 hours after the CS0 event for the SO-km location (Fig. 4.1). The
probability of infection curves for both virus groups are shown in Fig. 4.1a (ZS-km

location) and Fig 4.1b (SO-km location).

Results

Public health risks associated with recreation in viral contaminated surface
water were estimated for days followed a CSO event and on days without such an
event. Results of risk estimation through point risk and interval risk estimation
methods were compared. Sensitivity analyses were carried out to determine the most
influential factors contributing to health risks.
Point risk estimate

Point risk estimate of swimming at recreational sites located 25 km and 50 km
downstream of discharge point were calculated based on the mean summer river
discharge, mean summer CSO discharge volume, mean CSO discharge duration and

two virus concentrations (i.e. 6O viruses/ml and 1322 viruses/ml). Incidental

 

consumption volume of 50 ml/event was used. Annual risk of infection was calculated

153

based on a maximum of 12 swim days per year. All parameters used are summarized in
Table 4.3 (p. 143).

In general, risk of rotavirus infection is 0.7 to 1.6 loglo higher than risk of
echovirus 12 infection, based partly on the potencies of the virus models chosen for
moderately and highly infectious viruses. In the scenario in which CSO virus
concentration was 60 viruses/ml, the mean risks for rotavirus infection when
swimming were 1.4 x 10'1 and 8.6 x 10'2 for the 25-km and SO-km locations,
respectively; the mean risk of echovirus 12 infection for these two locations were 3.9 x
10'3 and 2.1 x 103. For the scenario in which CSO virus concentration was 1322
viruses/ml, mean risks of rotavirus infection were 2.8 x 10'l and 1.9 x 10", while the
mean risks for echovirus 12 infections were 5.4 x 10'2 and 3.2 x 102, for the 25—krn and
SO-km locations, respectively. The peak risk occurs approximately 20 hours after the
CSO event for the 25-km location and 40 hours after the CSO event for the 50-km
location (Fig. 4.1). The probability of infection curves for both virus groups are shown

in Fig. 4.1a (ZS-km location) and Fig 4.1b (SO-km location).

154

 

 

 

.7on fix me 8803 no 558305 38¢ 53, Taxman mm canoe? .«o 3:332.“ :88 .3583? 3.0 mm nougnooaoo 95> 988
A5 com .70? En mo :38? mo mum—Banana M89 £5 Tots: mm caveman: me 3:338.“ :88 4:50.045.» 86 mm momma—888 35>

808 .3 8m ..Eog owhfiomne mo 80.5838 Ex mm 03286 OmU 8...? 2:3 38> 8383 3238 me 323305 2d edema

155

EV omega? we; 8am 08E.
ow mm om mm om m fl 2 m o
, u 1 q

0

 

   
    
   

r
“i
o

.3 ;
O 0
“0:10am! JO filmqvqwd

1
°°.
O

 

 

 

 

.8888 203 8338: 82 av Ea 835:: 8 8% 828888 25.. 08 .83 0828:.

.6 88:38 3 3. 8.8 E on Ba 9 3. :8 E a .8236 08 8.. one 322, 8888 9:88 .8 8:82.: 3. 05mm

..-ans .8 888.8: .8 3:538: 8: 55 1:88; mm 8888 .8 52338: :88 ”_8\8m=.:> 36 E 888883 88> :88
A5 .5: $385.38 8888 .8 @2338: x8: 53, «-2wa mm 8888 .«o ham—BB8: :88 x8888? Ed mm 8888088 98>

888 .3 3m .88: ems—~86 808 88838 :8 on 08886 Omo 8% 083 883 88.88 2:358 mo 5:338: 3 .v oBmE

:6 8888:: 98> 88mm 885.

 

   

     

2. oo on ov cm on E o
4 q . _ 1 _ O M
O
r 1 . 0..
No m.
- -von
0
U
.. 1 B. .m...
«1w
- .. . m.
we u

_ . _ . _ .

 

 

 

156

Mean risk of infection decreased as distance from discharge point increased,
the mean virus infection risk fi’om swimming decreased approximately 0.22 loglo
(range: 0.16-0.27 loglo), over 25 km, this decrease is associated with the attenuation of
viruses during transport.

The duration of elevated risk correlated directly with virus infectivity, travel
distance, and tolerable risk levels (Table 4.6 and Table 4.7). The US EPA tolerable risk
level for recreational fi'eshwater is eight illness per 1000 recreational events (i.e.
Pmorbidity of 0.008, which is equal to Pi of 0.016 (eq. 6)). The rotavirus model, which
represents highly infectious viruses, gives a longer duration of elevated risk than the
echovirus 12 model, which represents moderately infectious viruses. At the 25-km
location, duration of elevated risk based on US EPA recreational standard for
freshwater using rotavirus model began at approximately 12 h after virus release and
ended at 33.5 h, compared to a “duration of elevated risk” that began at 14.5 h and
ends at 27.5 h after virus release for echovirus 12 model. At the 50-km location,
durations of elevated risk were 27-55 h, and 31- 48 h after virus release, for rotavirus
and echovirus 12 models, respectively. The difference between duration of risk
obtained from the different virus dose response models showed that the selection of an
appropriate dose response model is crucial in decision making regarding duration of a
beach closure to protect public health.

In cases where a more stringent standard is necessary, the US EPA tolerable
risk level for drinking water could be used as a guideline. Here, duration of elevated
risk based on tolerable risk level for drinking water (infection risk, P, of less than 1.0 x

10“1 per annum) were determined and compared. At the 25-km location, estimated

157

duration of elevated risk associated with rotavirus infection began at approximately
9.5 h after virus discharge and lasted through 39 h after virus release, compare to
duration of elevated risk that began at 12 h and ended at 33.5 h when US EPA
recreational risk standard was applied (Table 4.6 and Table 4.7).

Risk estimates for wading are lower than risk estimates for swimming at both
locations, because of a lower consumption rate used for wading activities
(10 ml/event). Risk estimates and duration of elevated risk for wading downstream
from CSO discharge point are summarized in Table 4.8 and 4.9. In general, infection
risk using rotavirus model was 1-2 loglo higher than infection risk when echovirus 12
model was applied. At the 25-hn location, mean rotavirus infection risk were 8.7 x
10'2 and 2.2 x 10'1 for low (60 viruses/ml) and high (1322 viruses/ml) CSO virus
concentrations, respectively. For echovirus 12 model, mean infection risk was 8.0 x
10'4 for the low virus concentration and 1.54 x 10'2 for the high virus concentration. If
echovirus 12 dose response was used a reference for risk estimates, infection risks
were lower than the US EPA tolerable recreational risk level (Pi: 1.6 x 102) at both
locations (i.e. 25- and 50-km) when CSO discharge contained 60 viruses/ml, which
means wading activities down stream are not affected by CSO discharge at all when
the CSO virus level is 60 viruses/ml or less.

