 

V .

whpvnwun‘:£lm.=~.

1. . Citrw .

in; A». an $6..
3.2:. a.

Q

‘-
. .
Lam:
A~QA
.K-"i- u‘.

JM} 1.: ‘1.

mime

I. V
5:21....
p 7.4.”
., . A

. o.

lt.!l..$ i»: :315

WEN:
3

 

. a.“ 5?.

Wm?

1...an
.

3

. 5.. ..
fgﬁ.

. raw. . ..
4.. «:3sz 1.)”...
. 1: R. n .
1w”. .3 «awﬁwmﬁﬁ.

«w a

{1: f»...
,
.9 :

w-‘
\

”51
..\

 

meals

‘3'
20C l

 

l LIBRARV
Michigan State
University

 

 

 

This is to certify that the
thesis entitled

The Descriptive Epidemiology
of Norwalk-Like Viruses

presented by

Diane Gorch

has been accepted towards fulﬁllment
of the requirements for

 

 

Master's degree in Epidemiology

W, ﬂ 624; mp

Major professor

 

Date November 15, 2000

0-7639 MS U is an Afﬁrmative Action/Equal Opportunity Institution

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINE return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11m c/CIRODaieDuepss-pu

 

THE DESCRIPTIVE EPIDEMIOLOGY OF NORWALK—LIKE VIRUSES
By

Diane H. Gorch

A THESIS

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

MASTER OF SCIENCE
Department of Epidemiology

2000

ABSTRACT
THE DESCRIPTIVE EPIDEMIOLOGY OF NORWALK-LIKE VIRUSES
By

Diane H. Gorch

Norwalk-like viruses are responsible for causing millions of cases Of gastro-
enteritis each year. They are ubiquitous and are highly infectious. Norwalk-like viruses
cannot be cultivated outside of man, and prior to the development of novel molecular
detection methods, the virus was difﬁcult tO detect in specimens. Norwalk-like viruses are
commonly spread by several modes of transmission. The consumption of raw molluscan
shellﬁsh is a common mode of infection. Infected food handlers are responsible for
transmitting the infection to large numbers of diners in food service establishments.
Sewage-polluted water supplies have also been responsible for causing large-scale
outbreaks. Most recently, hard epidemiological evidence for airborne spread via aerosols
has been established. Recommendations for prevention include better monitoring Of
shellﬁsh beds, using a viral indicator; improved food handler education regarding personal
hygiene and the prohibition of bare-hand contact of ready to eat foods; more strict
monitoring of well construction and sewage disposal systems, and meticulous
environmental decontamination of spaces where potential cases have vomited. Improved

surveillance will enhance our knowledge of the epidemiology of these viruses.

To my husband Joel and my daughters Chrissie and Jackie, whom I love.

iii

ACKNOWLEDGMENTS

Without my ﬁiends and family, the writing of this review would have been much
more difﬁcult. During my work at Ingham Country Health Department, I have had the
Opportunity to work with many capable, energetic and dedicated public health
professionals. While investigating an outbreak of foodborne illness which eventually was
proven to be caused by Norwalk-like viruses, I decided to research further into the
epidemiology of these viruses to answer the many questions which came to mind in the
course of our work. The administration of Ingham County Health Department offered me
unﬂagging moral and material support during my studies. I would like to thank F. Robert
Godbold, Terry Anderson, and Bruce Bragg for their support and encouragement. I
would also like to thank Wendy Little and Lora McAdams for their word process
assistance which saved me countless hours Of ﬁ'ustration. And thanks to my mentor and
ﬁiend Walter Mack, for whom 20 years of exhortation has at last paid Off. I would also
like to thank my family, who have patiently endured and encouraged me during the
researching and writing of papers, studying for midterrns, and the other pressures which

academic study can exert on an active family’s life.

iv

TABLE OF CONTENTS

LIST OF TABLES ......................................................................................................... vii

LIST OF

FIGURES ..................................................................................................................... viii

LIST OF ABBREVIATIONS ......................................................................................... ix

INTRODUCTION ........................................................................................................... 1

CHAPTER 1

DESCRIPTION OF NORWALK-LIKE VIRUSES .......................................................... 2

CHAPTER 2

CLINICAL FEATURES OF NLV INFECTION .............................................................. 3
Description of Symptoms ...................................................................................... 3
Histopathological Features of NLV Infection ........................................................ 3
Immune Response to NLV Infection ..................................................................... 6

CHAPTER 3

LABORATORY METHODS FOR IDENTIFICATION OF NLVs ................................ ll
Immune Electron Microscopy (IEM) .................................................................. 13
Radioirnmunoassay (RIA) and Biotin-Avidin Immunoassay ................................ 14
Enzyme-Linked Immunosorbent Assay ............................................................... 16
Reverse Transcriptase Polymerase Chain Reaction .............................................. 18
Testing Food Specimens ..................................................................................... 20
RT_PCR and Phylogenetic Analysis .................................................................... 22

CHAPTER 4

SEROPREVALENCE ................................................................................................... 28

CHAPTER 5

MODES OF TRANSMISSION ..................................................................................... 39
Contaminated shellﬁsh ........................................................................................ 4O
Infected Food Handlers ....................................................................................... 51
Waterbome Transmission ................................................................................... 59
Person-to-Person Transmission ........................................................................... 66

CHAPTER 6

SURVEILLANCE ......................................................................................................... 75

CHAPTER 7

BURDEN OF DISEASE ................................................................................................ 79

CHAPTER 8

SIGNIFICANCE OF NLVs ........................................................................................... 81
CHAPTER 9

RISK FACTORS ........................................................................................................... 86
CHAPTER 10

SUMMARY .................................................................................................................. 90
BIBLIOGRAPHY ......................................................................................................... 92
INDEX ........................................................................................................................ 108

LIST OF TABLES

Table] - Instructions for Collecting Specimens to Examine

for Agents of Viral Gastroenteritis ................................................................... 12
Table 2 - Age Distribution of Healthy Adults Tested Worldwide

And Percent Testing Positive for Anti-NV IgG Antibodies ............................... 30
Table 3 - Summary Of Foodborne Outbreaks of NLV Gastroenteritis .............................. 52
Table 4 - Summary of Waterbome Outbreaks of NLV Gastroenteritis ............................ 60
Table 5 - Summary of NLV Outbreaks Reporting Secondary (2°) Spread ...................... 71

vii

LIST OF FIGURES

Figure 1 - Dendograrn of phylogenetic relationship between HWA, Hawaii Agent
and 30 other strains of NLV. The length of the abcissa to the connecting
node is proportional to the genetic distance between sequences. DSV,
Desert Shield Virus; SOV, Southampton Virus; BRV, Bristol Virus; NV,
Norwalk Virus; SMA, Snow Mountain Agent. Line indicates 10%

variation in nucleotideidentity. From Ando et a] (4) ....................................... 23
Figure 2 - Seroprevalence of NLVs ............................................................................... 32
Figure 3 - Seroprevalence in Children: Norwalk Virus ................................................... 34

viii

LIST OF ABBREVIATIONS

CDC ........................................... Centers for Disease Control
ELISA ................................. Enzyme Linked Immunosorbent Assay
EU ..................................................... European Union
G1, 611 .................................................. Genogroup I , H
HAV ................................................... Hepatitis A Virus
HACCP ................................. Hazard Analysis Critical Control Point
IEM .......................................... Immune Electron Microscopy
NLV ................................................ Norwalk—like Viruses
NS SP ................................... National Shellﬁsh Sanitation Program
NV ...................................................... Norwalk Virus
PCR ............................................ Polymerase Chain Reaction
PFU .............................................. Plaque Forming Unit (3)
RIA .................................................. Radioimmunoassay
rNV VLP ........................ Recombinant Norwalk Virus Virus-like Particles
RT-PCR ....................... Reverse Transcription Polymerase Chain Reaction
TD ................................................... Travelers’ Diarrhea

ix

INTRODUCTION

A viral etiology for outbreaks Of acute infectious nonbacterial gastroenteritis was
ﬁrst conﬁrmed in 1972 (89), when virus was isolated from an outbreak in Norwalk, Ohio.
Christened Norwalk virus, it was subsequently shown to be a frequent cause of foodborne
and waterborne gastroenteritis in the United States. During the 1970’s and early 1980’s,
the clinical and epidemiological features of such outbreaks were characterized. An
increasing number of different viral agents, with clinical aspects similar to that Of Norwalk
virus, were also identiﬁed during this time. This group became known as “small round-
structured viruses” or Norwalk-like viruses (NLVs).

Laboratory identiﬁcation of these viruses was hampered by the fact that the viruses
do not grow in cell cultures or animal models, so conﬁrmation of outbreaks was rare.
Sew-epidemiology has conﬁrmed that infection with NLV is extremely prevalent
worldwide, and generally occurs at younger age in developing countries. Surveillance and
reporting of NLV outbreaks have also been deﬁcient, and the role of NLV illness in the
United States and worldwide has probably been grossly underestimated. Recently
developed laboratory methods have enabled conﬁrmation ofNLV etiology, and can be
expected to conﬁrm the suspected role of NLV as a major cause of infectious nonbacterial
gastroenteritis worldwide. Public health professionals working “in the trenches” as health
department sanitarians and disease control professionals may enhance their investigational
abilities relative to gastroenteritis outbreaks by having a clear understanding of the clinical
and epidemiologic features of Norwalk-like viruses. This literature review will hopefully

enchance this understanding.

Chapter 1

DESCRIPTION OF NORWALK-LIKE VIRUSES

Norwalk virus (NV) is the prototype of the Norwalk-like group of viruses (NLVs).
This group is also called “small round structured viruses” because viewed using electron
microscopy they have an amorphous surface with a feathery ragged outline that lacks
geometric symmetry. These 27-35 nm viruses also share other characteristics: they have
similar densities in cesium chloride (1.34 to 1.42 g per milliliter), they cannot be cultivated
in cell culture, and they derive ﬁom epidemics Of gastroenteritis. It is estimated that an
infectious dose ofNLV may be as little as 10-100 virus particles. Many strains are named
after the location of the outbreak. NLVs possess single stranded RNA of positive sense,

and are members of the family Caliciviridae (3, 12, 84).

Chapter 2

CLINICAL FEATURES OF NLV INFECTION

Description of Symptoms

Outbreaks caused by Norwalk-like viruses are characterized by median, dose
dependent incubation periods of 24-48 hours (range 2 hours to 7 days) , median duration
of 12-60 hours, and a high percentage of patients with nausea, vomiting, cramps and
diarrhea (9, 19, 43, 74, 91). Sudden projectile vomiting has been particularly associated
with NLV illness without a deﬁned prodrome (19, 21). Children and adolescents are
likely to experience vomiting more frequently than diarrhea, while adults frequently
experience higher rates of diarrhea than of vomiting (1, 94, 158, 169). Myalgia, sore
throat and chills are less commonly reported (91). Headache (43%) and fever (25-49%)
are also reported frequently (8, 35, 43). The disease is Self-limiting and generally benign,
occurring most frequently from September to March in the northern hemisphere. It

seldom requires medical treatment and has no known sequelae (19).

Histopathological Features of N LV Infection

The histological response to NLV infection was characterized shortly after the
infectious agent was characterized (2, 13, 147). Typically, studies involved the use Of
volunteers who ingested bacteria-free ﬁltrates prepared aﬁer passage through a human
agent. Baseline biopsies were performed, establishing that entirely normal tissues were
present before the experimental infection with NLV. Pre-exposure biopsies showed that

overall duodenO-jejunal cell architecture was normal at baseline, featuring tall villi and

short crypts, and a villus-height to crypt-height ratio of at least 3:1. There were no
absorptive cell abnormalities, and polyrnorphonuclear leukocytes were absent (2, 147).

Post-exposure biopsies Showed a consistent pathology in subjects who became ill.
A mucosa] lesion of the proximal small intestine was seen 48 hours after ingestion Of the
NLV agent. The virus replication appears to occur in the mucosa] epithelium, resulting in
a broadening and ﬂattening of villi; the villus-height to crypt-height ratio was reduced to
2:1. Damage to mucosa] epithelial cells resulted in crypt cell hyperplasia. Many poly-
morphonuclear leukocytes and increased numbers of mononuclear cells were seen
throughout the lamina propria (2, 147).