Overall, point risk estimates based on rotavirus dose response model
(representing highly infectious viruses) gives more conservative results and are more
protective of public health than risk estimates based on the echovirus 12 dose response
model (representing moderately infectious viruses). Rotavirus model gives a longer

duration of elevated risk compared to the echovirus model when the same initial virus

158

 

 

 

concentration was used. Point risk estimates based on rotavirus dose response show

that, swimming and wading activities are not advisable at the beginning of the 12th h
through the 34th h post CSO discharge 25 km downstream from discharge point and

between the 27th h and the 55th h 50 km from discharge point; whereas point risk
estimates based on echovirus 12 dose response gives a shorter duration of elevated

risk, swimming and wading activities are not advisable at the beginning of the 14th h
through the 28th h post CSO discharge 25 km downstream from discharge point and

between the 31St h and the 48th h 50 km from discharge point (Table 4.6 and Table

4.7). In general, duration of elevated risk is dependent on initial virus concentration

and the potency of the virus (dose-response model) used in the risk estimations.

159

 

 

 

.3305; 333 3838283 330 .20 we a: x o.” 3 30 £2038 a 8 833 30 8333 3 8339 a

.3323” 333 0.3330 Em we L: x 3 55 3.23 3: 3083 335 8 Ban 0% 3333 3 8003 a

 

 

 

160

b23038
.1232 03:52 00:30.3. 1:33 .3 3a
afieoa
323.: 72353 N.2052 33-2 N333 1:33 3 3E
£5003:
33.32.: 7230.3 725m "-2363 “3.3.2 0233 N3:53 02053 .3 fix
2 2:33.30
.0333
M333 was: 33:33 “-2me .3 3E
afieoa
33.3 1:33 7235: 33.3-3.3 .333 023: 3 3E
8300.3:
33.3 72050.3 2:3 woes: 32: Texas .3? E .3251 3 3E
“Essen
3 02m 0.0.:— 0.3m 0.5.
an: 333.0 03.. :3: :3: A5 02. .3333 3: :3: r3:
.3 559:5 .25: 0.00.— 502 .3 gem—3:: .232 0.00.— 502 we;

GEFQE? nun: 5605:0050 25> gamma 95000:»? :8 5505:0250 2...? .59—

*3 00m E0>0 .530 0:0 28 29% 2 mo 0.50098 Suzanna a 55 J 08:03.3 93 3 >02 50253 08000. 8150 a no 6003
000030—00 0.53 £3 :39? 000200 95> 00% 08: 8 E090.— xfi: 00230—0 mo 003035 .0509 092006 OmU 0 no 8005338

EM mm wEEEBm .8.“ @5588 was 563.88 50:00.3: Q 2:30:00 98 m§>88 mo amt 32:8 98 526 ofiwfim ed 033.

 

35:50 .833 3850060.. 5005 <mm we «-2 x Qm mo 52.— 352958 a no B23 505 “00:30.0 .50 350.59 n

332% 333 0533. Em 25 L: x 3 55 3.32 3 33 8355 338 3 335 3: 3333 3 .3335 a

 

 

 

161

£333
#533 3:53 “-232 95:53 3 38
53333
“.53 m 75:52 N.5253 3551.: 05:52 M.5253 3 33
8353
«3-8 7205.3 7255 "5:555 3.3-03 ”.2535 3:55 M5:52 3 32
2 3.23235
322.88
05:2; 05:35 3:53 “-533 3 32
33333
33-8 .5353. 05:50.5 33.2 75:2 .0 ”-533. 3 32
5500.35
«3-02. 75:25 75:55 75:53 3.2-2 7530.0 7255. 05:35 3 32
nae-36M—
fim 0.5m 0.5. 33
E 3... .8333 an: 5.35 b3: 5 V.3. 33.33 a: 5.5 5.5
.3 1339::— —a===< u—aoh fine: a: flofluhfln— :25: iv.— flno—Z want!

95302:? Nam: ace-5:00.80 25> aw:— A_E\000=5> c3 act-5:00:00 25> 30.—

.>3 Ba E0>0 530. 0:0 0:0 99% 3 mo 0.30098 825588 0 553 .5 .558030m can 2“ >02 5050: 30000 530 0 no @003
3330—00 803 235 3:59. 00020." 25> 50:0 05“ 3 20.50." 033 302320 mo scrub—D .53 033026 Omo 0 .3 80050552.

:2 53 033.3% 35 £333 05 .5352: 58805 2 2:328 Ba 0:338 3 3.: 3:53 Ba 333 205m 3 033.

 

.6385.“ as; 38228“ new Em we a: x 3 mo 2% £282: a 8 893 2% 8333 .3 855 a

.65qu 533 225% fim 2: L: x 3 as 32.3 “an :88? Essa 8 823 2% 35% Mo 853a a

 

 

 

2358
BE? 0-2on 9263 E33 we fix
$2988
$2-3 _ .2me vote: 52 ”-2sz :22? mo 22
. . ”2380.2:
$.26: 1:23 ”.223 $232 “33.2 ~2wa «.2256 vote? mo V72
2 ӣ55m
gates
23$ v22: v.25; wove? we fig
@2958
933.2 722.3 70:22 £24; 7253 _ ”-233 mo 22
2058.2:
«2-2 vote? 7252 BEN.” “3%: 722m.” 722.2“ «.222 mo 22
“3:33.
:5. in , v.5— in
5 an: 8.3% £5 £3 3 an: 3.5% 35 ES
.3 H8359 3552 23m :32 he 5:959 ...—:54 :3.— 532 25>
£82.35? NNME 5522.359 25> :wE Aiken—...? :3 55233.59 25> 33

 

Sac En Eo>o mafia? one can 99% om Mo 82893 82552: a at? .uonfioaom 38 .32 50253 comma waawaa 5 no 323
33:23 203 3% 3:92 .0322 25> Ban 2:: 8 2&2 fit @3963 .«o 8:825 .253 0333.6 Omo a mo Each—ago“.

=2 R 2:33 5 £388 Ea .bfieos 5:83 .2 3328 Ba mega mo ME 325 2a a?» 23m we 0:3

162

 

.Qauqfim 533 303382 swam aim ma we x 9w mo amt mefioE a no woman Mac 32:63 no conga a

.9383“ has 3% $5 we L: x 3 as 30:0 um: 8:82 gas 8 83; am: 8:320 we 8:85 a

 

 

163

 

£388
923$ 923$ wens? 1:32 L8 in
béeoe
$1.2 vein; 2:53 A22 1:53 wo_xo_.~ we fix
sarcoma
“32.3 72me “bus: 1.3% “33.3 meg? ”base .123? .8 fix
a @2325
£388
«baa? 9233 2:53 1:wa go via
£2988
3-3 239m "-233 $51.3 72:? $32 Mo vim
8803
«293 339m To}: 2:52 “33.3 7253 72x3“ ”.263 go in
negated
:5. “E: in an.
as an: 82.8.9 at £5 £5 5 at 8.5% an: £5 £5
a: 539:5 .252 xaom :32 .3 acuausn 1352 :3.— ..32 25>

Gasman—...? nun G 5595:0950 25> auflu GEES—E.» :8 gush—52.8 25> .39—
$3 Ba 80>» many?» 28 98 mag om mo 0.58%». 835me a fit» .uonaoaom can 32 59.32— 538 mafia 5 no woman

c8333“. 20? 86.: 3:52 .0322 25> bum 0:5 8 flame." amt 33%—o no cougm Asa 09.38% OmU a mo gobmgou

:2 on may“; 8.“ Euros Ea .bfieoa e383 S 3328 Ba 8:58 go an 3&5 Ba “:95 23m 3 03¢

 

Interval risk estimates

Interval risk estimates associated with swimming and wadding in virus-polluted
water were computed using rotavirus (represented highly infectious viruses) and
echovirus 12 (represented moderately infectious viruses) dose response models. For
volume consumed during swimming activity, a lognonnal distribution with a mean of
38 ml and 19 ml were used for non-adult and adult swimmers, respectively (Dufour et al.
2006). Volume consmned during wading activity was set at 10 ml per event. River
discharge, CSO discharge volume and CSO discharge duration values in summer along
with their means, distributions and standard deviations were used in all simulations
unless specified otherwise (refer to Table 4.4). Other parameters and their values used are
listed in Table 4.4. A Monte Carlo analysis consisting of 10,000 trials was performed for

each simulation.

 

Rotavirus infection risk distributions obtained from the Monte Carlo analysis did
not follow a normal distribution and are highly right-skewed (Fig. 4.2), resulting in mean
infection rates that are much higher than the median and 75th percentile risk values in
some cases. However, in order to maintain uniformity and not to underestimate health
risk, the mean rotavirus infection probability for each exposure group was used for
reporting and comparisons. Mean rotavirus infection risks correlate positively to water
ingestion volume and negatively with distance from discharge point. Mean rotavirus
infection risks were higher than the US EPA tolerable risk level for fi'eshwater 25 km

downstream fi'om discharge point for at least 100 h afier virus release with infection risk

peaked between the 18th and the 30th h after virus release (Fig. 4.3). At this location, the

164

peak rotavirus infection risks were the highest for non-adult swimmers (1.73 x 10"),
followed by adult swimmers (1.41 x 10") and waders (1.27 x 10").

For the SO-km location, mean rotavirus infection risks were below the US EPA
tolerable risk level for recreational freshwater for the first 10 h after virus release. At this

location, mean rotavirus infection risk peaked between the 30th and the 48th h after virus

release (Fig. 4.4). Peak rotavirus infection risks for the SO-km location follow a similar
trend as for the 25-km location though mean peak infection levels were lower than those
observed at the 25 -km location. The mean probability of rotavirus infection was the
highest for non-adult swimmers (1.06x10'1), followed by adult swimmers (8.39x10'2) and
waders (7 .47x10'2). This trend can be explained by the consumption rate for each
population group during recreational activities, non-adult swimmers tend to consume
more water than adult swimmers and waders.

The length of duration of elevated risk was positively associated with distance
from discharge point. For the 25 km location, the probability of rotavirus infection
remains elevated (i.e. above US EPA tolerable risk level of 8 illnesses/1000 events) for at
least 96 h after virus discharge (Fig 4.3). Duration of elevated risk 50 km downstream of
the discharge point has a longer tail (slower decline) than duration of elevated risk curve
at the 25-km location, which is in coherent with the virus transport pattern observed in
bacteriophage P22 tracer study (Chapter 3). At this location, rotavirus illness risk level

remains elevated for at least 140 h (Fig 4.4).

165

 

 

 

 

0.80 T

0.60 ..

0.40 -

Relative Probability

0.20 "

 

0.00 r
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Probability of infection

Figure 4.2 Distribution of rotavirus probability of infection by its relative probability
(occurrence), 24-48 h after release, 50 km downstream of discharge point. Median

probability of infection is lower than mean probability of infection.

166

 

 

.0322 we; 5% s cm “$2 am now 3.3 x m we _o>o_ amt $05: 383882 03828 <mm m3 Bot 833038
$3 on o; 92: conwE 33 can "850%: 502 .0328 “can a 50m Ea am“ 5033: Unison 5:82 we Mat :82 demon 03388

8on 885338 Ex 3 .omaaomé Omo 5% fine: £55 83338 383852 8m cocoofi: ESE .«o can :82 m6 0.3mm

 

A5 ease—o.— ..8uu ESP
8 8 on em 2 o

§
8
e
{2
Le

 

    

 

E
O
"0999:“! J0 )lsu mw

_o>£ <mm mD .III
gag. . . . NS
guacamnwfiikm II I. 36
/ \
asegsm Ill .
2 o ,
/ II\ \ , ,

 

167

 

.0322 35> “can a 9: 3.3— 3 you 9-3 x m we _o>o_ Mat 32:: 3533.88 03828 55 m3 88w 333038 we
x o.“ 55 Hog—mm: 83 fit 228095 502 ammo—8 “can a 5%. can 58 3253 c8309 b25389 5:985 502 430m 0328:“

88m 58.5338 Ex om .ombfioflw Omo baa mica find—m3 36158 3:05.202 .8.“ couoomfi £5238 mo amt :82 iv oSmE

 

cc 0322 .33.. «Ema.

 

 

    

 

lxy|r\~l.|\rpl‘}l IIL LL’L!\‘L|L I'r‘f iv 80
..l: .
I I [ Nod
, 3.0
y 86
33‘— <mm mD |
mafia}? - . . / I 86
aegggg | I / / \ .
Ea.m§3ml / I n\ I 2 o
[ N;

Home}!!! 10 my “new

 

168

Compared to point estimate, interval estimate has an added advantage as one can
include uncertainties associated with each parameter in a single simulation and an overall
picture of risk (i.e. distribution, and range) can be obtained. A Monte Carlo simulation
takes into account uncertainties and variability associated with each input parameter
based on their distribution types. When compared to results from point estimate, it was
observed that interval risk estimate based on mean rotavirus infection risk gave an
extended duration of elevated risk. For the 25-km location, duration of elevated risk for
rotavirus infection by the point estimate (based on an initial CSO virus concentration of
1322 viruses/ml) began at 27 h after virus release and lasted to 55 h after virus release,
whereas for the interval estimate, the duration of elevated risk began approximately four
hours after virus release and infection risk was elevated till at least 96 h after virus release
(Table 4.6 & Fig 4.3). In other words, interval risk estimate gave a more conservative
estimation; rotavirus infection risk level was elevated beyond the US EPA tolerable
recreational freshwater level for 28 h and 92 h, according to the point estimate and the

interval estimate, respectively. For the SO-km location, risk level for rotavirus infection
was elevated between the 27th and the 55th h after virus release by point estimate, while it

was elevated beginning at the 10th h and declined below the US EPA risk level at
approximately 140 h based on the interval estimate (Table 4.8 & Fig 4.4).

By defining a large range for the “time after release” parameter, sporadic
“elevated risk event” at time ranges beyond the duration of elevated risk calculated based
on the mean infection risk can be observed and may be included in risk management (Fig
4.5 and Fig 4.6) . For the 25-km location, it was shown that a random swimmer may have

a 30 or 40% chance of getting a rotavirus infection from swimming in the contaminated

169

 

 

 

water approximately 140 hours after the discharge event (Fig 4.5). For the SO-km
location, a sporadic elevated risk event may happen for more than 300 h after the

discharge event (Fig 4.6).