In general, the epithelial absorptive cells were decreased in height and became
disorganized. There were focal areas of epithelial cell vacuolization with non-lipid staining
material. The epithelial cells showed a dilation of the rough and smooth endoplasmic
reticulum, along with an increase in multivesiculate bodies. Increased numbers of
lysosomal bodies were seen. In most cases, the microvilli were shortened, and
intercellular spaces were widened and ﬁlled with an amorphous electron-dense material.
Increased numbers of mononuclear cells and some polyrnorphonuclear leukocytes had
also inﬁltrated the intercellular spaces between epithelial cells (2, 147). Direct evidence of
viral maturation was lacking; no viral particles were Observed when specimens were
examined using electron microscopy (2, 20). In contrast to infection by invasive bacterial
agents such as Shigella and enteropathogenic E. coli, the mucosa remained intact.

Rectal and colonic biopsies taken at the height Of illness showed a normal
histologic pattern. This evidence, along with the absence of fecal leukocytes in Norwalk-

induced disease, suggests that the colonic mucosa is not involved in NLV syndrome (2,

12). Gastric lesions are not seen, and gastric secretion of hydrochloric acid, pepsin, and
intrinsic factor are unaltered. Gastric emptying was markedly delayed, however, and this
abnormal gastric motor ﬁrnction probably explains the frequent nausea and vomiting
characteristic Of NLV illness (12).

Peak viral shedding generally occurs 25-72 hours after inoculation, and has been
shown to persist up to and beyond 7 days (57, 133). This is true of both symptomatic and
asymptomatic infections: those with no symptoms have been shown to shed virus for up
to 6 days after challenge (57).

By 48 hours after ingestion of NLV, two of four asymptomatic volunteers had
developed mucosa] lesions which were indistinguishable from those seen in the two
volunteers with overt clinical illness. The histologic abnormalities were still present 5 days
after ingestion of NLV, about two days after symptoms had cleared. However, though
still present, the epithelial cell abnormalities and polymorphonuclear leukocyte inﬁltration
were less severe (2, 147). Convalescence biopsies done two weeks after inoculation
showed that the mucosa had returned to normal in all cases.

TO study changes in enzymology, Agus et a] (2) assayed brush border enzyme
activities in post-inoculation and convalescence biopsy specimens, and compared changes
experienced in each subject to that subject’s own baseline information. They found that
percent changes of alkaline phosphatase in ill volunteers (-49.3°/o ) and well volunteers
(+14%) were signiﬁcant (p<0.01). Similarly, the percent change in trehalase activity at
the time of illness (-61.%) was signiﬁcantly different (p=0.02) from what was seen in well
volunteers (-l6.2%). Percent changes in sucrase and lactase levels were also decreased

during illness, though the changes were not statistically signiﬁcant (2). Transient

malabsorption of D-xylose and fat has also been demonstrated during acute NLV
infection; the malabsorption is no longer present at 9-11 days aﬁer illness (2, 13, 147).
Several researchers have Observed virus capsid protein antigens and soluble
antigens in stools of patients infected with NV (57, 66). It appears that soluble antigen,
which apparently is lacking in viral nucleic acid, is present in higher concentrations than
capsid proteins. It is not known whether the soluble antigen is simply a degradation
product whose appearance may be variable, or if it has a speciﬁc biologic function in viral

replication (57).

Immune Response to NLV Infection

Studies of serological response to NVs using recombinant capsids indicate that
infection results in a rise in serum IgG and IgA antibodies, with IgM being observed in the
majority of volunteers (1 60). Several investigators have demonstrated that, paradoxically,
volunteers challenged with NLV who became ill were more likely to have high, rather than
low, prechallenge serum antibody titers (11, 57, 58, 73, 113, 160). Treanor et a] (160)
speciated the antibodies, showing that mean pre-challenge serum levels of IgG , IgA and
IgM were higher in individuals who subsequently developed illness. Serum IgM
responses were more common in ill individuals than in those who had subclinical infection,
but IgM responses were also noted in those with asymptomatic NV infections. It has been
suggested that IgM responses may be more strain-speciﬁc than IgG (83). It has also been
Shown that a wide distribution Of antibody levels exists in the initial serum specimens from
patients involved in outbreaks of gastroenteritis, ﬁ'om less than 2,000 to greater than

240,000 IgG units (122), and that the absence or presence of antibodies in the initial serum

did not correlate with protection against the occurrence Of a subsequent seroconversion.
These ﬁndings are consistent with the suggestion that serum antibody is not protective
against NV infection (11, 58, 88, 138,160). It should be noted that conﬂicting results
have been Obtained in studies of naturally acquired NV infection in Panama and
Bangladesh (10, 142). Interestingly, in a subject who was monitored continuously over
two years, antibody levels declined rapidly to the pre-challenge level after one year (3 9).

Graham et a] (57) described an association among infected groups wherein
increases in antibody titers in convalescent sera were signiﬁcantly higher in subjects who
had vomiting or vomiting and diarrhea. The presence of vomiting and nausea, and
headache or body aches correlated with the magnitude Of the seroresponses. They also
noted that titers Of pre-existing antibody were signiﬁcantly higher in subjects who
excreted virus.

Parker et a] (134) found that serum IgA response among volunteers was variable
and unpredictable. One ill subject mounted a signiﬁcant IgG response but serum IgA
could not be detected in pre- or post-challenge sera. Two ill subjects who showed no
signiﬁcant IgG response showed a signiﬁcant IgA response. One asymptomatic subject
showed signiﬁcant rises in both IgA and IgG titers. In summary, the mechanisms of
seroresponse to NLV infection are poorly understood.