0.9 .-
0.8 '
0.7 -.
0.6 3 '
0.5 .. : ‘
014 , 1
0.3 '

Probability of infection

0.2 ~-

0.1 "A." o .' ... .

 

 

0 --'-r--"r ' r. 7 7 1
0 50 100 150 200 250 300
Time after virus release (h)

Figure 4.5 Probability of rotavirus infection versus time after release (h) 25 km
downstream from discharge point. Elevated risk event could happen after more than
100 h afier virus release. River discharge and CSO variables in summer were used for

this simulation.

170

Probability of infection

 

 

--“‘-t—"-+-‘ ;'; ;‘: :
0 50 100 150 200 250 300 350

Time after virus release (h)
Figure 4.6 Probability of rotavirus infection verses time after release (h) 50 km

downstream from discharge point. Elevated risk event could happen after approximately
300 h after virus release. River discharge and CSO variables in summer were used for

this simulation.

Sensitivity analysis showed that river discharge is the primary contributor to
uncertainty in probability of virus infection for both swimming and wading activities. A
trend between river discharge and probability of rotavirus infection 25 km downstream
from discharge point was found (Fig 4.7). In Fig 4.7, probability of rotavirus infection
25 km downstream of virus discharge point (time after virus release was set between 30-
40 hours after virus release) peaks when river discharge is approximately 35 m3/ s. The

relationship between river discharge and probability of virus infection is dependent upon

171

 

 

distance from discharge point and time after contaminant release. For example, summer
river discharge contributed to approximately 33 % of the single event rotavirus infection
risk and 40 % of annual rotavirus infection risk in adult swimmers 25 km downstream of
virus discharge point between 30—40 h afier virus release. CSO duration contributed to
15% of single event infection risk and 18% of annual infection risk in adult swimmers
under the same distance and time conditions.

In addition, the river discharge also highly influenced the duration of the elevated
risk at each location. As river discharge increases, the duration of elevated risk decreases.
In this case, when river discharge values for the summer (mean: 66.87 m3/s, std dev:
50.31) was substituted with the overall mean river discharge values (mean: 106.64 m3/s,
std dev: 33.50), which has a higher mean and a lower variance compared to river
discharges in the summer in risk simulation, duration of elevated risk at both locations

(i.e. 25 and 50-km downstream) decreased. Mean infection probability peaked between
the 14th and 16th h after release. For the 25-km location, mean rotavirus infection risk was

higher than 1.6 x 10'2 (calculated from US EPA tolerable recreational illness risk level of
8 x 103) for less than 40 h after virus release (Fig 4.8), compared to approximately 96 h
afier virus release when river discharge for summer was used (Fig 4.3). Peak mean
rotavirus infection risks were also higher in this simulation. They were 2.71x10'l,
3.28x10'l and 2.47x10'l for adult swimmers, non-adult swimmers and waders,
respectively, compared to 8.39x10'2 (adult swimmers), 1.06x10’I (non-adult swimmers),
and 7 .47x10’2 (waders) when river discharge values for summer were used. Other

parameters such as seasonal virus concentration in CS0 discharge and CS0 discharge

172

 

 

duration, do not show significant influence on duration of elevated risk and risk

distribution.

0.9

0.8
0.7
0.6
0.5

0.4

0.3

Probability of infection

0.2

0.1

0.0

V
.y.
‘I’.

Zeta:

’3’9’5’.‘
o 0"
.U

‘-

I.-

Q

.z‘.“ 5"”
..‘I 3111?: °
F

'9

 

O

l?

50

i i A: ‘ 1‘ “‘ i ‘ ‘i
100 150 200 250 300 350
River Discharge (m3/s)

Figure 4.7 Scatter plot of probability of rotavirus infection versus river discharge (ms/s)

25 km downstream of virus discharge point between 30 and 40 hr afier virus release. The

probability of infection peaks when river discharge is approximately 35m3/s.

173

 

 

1v

 

  
 
   

 

 

 

0.35 7
a
g 0‘30 1 / \ — Swimming-adult
g 0.25 ~ — - Swinnning—nonadult
.5 ' ' " Wading
“a 0.20 -
— US EPA level

g 0.15 —
'8
A 0.10 .
E
9" 0.05 .

0.m I I l l I l l I

0 5 10 15 20 25 30 35 40
Time afier release (h)

 

 

 

Figure 4.8 Mean risk of rotavirus infection for recreational activities within hours after
CSO discharge, 50 km downstream from discharge point, simulated with the overall river
discharge values (including all seasons, mean: 106.64 m3/s, std dev: 33.50). Mean

infection probability peaked between the 14th and 16th h after release. Peak mean

rotavirus infection risks were 2.71x10", 3.28x10'l and 2.47x10'l for adult swimmers,
non-adult swimmers and waders, respectively. Mean infection risk was higher than 1.6 x
10’2 (calculated from US EPA tolerable recreational illness risk level of 8 x 103) for less

than 40 h after virus release.

The last part of this risk analysis evaluated risk for swimming in the lower Grand
River on days without a CSO event. For this scenario, outcomes from point estimate were
compared to those of interval risk estimates. Only two parameters: ingestion rate during
swimming and virus concentration were used in the analysis. In point estimate, ingestion

rates of 50 ml and 10 ml were used for swimming and wading, respectively. Dose

174

 

response models for rotavirus and echovirus 12 were used for risk comparisons between
highly infectious viruses and moderately infectious viruses. An average virus
concentration of 7.76 viruses /ml was used. The annual risk for swimming was estimated
based on a maximum of 94 days and the annual risk for wading was based on an annual
exposure of 140 days. For the partial Monte Carlo analysis, distributions for virus
concentration in surface water (lognormal; mean: 7 .76 viruses/ml, standard deviation:
19.8) and maximum exposure days for both swimming (min:1, mode212, max:94) and
wading activities (triangular distribution: min:1, mode:27, max: 140) were added. In the
full Monte Carlo analysis, distributions of water ingestion rate during swimming for
adults (mean: 19 ml, standard deviation: 19 ml) and non-adults (mean: 38 ml; standard
deviation: 31 ml) were included in the simulation. All parameters used are listed in Table
4.5.

In point estimate, a single swimming event risk of 8.4x10" and 3.4x10'l was
computed for rotavirus and echovirus 12, respectively (Table 4.10). For a single wading
event, infection risk were slightly lower, 7.5 x10'1 for rotavirus and 1.2x10‘l for
echovirus 12. These risk values shows that even during normal days, risk of virus
infection is elevated in the lower Grand River. However, it must be noted that the use of
the average virus concentration in surface water for this risk calculation most likely
caused overestimations because only 6 out of 20 surface water samples analyzed were
positive for adenovirus DNA (non-detects were substituted with detection limit*0.50). In
addition, the assumption that one virus DNA copy represents one viable virus may
contribute to risk overestimation as well. Moreover, other factors such as virus

attenuation in the water were not included in the model.

175

 

 

 

Results from partial Monte Carlo and full Monte Carlo analyses were less H
conservative (have lower mean risk values) than results from point estimate, _
demonstrating the influence of uncertainties associated with virus concentrations in
surface water and ingestion volume. However, with partial and full Monte Carlo
analyses, distribution and range of infections could be obtained. Partial Monte Carlo
analysis gives a mean single event risk of infection from swimming of 7.7 x 10'1 (min:

3.7 x 10']; max: 9.3 x 10“) and 2.3 x 10'1 (min: <l.0x10‘3; max: 7.9x10") for rotavirus
and echovirus 12, respectively (Table 4.10). In partial Monte Carlo estimation, single
event risks for wading are lower than point estimates for both viruses as well, 6.5 x 10'l
(min: 5.0 x 104; max: 9.2 x 10") for rotavirus and 9.0 x 10‘2 (min: <1.0x10'3; max:
7.2x10") for echovirus 12. Annual infection risks for rotavirus were higher than 9.9x10'l
for all scenarios (both swimming and wading activities) while annual infection risks for
echovirus 12 was 9.1 x 10" for swimming and 7.8 x 10'l for wading.

Full Monte Carlo analysis was conducted to evaluate the influence of uncertainty 1
in ingestion volume, in addition to uncertainty associated with virus concentrations in
surface water. Ingestion volume data for adult and non-adult swimmers were extracted
from Dufour et al. 2006. Full Monte Carlo analysis gave a slightly lower risk estimates
than those fi‘om partial Monte Carlo, showing minimal influence of ingestion volume on
virus infection risk. The overall rotavirus and echovirus 12 infection risk from swimming
are 7.2x10'l (adult: 6.7 x10'1; non-adult: 7.3x10'1) and 1.7x10'1(adult: 1.2 x 10";
non-adult: 1.8x10'1), respectively (Table 4.10). Full Monte Carlo analysis was not
performed for wading activity because a distribution for ingestion rate was not available.

Annual infection risks for rotavirus were higher than 9.9x10'l for all scenarios (both

176

 

swimming and wading activities) while annual infection risks for echovirus 12 ranged
between 7.2 x 10’1 and 8.3 x 10". In all analyses, single event infection risks estimated
using both rotavirus and echovirus 12 dose response models were higher than US EPA
tolerable risk level for recreational freshwater. This suggested that the lower Grand River
is contaminated by enteric viruses at levels where by the risk of contracting viral-induced
acute gastrointestinal illness from recreating in the river without a CSO event is higher
than 8x103.

With the use of Monte Carlo analysis, difference in risk distribution for both
viruses can be observed. In interval risk estimation, distribution of risk is important in
determining a reference risk value to be utilized. Distributions of both rotavirus and
echovirus 12 infection risk were fitted using Risk Solver®. In the full Monte Carlo
analysis, probability of rotavirus infection is heavily left-skewed (i.e. has relatively few
low values) and fits into a MinExtreme distribution. For example, the probability of
rotavirus infection for adult swimmers has a mean (6.7x10") that is lower than its median
(7.0x10'1) and mode (7.5x10'1) (Fig 4.9). Probability of echovirus 12 infection, on the
other hand, is heavily right-skewed and fits an exponential distribution. The probability
of echovirus 12 infection for adult swimmers has a mean (1 .2x10") that is higher than its

median (6.8x10'2) and mode (6.0x10'3) (Fig. 4.10). So, depending on the desired level of
protection, mean, median, 95th percentile or other risk percentile fi'om a risk distribution

may be used in risk reporting and evaluation.

177

 

 

00.5808 :0 00.0 L: x ad A 5300MB 03388 .8.“ Mat :3ch ....I.

.608 ..m e 0505 :80

903 Edam 00m 0823 noum0wfi was 2023005800 033 53 Ho.“ 8:35me induceme— wEms 0805980 00.3 230 0:52 :3 ......

.82.. 3280 wq30§60 How 000: 0003 canvas“ .NHH0coE JHEEV mafia—£30 was 8:"wa .bmnbwofi JHEEV mica?
Ho.“ 05898 no 930 Sm 2853506 Esmswth Awd— “coarse 05:80 .3535 21.5 ”50.5 .0003 0009.50 E seven—50:00 25>
mo aorBEflc _mnuoewE 98 836%: cm 033:8 2 no 0823 53009: 0m§0>m 5 wfim: 00005980 003 2000 2:22 Bub:— ...

 

.030

 

7205.0 1:05. V .03; €282 7203.0 .030 1:065 2:382
.023 725; 1:05. V 72x2 033‘ Tex? 7033 703.0 :35.
Li; 720.3 .025. V 72x: =§>o0§m 72x2 N2.5.0 Texas =§>o0£se5m
3.30 00.32 :5— :230 0:52 :5—
3:st 72x? .-ocsV Nbead was; 78de N.2056 72x3 0523
72*; Texas n0:5. V .030 mafiafim 7ng we”; .25. 0:05:50.
2.30 3.82 350.5 02.30 8.32 Eta.—
.-Sxfi 0533 72x2 0535
72.5.0 .03; mgfim 7203.0 05585.0.
3.25% .58 38.08 .5...—
ttexmt an: flaw.—
..552 :2 a: :82 S 25350 :2 £2 532 “E353.

 

 

.032 @920 0032 2: E 00003 000M050 000330.380 E N: 0:030:00 can 2:388 Sm 2000

2:02 :3 05 230 00:02 Evan 600850 “Eon mama: 02:58 3:2002000 08 28:00.? mo amt 508 we 50.5an0 36 033.

178

Relative probability
.0 o .o P
S: S G '5‘

_o
o
o

 

0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
Probability of infection

Figure 4.9 Relative probability of risk level versus probability of rotavirus infection for
adult swimmers in the full Monte Carlo analysis shows distribution of rotavirus infections

in surface water without a CSO event. Mean: 6.7x10", median: 7.0x10", mode: 7.5x10".

179

 

0.25

0.20

0.15

0.10

Relative probability

0.05

0.00
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90
Probability of infection

Figure 4.10 Relative probability of risk level versus probability of echovirus 12 infection
for adult swimmers in the full Monte Carlo analysis shows distribution of echovirus 12
infections in surface water without a CSO event. Mean: 1.2x10", median: 6.8x10'2,

mode: 6.0x10'3.

Discussion

Recently, quantitative microbial risk assessment models have been used to
evaluate and describe risk associated with exposure to pathogens in different
environmental matrices (Donovan et al. 2008; Eisenberg et al. 2008; Rose et al. 1991;
Ryu and Abbaszadegan 2008; Soller et al. 2003; van Heerden et al. 2005). This risk
analysis sought to fill gaps in previous risk analyses associated with virus contaminated
water. This is the first risk analysis that incorporated a virus transport model to estimate
virus attenuation downstream fiom discharge point. In addition, while risk analysis have

been performed to estimate health risk of CSO discharge, this is the first study that

180

 

measured the concentration of a human virus pathogen in CSO discharges and applied the
concentration in a risk assessment.