The secretory immune response to NLV infection is also complex and poorly
understood. An early study evaluated the intestinal immune response to NV, measuring
the blocking activity of duodenal ﬂuid in 14 volunteers, using a method which did not
distinguish between speciﬁc antibody classes (11). In this study, titers of duodenal IgA for

ill and well persons were Similar, and none Of the volunteers developed a fourfold increase

in titers. Furthermore, all persons who had blocking activity in their duodenal ﬂuids
before NV challenge became ill after the challenge, while only 46% of those without
blocking activity became ill after the challenge. Okhuysen et a] (130) similarly
demonstrated that the presence of high titers of speciﬁc anti-NV duodenal IgA in
volunteers before challenge correlates with the likelihood of development of clinical illness
and the failure of protection against subsequent challenge.

Parrino et a] (138) observed that serum antibody titers to NLV were not
protective against illness in a study of 12 volunteers repeatedly challenged with NV. The
6 who remained well did not experience gastroenteritis upon rechallenge. Four volunteers
who became ill were challenged a third time 4-8 weeks after the second challenge. Only
one became ill: this volunteer had a high antibody titer before the third challenge.

Parrino et a] (138) also described a distinct group of individuals with low initial
titers who were resistant to experimental infection even after three challenges. These
individuals maintained low levels of serum antibody to NLV and failed to become infected.
This ﬁnding has been conﬁrmed by other researchers (57, 58, 88). It has been postulated
that some subjects (identiﬁed in studies in the US) lacked a receptor that would allow
NLV to enter the mucosa] cells of the small intestine and were thus intrinsically
nonsusceptible to NLV infection (88, 138). Whether speciﬁc immunologic, genetic, or
other host factors cause this resistance is unclear. Volunteer studies conducted in the US
using adults showed a lack of protective immunity, while apparent protective immunity
has been reported among children in developing countries (10, 142). These children also
had a history of rotavirus infection, perhaps of serial infections, possibly even concurrent

with NV infection. The number and type of infections, and age at infection may play a role

in the development and retention of protective immunity against NLV. Temporary
resistance to NLV may be passed to infants from mothers by breast milk containing virus-
speciﬁc IgA (44, 125).

Some degree of acquired protective immunity has been observed. Parrino et a]
(138) observed a short-term immunity lasting up to 14 weeks. Such immunity was
evidenced in 3 out of 4 volunteers in Parrino’s study. Subjects were rechallenged with
NV 28-42 months after the initial challenge, becoming ill both times. When challenged a
third time 4 to 8 weeks later, they did not become ill. The fourth subject became ill after
all three challenges. This apparent short-term immunity is consistent with other reports
(13, 58). In a larger volunteer study, Gray et a] (58) found that this apparently acquired
short-term immunity lasted up to 6 months in most individuals. They suggested that
acquired resistance to infection may depend on the rate and frequency of previous
exposure of the individual to NV.

The lack of protection afforded by preexisting serum IgG and IgA antrhodies (11,
58, 88), the unexplained, paradoxical association of high levels of serum antibodies with
clinical illness after challenge, the development of short-term immunity (5 8, 138) and the
occurrence of long-term immunity suggest that resistance to infection with NV may be
mediated by cellular mechanisms, probably in the gastrointestinal tract (130).

Recent research toward vaccine production has used the recombinant Norwalk-
virus-like particles (rNV VLP). Jiang et a] (84, 87) discovered that the capsid protein of
NV was one long molecule which could be produced in insect cells, and which self-
assembles into a hollow but complete viral capsid. Ball at a] (7) tested irnmunogenicity of

this rNV VLP in mice, wherein IgG response arose by 9 days post-exposure, and IgA

response arose after 24 days. The rNV VLP is thus irnmunogenic, but as was shown in
volunteer and outbreak-related serological studies, the induction of protective immunity

appears to depend on cellular, host, and other factors which are not yet well understood.

10

Chapter 3

LABORATORY METHODS FOR IDENTIFICATION OF NLVS

The 27-nm Norwalk virus was identiﬁed by immune electron microscopy in 1972
in fecal material derived from a 1968 outbreak of gastroenteritis in Norwalk, Ohio.
Classiﬁcation and biochemical and molecular analysis of the NLVs have proven difﬁcult
because they could not be adapted for growth in tissue culture. Although NV induced
subclinical infection in pygmy chimpanzees, all attempts to develop a practical laboratory
animal model have been unsuccessful (62,164). Volunteer studies that were instrumental
for the initial identiﬁcation of this 27-nm virus were thus continued not only to gain an
understanding of the natural history of NV and related viruses, but to generate fecal
material containing NV for use as an antigen reagent.

Laboratory conﬁrmation of the cause of outbreaks of food-home and waterborne
viral gastroenteritis requires either the detection of virus in stool or demonstration of a rise
in speciﬁc antibody. Virus can be identiﬁed by detection of viral antigens, or visualization
of the virus by electron microscopy, or by ampliﬁcation of the viral RNA using PCR
techniques. Proper collection and storage of specimens for testing, thus preserving the
integrity of the viruses present in the specimen, is prerequisite for recovering evidence of
viral infection. Instructions from the Centers for Disease Control and Prevention (CDC)
for collecting specimens to evaluate outbreaks of viral gastroenteritis are summarized in

Table 1 (31,74).

11

Table 1. Instructions for collecting specimens to examine for agents of viral

 

gastroenteritis.
Parameter Stool Serum
Source 10 ill persons; 10 controls for Same persons from whom stool
comparison (optional) was collected (controls Optional).
Specimen At least 10 ml/person in clean 15 ml (adults) and 3 ml (children)
dry containers blood specimens collected in tubes
containing no anticoagulants.
Time Within 48-72 hours after onset Collect acute phase specimens at
of illness same time as stool; collect conva-
lescent-phase specimens 3-4 weeks
after onset of illness.
Storage and Immediately refrigerate at 4°C (39°F) Refrigerate tubes of serum until
Shipping Place bagged and sealed specimens shipped with frozen refrigerant
with frozen refrigerant packs in insu- packs in insulated box. Keep
lated box. Send by overnight mail. specimens frozen by shipping
DO NOT FREEZE. on dry ice. Send by overnight mail.
Adapted from the CDC (31).

12

Immune Electron Microscopy (IEM)

Norwalk virus was ﬁrst identiﬁed in ﬁltrates of stools from people who were part
of an outbreak of nonbacterial gastroenteritis in Norwalk, Ohio (1). Convalescent stage
sera from these patients were used to aggregate virus particles, thus concentrating them
on the microscope grid for easier detection by transmission electron microscopy (89).
The [EM technique, developed in the early 1970’s, afforded important advantages to the
search for viral agents in stools. The ability of convalescent stage sera to aggregate virus
allowed IBM to be used both as a serologic assay and for virus detection. The aggregated
clusters of virus particles coated with antibody were critical in identifying particles of
potential interest and excluding irrelevant particles of similar morphology. IEM also
permitted the ﬁrst means with which to assess the presence of virus-speciﬁc antibody in
serum specimens, at least on a semiquantitative basis. The sensitivity of IBM was 10-100
times better than normal transmission electron microscopy (22); however, it remained
relatively insensitive and its utility is limited (74).

IBM requires highly specialized facilities and technical expertise, is extraordinarily
labor and reagent intensive and is unsuitable for the examination of large numbers of
specimens (43, 47). Norwalk virus (NV) was identiﬁed in stool by IBM in only three of
seven outbreaks investigated ﬁom 1977 to 1982 in which NV was serologically
determined to be the causative agent (68, 69, 70, 92, 101, 104, 157). In two of these
outbreaks, NV was identiﬁed by IBM in only 1 of 18 and 2 of 30 stool specimens.
Reports of better success in detecting NV by use of [EM appear to be the exceptions (69).
Because electron microscopes scan a ﬁeld only 0.000001 meters wide, between 105 and

10‘5 virus particles per ml of stool must be represent to be detectable (22). Within 48 to 72

13

hours after onset of symptoms, the virus concentration in stool declines below levels
detectable by IBM (159). For these reasons, IBM is irnpractica] for most public health

investigations.

Radioimmunoassay and Biotin-Avidin Immunoassay

IEM remained the best method of virus detection for NV until the development of
a radioirnmunoassay (RIA) technique at the National Institute of Allergy and Infectious
Diseases in 1978. The assay relies on the comparison of the binding of antigen to wells
coated with pre-challenge serum to binding in wells coated with post-challenge serum of
the same subject. Microtiter plates are pre-coated with Norwalk antibody are inoculated
with crude stool suspension containing Norwalk antigen. After overnight incubation, they
are washed and inoculated with a dilution of the serum to be tested (either pre-or post
exposure). After overnight incubation, an indicator consisting of radiolabeled ”’1 anti-
Norwalk IgG (convalescent serum)is added. The plates are incubated and washed, and
individual wells are placed in a gamma radiation counter. A 50% or greater reduction in
residual radioactivity produced by a serum sample as compared with a buffer control is
considered evidence of the presence of NV antibody. Differences in binding between
wells coated with post-challenge and pre-challenge serum are measured, and ratios of
binding with post-challenge to pre-challenge serum indicate the presence of antigen in the
specimen tested (64). Using a modiﬁcation of this assay, testing stool samples with
serum samples of known value, viral antigen was detected in the stools at the onset of
illness, but not before, and for a short interval after illness, usually 1-3 days (47, 66).

Antigen detection by RIA is about 10-100 times more sensitive than antigen

14

detection by IBM (47, 69, 70, 101). In addition to detection of NV antigen, one of the
ﬁrst applications of RIA was for the retrospective detection Of antibodies to NV in sera
collected during 25 outbreaks that occurred between 1966 and 197 7 (67). 111 persons
from 11 of these outbreaks had fourfold or greater rises in antibody titer between acute
and convalescent phase sera, a recognized standard indicating seroconversion. This study
and a subsequent evaluation of further outbreaks of acute infectious non-bacterial
gastroenteritis reported to the CDC between 1976 and 1980 suggested that NV was a
major cause of such outbreaks in the US (90). The detection of a rise in antibody by RIA
has been sufﬁciently sensitive to provide useﬁrl epidemiologic data. In several studies, the
number of patients with signiﬁcant antibody titer rises was apparently reduced by the late
collection of acute-stage sera a week or more after the onset of symptoms (8, 105).
However, in these studies and others, geometric mean antibody titers in convalescent
stage sera were greater in patients that in controls, supporting the conclusion that NV was
the cause of the outbreak (68, 94, 104, Ill).

Antibody to NV begins to develop within 5 days after the onset of illness. It peaks
within 3 weeks and begins to decline by the sixth week post-challenge (74). The early
antibody response of IgM and IgA peaks 2 weeks aﬁer onset of illness. The presence of
pre-existing IgG antibody in about half of the US population precludes the use of single
serum specimens to document recent infection in most instances (64). The necessity of
obtaining paired acute and convalescent phase sera complicates the use of techniques
demonstrating the rise of antibody titer, since it requires repeat contacts with patients aﬁer
they have recovered from their illness. Patients are often reluctant to provide serum

specimens for a mild, self-limited illness from which they recovered weeks before (22).

15

A disadvantage of RIA is that it utilizes potentially hazardous radioactive isotope-
labeled reagents. To eliminate the use of radioisotopes, CDC developed the biotin-avidin
immunoassay in 1984, where the 125I label was replaced by a biotin compound which was
safer to handle (74). The assay has a sensitivity comparable to RIA After its initial use of
to conﬁrm an outbreak of food-bome NV gastroenteritis in a school in 1984 (76), the
biotin-avidin immunoassay became CDC’S standard method of testing for NV antigen and
antibody for almost a decade (31).

Further shortcomings lie in that the RIA and biotin-avidin assays take six days to
perform. Also, antibodies used in the RIA or biotin-avidin assays can detect only the
Speciﬁc strain of virus involved in a particular outbreak, not the full spectrum of NLV
agents which have been identiﬁed. In a survey of 100 gastroenteritis outbreaks of
suspected viral origin submitted to CDC between 1985 and 1988, NV was identiﬁed by a
fourfold rise in antibody titers in approximately 20% of outbreaks. Approximately 40%
showed partial rises (less than half of persons with fourfold rise), suggesting that an
antigenically related agent was involved. The remaining 40% showed no titer rises at all,

indicating that an agent serologically distinct from NV was involved (22).