As with any microbial risk assessment model, this risk assessment comes with
several limitations and contains assmnptions on pathogen, host behavior and efficiency of
pathogen detection assays. The most important limitation of this risk assessment is
limited field data. In this assessment, average virus concentrations in CSO events were
computed from only six samples collected in winter (one event) and spring (five events)
2008. As reported by other virus monitoring studies, virus concentrations may vary from
season to season, affected by differences in temperatures, host discharge rates, and
microbial composition of water (Krikelis et al. 1985; Skraber et al. 2004; Tani et al.
1995). In order to better characterize seasonal trend and distribution of virus
concentrations for more accurate estimation of health risk, long-term monitoring of virus
concentration in C808 is warranted.

The high background risk estimated by this study (>6.7x10'1 for rotavirus and
>1 .2x10'l for echovirus 12) based on average virus concentration in surface water is
probably due to the high detection limit and insufficient field data (n = 20). The virus
detection method used in this study has a quantification limit of 10 viruses/L, an
improved virus detection method with a lower detection limit is essential for getting a
more accurate picture of virus concentrations in impacted and contaminated surface
water. In addition, because of the limited number of samples collected, a good
distribution of virus concentration cannot be determined. A more frequent and rigorous
monitoring of enteric virus concentrations in the lower Grand River is necessary for a

more precise estimation as well as identification of pathogen hot spots (e. g. illicit

181

 

contaminant discharge point or transient storage area) along the river. Furthermore,
because of a lack of knowledge on enteric virus-associated illness prevalence in the
population, the attributable risk by water-related recreational activities in the area cannot
be determined and compared with results from this assessment.

This risk assessment was built on several assumptions that may cause
underestimations or overestimations of risks. Assumptions regarding attenuation of viral
pathogen in surface water, viability of recovered viruses, infectivity of pathogens, and
ingestion volume were used. Because it is impossible to inoculate river water with
pathogenic microorganisms, the transport behavior and attenuation of a surrogate virus,
bacteriophage P22, in surface water was modeled and used in this risk assessment. In this
study, it was assumed that human viral pathogens (in our case, adenoviruses) are as
persistent as bacteriophage P22 while the actual survival of these viruses in river water
has not been compared.

Secondly, in order to generate a worst case scenario, it was assumed that all
viruses isolated from environmental samples were infectious and that one DNA copy in
PCR represents one virus. In reality, this may not be true. There is a possibility of
quantification of partially degraded DNA from environmental samples as DNAs are
known to outlast virus viability in the environment (Masago et al. 2008; Wetz et al.
2004)

In addition, assumptions regarding immunity or sensitivity of hosts that may
cause underestimations of risk in certain population groups were used. For example,
because of a general lack of dose response data, a single dose response model was used

for both adult and children though children may be more vulnerable to these pathogens.

182

 

 

The recreational risk of other sensitive populations, such as pregnant women, elderly and
immuno-comprornised patients are not specified in this assessment.

This assessment is based on monitoring data for adenoviruses, although they are
thought to be more prevalence and persistence than other waterborne enteric viruses, the
cumulative risk from more than 100 types of enteric viruses, parasites and bacteria
pathogens, which are regularly present in waste water is hard to determine. Nevertheless,
this risk assessment provides a feasible framework to estimate risk from contacting with
an important group of pathogens that were released into the river during a sewage spill or
CSO event.

Overall, this risk assessment could be improved by a detailed study of major
uncertain parameters in the model: 1) a long-term and more rigorous assessment of
enteric virus concentration in the river during dry season and in CS0 discharge; 2)
comparison of viable virus counts to qPCR virus concentration; 3) survival of enteric
virus versus bacteriophage P22; 4) survey of virus transport and dispersion at Lake
Michigan beaches (plume studies).

Compared to other risk analysis on recreational water, this risk assessment
estimated higher background virus infection rates, possibly due to the high background
virus concentrations detected in the lower Grand River and some of the assumptions
mentioned previously. In a risk assessment of adenoviruses in recreational water in South
Africa, daily adenovirus infection risks of 1.71x104 and 3.12x10'5 were calculated for
river water and darn water, respectively (van Heerden et al. 2005). The low infection
risks were linked to the low adenovirus concentrations in the studied river water and darn

water, virus levels were between 103 and 104 times lower than those found in this study.

183

 

 

Risk analyses related to recreational water have also been performed applying
different methodologies. In an epidemiological study in which beachgoers were
interviewed for health symptoms after recreating at four Great Lakes beaches polluted by
sewage, Wade et al. 2008 found that gastrointestinal illness rate is directly correlated to
daily average Enterococcus exposure. They observed a 3.4 % increase in gastrointestinal
illness incidence in swimmers who were exposed to 100 Enterococcus compared to non-
swirnmers. During the 10 to 12 days follow-up period in the study, the incidence of new
GI illness among swimmer and non-swimmers were 8.3% and 6.0%, respectively (Wade
et al. 2008). Soller et al. (2006) took a different approach in their risk analysis, they built
a water quality model based on fecal indicator data in the Newport Bay area for
estimating site-specific concentrations of male specific coliphage (used as a surrogate for
human enteric viruses). A much lower gastrointestinal illness rate (0.09 %) from
recreating in marine beaches in South California was simulated based on a disease
transmission model. This phenomenon is possibly because exceedances of the water
quality standard based on bacteria indicators levels most commonly occur during the time
of the year and in areas where recreational usage is low.

Until recently, few microbial risk assessments has been performed on health
effects of recreation in contaminated freshwater, especially in the afiermath of a
contamination event, such as a CSO event. Donovan et al. (2008) found high bacteria
indicators and Giardia concentrations in two water samples collected within the vicinity
of a CSO outlet in the Lower Passaic River but detailed analysis on other pathogens was

not available. Based on their CSO data, Donovan et al. (2008) estimated annualized

184

 

infection risks of 0.88 for both Enterococcus and fecal Streptococcus, and >0.99 for
Giardia.

Conversely, this risk analysis provides a comprehensive view of the health risk
associated with recreation in surface water affected by CSO events; with sufficient
monitoring and transport data, health effects associated with other pathogens can be
evaluated. Different scenarios can be developed to demonstrate benefits of management
choices, such as improved wastewater treatment, diversion of contaminated flow, public
education or beach advisory/closure. With this model, it was suggested that duration of
elevated risk and risk levels at a site are influenced by river discharge, CSO volume and
CS0 duration. Thus, instead of repeated measurement of microbial indicator levels,
which is laborious and time consuming, a predictive system based on CSO discharge
information and river discharge can be used by environmental managers for informed
regulatory decision making such as duration of beach advisory or beach closure. In
addition, this risk assessment fi'arnework may be used to evaluate CSO effects at other
sites, provided that sufficient site-specific data (i.e. background bacteria indicators or
virus concentrations, hydrogeology data, CSO discharge data) are available.