Enzyme-Linked Immunosorbent Assays (ELISA)

During the late 1970’s and early 1980’s researchers continued to develop
improved serological tests to detect NLV outbreaks. Like RIA, ELISA was developed
for detecting antibody to NV using specimens from experimentally infected human
volunteers. \Vrdespread application of these assays was limited by the difﬁculty in

obtaining, purifying, and standardizing human-derived antigen and antibody reagents

16

(122). In addition, reagents collected ﬁ'om human subjects could be broadly reactive,
reﬂecting the accumulated exposure history of the patients with the various antigenically
distinct strains ofNLVs (86). For a basic ELISA assay, a 96-well plate is coated with
puriﬁed antigen and incubated with the human serum to be tested. The plate is washed,
removing any unadsorbed human serum. It is then incubated with enzyme-labeled goat
anti-human antibody, which adheres to the antigen-antibody complexes. The plate is
washed again and the amount of adsorbed labeled antibody is then measured
colorimetrically (79, 87).

Recently, a baculovirus vector system was developed which could synthesize NV
capsid protein (83, 87). The recombinant-expressed Norwalk virus (rNV) capsid protein
self-assembles into empty virus-like particles. This method represented a major
breakthrough because large quantities of speciﬁc, standardized, puriﬁed capsid antigen
reagent could be generated with relative ease for large-scale laboratory use without
passage through human subjects. These particles are useﬁrl as antigens to produce
hyperimmune serum for use in a highly sensitive ELISA for detecting NV antibodies in
experimental volunteers and in epidemiologic investigation (57, 62, 87). Similar ELISAs
have since been developed for the Snow Mountain and Hawaii viruses, representing two
other antigenically distinct genogroups of NLV (1 13, 122).

Because these tests were originally designed to use human convalescent phase
serum from patients, they were of relatively low sensitivity because of the low afﬁnity of
the antibodies for the viruses (83). In addition, sources of serum were limited. The use of
rNV has permitted the generation of hyperimmune serum to NV in animals (87). The

hyperimmune serum and baculo-virus-produced rNV were in turn used to develop highly

l7

sensitive ELISA tests for detection of NV-speciﬁc antigen and serum antibody. Graham
et a] (57) reported that using rNV generated reagents, the sensitivity of ELISA for
detecting rNV in serial dilutions of puriﬁed antigen was comparable to RT-PCR. The
ELISA is also highly speciﬁc; the NV antigen ELISA failed to detect Snow Mountain
agent, Hawaii agent, HuCV, Sapporo, and astrovirus, as well as feline calicivirus,
poliovirus and other enteric viruses (57, 62). Thus the high sensitivity of the new ELISA
techniques and the unlimited supply of hyperimmune sera and NV capsid antigen made
possible by the use of rNV permitted research to move forward. Shortly after, the capsids
of Mexico virus, Southampton virus and recently Lordsdale virus have been cloned using

the baculovirus system, enabling a wider range of strain detection.

Reverse Transcriptase Polymerase Chain Reaction (RT-PCR)

The use of molecular epidemiology methods to characterize NLVs was made
possible when Jiang et a] (83) sequenced the entire genome of NV in 1992. Using
Norwalk virus-speciﬁc complimentary DNA (cDNA), Jiang et a1 (83) showed that NV
contains a polyadenylated single-stranded RNA genome of about 7.7 kilobases. Based on
this knowledge, they developed a reverse transcription polymerase chain reaction (RT-
PCR) method for the detection of NV in human stool, a method in which enzymes are
used to amplify the presence of viral nucleic acids. Other researchers have subsequently
developed improvements in technique, and identiﬁed new primer sets based on conserved
gene sequences in other NLV genogroups, including GH Southampton virus and

Lordsdale virus, which are useful targets for ampliﬁcation (4, 121).

18

Polymerase chain reaction assays are conducted as follows. Brieﬂy, the stool
Specimen is ﬁltered, denatured, and lysed to purify and isolate the viral RNA, which binds
to a silica-based membrane. Individual primers or sets of primers targeted at speciﬁc
conserved regions in the capsid and RNA polymerase gene sequences of the viral genome
are selected, and added to the reaction tube with enzymes. Precisely timed cycles of
heating and cooling cause replication of the gene sequences and the number of amplicons
increases logarithmically. Gel electrophoresis separates the amplicons into bands based on
their identiﬁable, speciﬁc number of base pairs. The formation of the signature bands
indicates presence of virus (119). High quality puriﬁcation of the viral RNA is very
important for obtaining positive results (83). It has been found that small particles and
soluble substances intrinsic to the material being tested may inhibit the PCR reaction (83,
133)

The sensitivity and speciﬁcity of this method is exquisite. Irang et a] (83)
estimated that stool samples may contain approximately 105 viral particles per ml, and
could detect NV RNA in stool samples diluted up to 10*, this suggests a detection limit of
as few as 10 particles and certainly 100-1000 particles. Because speciﬁc conserved
regions represent the “ﬁngerprint” of the virus, using only one set of primers would detect
only one target strain. Ando et a] (4) simultaneously used sets of primers which would
differentially detect several strains belonging to the antigenically distinct Genogroups I and
II. This allowed a wider range of virus identiﬁcation and formed the basis of techniques
now practiced in diagnostic laboratories throughout the US (Steve Michalik, MDCH,

personal communication).

19

Testing Food Specimens

In 1993, Atrnar et al (6) ﬁrst adapted RT-PCR techniques to attempt to recover
viruses from experimentally contaminated oysters. In separate assays, they were able to
detect as few as 9-90 virus particles in serial dilution of stock virus lysate, 300-3000 virus
particles after the addition of virus to oyster extract, and about 4100-41000 virus particles
in bioaccumulating oysters. When oysters were seeded with NV extracted ﬁom a human
stool specimen, they estimated they were able to detect 50-500 virus particles in the oyster
homogenate. More recently, a multiplex RT-PCR was developed for the simultaneous
detection of human enteroviruses, represented by poliovirus, Hepatitis A, and NV. Three
diﬂ‘erent sets of primers were used to produce three size-speciﬁc amplicons for the
respective viruses (6). When tested on serial diltuions of mixed, puriﬁed virus
suspensions, the method achieved detection limits of 1 PFU for HAV and poliovirus, and
1 RT-PCR—ampliﬁable unit of NV. When perfected, this method might enable rapid and
cost-eﬁ‘ective assays of foods by allowing testing for different strains at the same time. It is
of note that tests were not run on food samples, but in simple suspensions.

There are serious limitations in the present use of molecular biological techniques
that can reduce their usefulness in measuring the virological safety of foods. Compared to
bacterial pathogens which can actively grow in foods, enteric viruses are merely harbored
by them. These viruses are present in foods only as a result of incidental contamination
during production, processing, preparation or service. This presents a sharp contrast to
clinical specimens where viruses occur in vast numbers, being concentrated within the
human body through active replication. In foods, viruses may be present in minute

quantities; numbers suﬂicient to cause illness but below the detection limit of the RT-

20

PCR, due to factors intrinsic to the food specimen which interfere with detection. Current
PCR methods cannot yet consistently overcome these intrinsic factors, and are not yet
capable of reliably detecting these small but infective numbers of viruses.

The viability of the virus is an important issue in food quality control. The virus
may be inactivated due to partial degradation during storage, processing, or by other
means (144). While NLVs have been shown to be resistant to inactivation by heating at
60° C for 30 minutes, resistant to chlorine, exposure to salt water, the UV light used in
oyster deputation, and extended periods of freezing, it is uncertain the degree to which
the viruses that are inactivated by damage to the capsid protein maintain their intact
genomic structure (45, 47, 98, 99, 116, 119). Viruses sustaining damage to their capsids
may test positive by PCR even though they are not infectious (144). Likewise, the nucleic
acids may also be damaged by environmental stressors. This has diﬂ‘erent implications for
routine quality control testing as compared to outbreak investigation. For food quality
control testing, the presence of inactivated virus would represent false-positive detection,
possibly resulting in the condemnation of virologically safe foods.

By contrast, virus-containing extracts from foods may contain extraneous material,
such as acidic polysaccharides, glycogen, and lipids that inhibit RT- PCR reactions. Such
is the case when enteric viruses are extracted from shellﬁsh. Further, techniques used on
shellﬁsh have not been tested on shellﬁsh ﬁ'om different geographical areas and different
seasons when the levels of compounds such as glycogen, lipids and algae are especially
high. The presence of these inhibitors affects the sensitivity of the test and has been

shown to cause false-negative results (6, 144).

21

In summary, laboratory techniques for detecting viruses in foods are still
developing. Concentration methods are usually cumbersome, and yields are less than
optimal. Irnmunoﬂuorescent staining is expensive, slow, and requires highly trained
personnel. Alternative methods such as ELISA and RT-DNA probes have been tested but
are limited by high detection limits (>103 infectious units), unavailability of reagents, and
poor sample quality. Despite enormous strides in the ability to detect human enteric
viruses in foods using PCR, the technique is still limited by the absence of effective
concentration methods, the presence of enzymatic inhibitors, and the ability to distinguish

between infectious and noninfectious virus particles (82).

RT-PCR and Phylogenetic Analysis

NLVS possess Single stranded RNA genomes and were ﬁnally classiﬁed deﬁnitively
as members of the family Caliciviridae through sequencing of viral genomes and RT—PCR
analysis (3 7-40, 62, 83). The availability of the complete cDNA nucleotide sequences of
two NLV strains, Norwalk virus (NV) and Southampton Virus (SOV) has allowed the
development of primers for the detection ofNLVs in fecal samples by PCR. Detection of
different strains NLV have been reported using NV-speciﬁc primers and “degenerate”
primers to accommodate sequence variation between diﬁ‘erent NLV strains (3, 4, 62, 63,
128). On the basis of sequence variation within the RNA polymerase gene, the capsid
gene and the 3’ end Open reading frame (ORF2) (3, 4, 107), it has been shown that NLVS
can be divided into at least two major genetic groups, based on gene sequences in the
RNA polymerase region of the genome. Genogroup I (GI) includes two subgroupings:

the Norwalk Virus cluster, containing the viruses named Norwalk (NV), Southampton

22

via-12
I old-6
_ “1.7

 

 

ukt-t
uk1-4

 

 

 

 

0.10

F igurc l - Dendogram of phylogenetic relationship between HWA, Hawaii Agent and 30
other strains of NLV. The length of the abcissa to the conmcting node is proportional to
the genetic distance between sequences. DSV, Desert Shield Virus; SOV, Southampton
Virus; BRV, Bristol Virus; NV, Norwalk Virus; SMA, Snow Mountain Agent. Line
indicates 10% variation in nucleotide identity. From Ando et a] (4).

23

(SOV), Cruise Ship Virus (CSV), Desert Shield (DSV), Venlo (VV), and many other
numbered strains such as UK2; and the Gwynedd Virus (GV) cluster, containing UK-l
and other numbered strains. (Note that subsequently, Noel et a] (128) placed GV in to
Genogroup II based on sequence data in the capsid region. Genogroup II (GII) contains
5 subgroups; the Toronto Vrrus (TV) cluster, the most closely related to the Genogroup 1,
including Mexico virus (MV); Hawaii Virus group (HV), sharing only about 48% amino
acid identity with NV; Lordsdale Virus cluster (LV), a large group including LV, Bristol
(BRV) and Carnberwell (CAV) viruses; and the Snow Mountain Agent (SMA) cluster,
containing SMA, Melksham (MK) other numbered viruses (3, 4, 50, 107, 128). Figure 1
portrays a dendogram showing relatedness of various strains ofNLV by pairwise
nucleotide sequence identity (4). By characterizing and analyzing the anrino acid and
nucleotide sequences of the various distinct stains of NLVS it becomes possible to study
their distribution and occurrence in outbreaks and sporadic cases, and to elucidate the
epidemiology of NLVs.