In short, this risk analysis provides valuable insights that can be applied in both
regulatory and operational contexts. Environmental managers can refer to this analysis
for regulatory decision-making, such as the posting of beach advisory or deciding
duration of a beach closure. Wastewater operators may use this framework to examine

the benefits of making operational changes or adding additional waste water treatments.

185

Conclusion

This risk assessment combined an analytical hydrological model with
environmental data in a QMRA framework using probability of infection models. The
result of this risk analysis suggests that contact with water in the lower Grand River after
a CSO event poses significant health risk. However, more monitoring data are needed for
evaluating seasonal risk from CSO discharges. Interval risk estimation gives a longer
duration of elevated risk than point risk estimate. However, careful examination of
interval risk estimation data, especially risk distribution, is important when interpreting
risk. In this study, stream flow was a good predictor for duration of elevated risk. Overall,
this analysis suggests that surface water quality in Grand Rapids can be highly impacted
by CSO discharges and the community attention on preventing and controlling CSOs is

warranted.

186

References

Bennett, J. V., S. D. Homberg, M. F. Rogers and S. L. Solomon (1987). "Infectious and
parasitic diseases." Am J Prev Med 55: 102-114.

Couch, R. B., T. R. Cate, P. J. Gerone, W. F. Fleet, D. J. Lang, W. R. Griffith, V. Knight
(1966). "Production of illness with a small-particle aerosol of adenovirus type 4."
floteriol. Rev. 30: 517-528.

Donovan, E., K. Unice, J. D. Roberts, M. Harris and B. Finley (2008). "Risk of
gastrointestinal disease associated with exposure to pathogens in the water of the
lower Passaic river." Appl. Environ. Microbiol. 74(4): 994-1003.

Dufour, A. P., Otis Evans, Thomas D. Behymer and R. Cantu (2006). "Water ingestion
during swimming activities in a pool: A pilot study." J Water Health 4(4): 425-
430.

Eisenberg, J. N. S., Kelly Moore, Jeffery A. Soller, Don Eisenberg and J. John M.
Colford (2008). "Microbial risk assessment framework for exposure to amended
sludge projects." Environ Health Perspect 116(6): 727-733.

Gray, G. C., T. McCarthy, M. G. Lebeck, D. P. Schnurr, K. L. Russell, A. E. Kajon, M.
L. Landry, D. S. Leland, G. A. Storch, C. C. Ginocchio, C. C. Robinson, G. J.
Demmler, M. A. Saubolle, S. C. Kehl, R. Selvarangan, M. B. Miller, J. D.
Chappell, D. M. Zerr, D. L. Kiska, D. C. Halstead, A. W. Capuano, S. F.
Setterquist, M. L. Chorazy, J. D. Dawson and D. D. Erdman (2007). "Genotype
Prevalence and risk factors for severe clinical adenovirus infection, United States
2004-2006." Clin. Infect. Dis. 45(9): 1120-1131.

Haas, C. N., J. B. Rose, C. P. Gerba and R. Regli (1993). "Risk assessment of viruses in
drinking water." Risk Anal. 13: 545-552.

Katz, M. and S. A. Plotkin (1967). "Minimal infective dose of attenuated poliovirus for
man." Am J Public Health Nation Heglm 57(10): 1837-1840.

Krikelis, V., N. Spyrou, P. Markoulatos and C. Serie (1985). "Seasonal distribution of
enteroviruses and adenoviruses in domestic sewage." Can. J. Microbiol. 31(1):
24-25.

187

Lepow, M. L., R. J. Warren, V. G. Ingram, S. C. Daugherty and F. C. Robbins (1962).
"Sabin type I (LSc2ab) oral poliomyelitis vaccine. Effect of dose upon response
of newborn infants." Am J Dis. Child. 104: 67-71.

Masago, Y., Tomoyuki Shibata and J. B. Rose (2008). "Bacteriophage P22 and
Staphylococcus aureus attenuation on nonporous fomites as determined by plate
assay and quantitative PCR." Appl. Environ. Microbiol. 74(18): 5838-5840

MDEQ (2008). Combined Sewer Overflow (CSO) & Sanitary Sewer Overflow (SSO)
2007 Annual Report (January 1, 2007 — December 31, 2007).
http://www.michigan.gov/documents/deq/deq-wb-csossoreport07_243725_7.pdf

Minor, T. E., C. T. Allen, A. A. Tsiatis, D. B. Nelson and D. J. D'Alessio (1981).
"Human infective dose determination for oral poliovirus type 1 vaccine in
infants." J Clin. Microbiol. 13: 388-389.

O'Loughlin, E. M. and K. H. Bowmer (1975). "Dilution and decay of aquatic herbicides
in flowing channels." J Hydr0126: 217-235.

Pacini, D. L., A. M. Collier and F. M. Henderson (1987). "Adenovirus infections and
respiratory illnesses in children in group day care." J Infect. Dis. 156: 920-926.

 

Papapetropoulou, M. and A. C. Vantarakis (1998). "Detection of adenovirus outbreak at a
municipal swimming pool by nested PCR amplification." J Infect. 36(1): 101-103.

Rose, J ., C. Haas and S. Regli (1991). "Risk assessment and control of waterborne
giardiasis." Am J Public Hlth Assoc. 81(6): 709-713.

Ryu, H. and M. Abbaszadegan (2008). "Long-term study of Cryptosporidium and Giardia
occurrence and quantitative microbial risk assessment in surface waters of
Arizona in the USA." J Water Health 6(2): 263-273.

Schiff, G. M., G. M. Stefanovic, B. Young and J. K. Pennekarnp (1984). Minimum
hmnan infectious dose of enteric viruses (echovirus-12) in drinking water. Enteric
viruses in water. J. L. Melnick. Basel, Karger. 15: 222-228.

Skraber, S., B. Gassilloud, L. Schwartzbrod and C. Gantzer (2004). "Survival of
infectious Poliovirus-l in river water compared to the persistence of somatic

188

 

coliphages, thermotolerant coliforms and Poliovirus-l genome." Water Res.
38(12): 2927.

Soller, J. A., A. W. Olivieri, J. Crook, R. C. Cooper, G. Tchobanoglous, R. T. Parkin, R.
C. Spear and J. N. S. Eisenberg (2003). "Risk-based approach to evaluate the
public health benefit of additional wastewater treatment." Environ Sci T echnol
37(9): 1882-1891.

Suptel, E. A. (1963). "Pathogenesis of experimental coxsackie virus infection." Acta
virol. 7: 61-66.

Tani, N., Y. Dohi, N. Jurumatani, K. Yonemasu. (1995). "Seasonal distribution of
adenoviruses, enteroviruses and reoviruses in urban river water." Microbiol.
Irnmunol. 39: 577-580.

Turner, M., G. R. Istre, H. Beauchamp, M. Baum and S. Arnold. (1987). "Community
outbreak of adenovirus type 7a infections associated with a swimming pool."
South Med J 80(6): 712-715.

van Heerden, J ., M. M. Ehlers, J. C. Vivier and W. O. K. Grabow (2005). "Risk
assessment of adenoviruses detected in treated drinking water and recreational
water." J Appl. Microbiol. 99(4): 926-933.

Ward, R. L., D. L. Bemstein, E. C. Young, J. R. Sherwood, D. R. Knowlton and G. M.
Schiff (1986). "Human rotavirus studies in volunteers: determination of infectious
dose and serological response to infection." J Infect. Dis. 154(5): 871-877.

 

Wetz, J. J ., E. K. Lipp, D. W. Griffin, J. Lukasik, D. Wait, M. D. Sobsey, T. M. Scott and
J. B. Rose (2004). "Presence, infectivity, and stability of enteric viruses in
seawater: relationship to marine water quality in the Florida Keys." Mar. Poll.

Bull. 48(7-8): 698-704.

189

 

CHAPTER 5.

SUMMARY AND CONCLUSIONS

Enteric viruses are important pathogens in surface water and ground water of the
Great Lakes region. Enteric viruses, such as adenoviruses and enteroviruses, are capable
of causing diseases affecting multiple organs in infected individuals and are major causes
of water-transmitted gastrointestinal disease in swimmers (Bosch 1998). In the current
study, it was hypothesized that human enteric viruses are present in high concentrations
in sewage and rivers receiving sewage effluents in Michigan and waste water treatment is
not efficient in removing these viruses. Real time PCR was used for adenoviruses
detection and quantification in raw sewage, treated waste water and river water.
Adenovirus was chosen as an index virus because of its high prevalence and persistence
in aquatic environments.

Monitoring results showed that adenovirus consistently present in waters in
Michigan and waste water treatment in East Lansing only physically removed less than
two logic of adenoviruses. The actual inactivation of adenoviruses could not be assessed
because of limitation of detection method used. Average adenovirus concentrations were
1.15 x 106 viruses /L in raw sewage, 1.12 x 106 viruses/L in primary treated effluent, 1.05
x 103 and 2.0 x 104 viruses/L in secondary treated effluent and 8.3 x 104 viruses/ L in
tertiary treated effluent. Enteric adenovirus type 41 (60%) was predominant in waste
water tested. Other adenovirus types detected were Adenovirus type 12 (29%), type 40
(3%), type 2 (3%) and type 3 (3%). Human adenovirus concentrations were above real
time PCR detection limit in 6/20 river water samples, with concentrations ranging

between 8x101 and 6.6 x 104 viruses/L (mean: 7.76 x 103 viruses/L). Six combined sewer

190

 

overflow (CSO) samples were collected from the Market Street Retention Basin in Grand
Rapids in 2008, adenovirus concentration in these samples ranged between 6 x 104 and
1.3 x 10‘5 viruses/L (mean: 5.4 x 105 viruses/L). Analysis of variance (ANOVA) shows
that virus concentration in CSO samples (average: 5.35 x105 viruses /L) were not
significantly different from virus concentration in raw sewage and primary-treated
effluent (p-value: 0.39). The high adenovirus concentration in CSO samples suggested
that the discharge of CSO into the Grand River after heavy precipitation events could
lead to adverse public health risk.

The main objective of this research was to quantify human health risk associated
with enteric viruses fiom recreating in sewage contaminated surface water afier a C80
event. In the current study, it was hypothesized that virus transport and attenuation in a
river can be assessed by running a tracer study using a bacteriophage (i.e. bacteriophage
P22) as a surrogate for waterbome viruses. Bacteriophage P22 morphologically
resembles adenoviruses and unlike most soluble chemical tracer, it presents as suspended
colloids and more adequately describes human virus behavior in water. As a part of the
exposure assessment and to model virus transport and inactivation afler CSO events, a
tracer study using bacteriophage P22 was conducted on a 40-krn reach of the lower Grand
River in Grand Rapids. The transport and attenuation of bacteriophage P22 in the tracer
study could be explained by sunlight inactivation, adsorption to suspended particles,
transient storage and land use. It was thus determined to be a suitable tracer for this
complex surface water system. From the tracer study, it was estimated that bacteriophage
P22 has an inactivation range between 0.27 and 0.57 per day. An analytical virus

transport model for estimating virus arrival times and concentrations after release was

191

then developed based on site specific information (i.e. river discharge, CSO discharge
volume, and CSO discharge duration) and virus attenuation rate from the tracer study.

Using adenovirus as an index virus to represent the likelihood of virus survival
through waste water treatments and virus concentration in CSO and surface water, it was
hypothesized that the risks for enteric virus exposure at Great Lakes beaches are elevated
to above US EPA acceptable risk level for recreational freshwater (8 illnesses per 1000
swimmers) during the swimming season, especially after a CSO event. With the field
survey data and virus transport model, a quantitative microbial risk assessment was
conducted to evaluate the probability of virus infection or gastrointestinal disease
resulting from incidental ingestion of contaminated recreational water at Lake Michigan
beaches that receive input from the lower Grand River.

The point risk assessment modeled with the lowest CSO virus concentration and
CSO discharge volume estimated that the probability of rotavirus infection and
gastrointestinal illness were 4.8x10'l and 2.4x10", respectively. The risk assessment
found that with the levels of virus discharged, contact with the water after a CSO event
poses significant human health risk. The high risk associated with recreating in water
receiving CSO discharge determined from this study emphasizes the importance of
proper wastewater treatment and retention. For all scenarios, estimated risks were higher
than the acceptable level set by US. EPA for recreational freshwater.

Infection risks related to swimming in beaches of the lower Grand River without a
CSO event were assessed using rotavirus (representing highly infectious viruses) and
echovirus 12 (representing moderately infectious viruses) dose response models. Risk of

rotavirus infection related to swimming in beaches in the lower Grand River without any

192

CSO event ranged between 8.0x10'2 and 9.5x10’l (mean: 7.2x10'1). The echovirus 12
dose response, on the other hand, gave infection risks range between 0.00 and 8.5x10’l
(mean: 1.7x10"). Regardless of dose response model use, estimated risks were again
higher than US EPA’s acceptable risk level for recreational freshwater.

Monte Carlo simulations were used to characterize uncertainties associated with
different discharge scenarios (i.e. seasonal stream flow, CSO discharge volume, virus
concentration, distance from discharge point etc). River discharge and CS0 duration
were identified as two main factors influencing risk levels and duration of elevated risk.
Summer river discharge contributed to approximately 31% of single event infection risk
and 38% of annual risk of infection in adult swimmers. CSO duration contribute to 15%
of single event infection risk and 18% of annual infection risk in adult swimmers. Effect
of a CSO discharge may last for several days depending on river discharge and distance
from CSO discharge point, i.e. up to 140 hours and 300 hours at 25 and 50 km from CSO
discharge point, respectively.

In conclusion, concentrations of enteric viruses in the lower Grand River make it
unfit for bodily contact recreational activities, especially after a CSO event. Analytical
virus transport model used in the current study served as a good starting point for
estimating virus survival, transport and associated health risk at a specific location as a
result of a CSO discharge or sewage spill. Availability of CS0 discharge information and
site specific data (i.e. river discharge, area of channel, and slope of river) make it possible
for a more accurate estimation of duration of elevated risk at a recreational site and

inform improved management options.

193