Several studies demonstrated that strains belonging to G1 and GH are not only
genetically distinct but are also antigenically distinct. (107). Noe] et a] (128) observed
very little cross-reaction in patient seroresponses to the expressed antigens between the
two genogroups. However, the seroresponse to the expressed capsid antigens in subjects
infected with G1 strains differed from the seroresponse in those infected with G1] strains.
Patients infected with NV, SOV, CSV and DSV, strains which showed up to 37%
nucleotide and 38% amino acid divergence from prototype NV, demonstrated relatively
homogenous seroresponse to the single NV antigen. This contrasted with seroresponses

in those infected with GH strains. For example, patients infected with TV and HV strains

24

exhibited good seroresponses only when the infecting strain showed less than 22%
nucleotide and 6.5% amino acid divergence from the respective antigens. Crossreactivity
has also been shown between HV and SMA(113). Studies in Brazil have shown that
infection with and immune response to one strain often does not confer protection against
infection with a different strain within the same genogroup (137).

Ando et a] (49, 100) used ﬁrst RT-PCR techniques to investigate an large
multistate outbreak of NLV illness associated with oysters from Grand Pass/Cabbage
Reef, Louisiana. Using RT-PCR and other techniques, NLV strains of the P1A
phlyogenetic group (GI) were detected in 32 (86%) of the specimens, from 8 clusters in 4
states. Of the 32 PCR products Obtained, 19 (59%) from all 8 clusters had an identical
genetic sequence. Given that these 19 identical products were detected from
geographically separate areas, but epiderrriologically linked to the same contaminated
oysters, they concluded that the 32 products detected could only be represented by a
single strain from a common source which was responsible for this large outbreak (3).

Sugeida (155) successﬁrlly used nested-RT-PCR to detect two different genotypes
of NLV from stool and from oysters implicated in an outbreak of NLV foodbome illness
in Shizuoka Prefecture in Japan. This ﬁnding conﬁrmed the epidemiological link of the
illness outbreaks with the oysters, and the nested technique permitted detection of two
distinct genotypes of NLV coexisting in the same Specimen.

When a large number of outbreak strains were characterized using RT-PCR and
genetic sequencing by workers at CDC, it was found that strains from 2 outbreaks which
occurred in the same area in Arizona 53 days apart, and strains ﬁom 2 outbreaks that

occurred 74 days apart in South Carolina, differed by 2 and 8 nucleotides respectively, and

25

by 0 amino acids. This suggests that the same stain may persist in the same community
and result in repeated outbreaks. It also suggests that the genetic sequence may driﬁ with
continued passage. Over time, they may accumulate more and more sequence changes,
ﬁrst as silent changes in the third base position, then as coding differences as well (127).
Thus NLVS possess a mechanism to cause the phylogenetic tree of NLVS to continue to
grow and change, with different strains emerging and becoming predominant. The speed
at which such genetic drift may occur is currently under study. It has been shown in one
instance that drift did not occur after 4 serial passages through volunteers (C. Moe,
personal communication).

Molecular analysis of NLVS associated with outbreaks has demonstrated both a
great diversity of strains in circulation and the presence of a single predominant strain
during given time periods (50, 106, 128, 165, 166). Partial genetic sequences now exist
for over 100 strains of NLV. Despite this, few surveillance programs use PCR techniques
for the detection of NLV’s and fewer still characterize the NLV strains that could
potentially link outbreaks to a common origin (127). Only about 17 State and regional
laboratories routinely use RT-PCR to detect NLVS at this time.

NLVS were ﬁrst found naturally occurring in animals other than humans by
Sugeida in 1998 (155), then subsequently in cattle by others. In a paper published in
2000, van der Poel et al (164) reported that using RT-PCR techniques and genome
sequencing, they found that three phylogenetically diﬁ‘erent strains of calf caliciviruses
formed a tight cluster closely related to human NLVS from Genogroup I. The nucleotide
sequence identities between the calf NLVS and GI NLVS were 63-70%, with 75-77%

nucleotide homology. Swine NLVS were clustered as a separate lineage within

26

Genogroup II, with 69-71% nucleotide and 79-83% amino acid identity. The calf and pig
strains did not hybridize with the probe nrixture used to detect human NLVs. However,
the genetic distances between the animal and human NLVS are similar to the distances
between the GI and GI] human strains (164).

The close similarity of porcine and bovine sequences to the NLVS infecting
humans indicates the possibility of an animal reservoir for human infection. This raises
two important issues. The ﬁrst is that epidemic Spread of NLVS from an animal reservoir
may be possible. Secondly, these ﬁndings may at last lead to the development of an
animal model for the NLVS of humans; a pig model would enable studies of mucosal

immunity following NLV infection (164).

27

Chapter 4

SEROPREVALENCE

Once the Norwalk agent was identiﬁed in the 1970’s, several researchers set out to
characterize the prevalence of the NV in various populations. Detection of the virus was
dependent upon the sensitivity and speciﬁcity of the means of identiﬁcation;
radioirnmunoassay (RIA) examination of serological specimens was the best method
available. At this time, the detection was targeted at prototype Norwalk virus only; the
genotypic and antigenic differences of other NLVS circulating had not yet been
characterized.

Greenberg et a] (64) conducted the ﬁrst study of NV seroprevalence in various
countries using RIA techniques. They tested serum specimens from urban volunteer blood
donors in the US, Belgium, Switzerland and Yugoslavia; specimens ﬁ'om a longitudinal
study of enteric virus infections from villages in Bangladesh; specimens from Ecuadorian
Indians, including the highly isolated Gabaro tribe; healthy Nepalese villagers; healthy
Yugoslavian children, ﬁ'om specimens collected for a serological survey for HAV; US
children’s samples, from hospital admissions for respiratory and other non-gastrointestinal
complaints; and ﬁom US adult male and female homosexuals, as part of another
serological study.

Antibody to NV was found in serum from individuals in all of the areas studied.

Of the 861 persons tested, 71% had antibody. There was not a great difference in
antibody prevalence in male and female adults, nor between Western countries (US,

Belgium, Switzerland,) and the less developed countries such as Nepal, Bangladesh, and

28

Ecuador. The prevalence of NV antibody among homosexuals did not differ
signiﬁcantly from the other US blood donors (64). It was found that by the fourth decade
of life, antibody acquisition was at least 65% in all countries studied.

Of interest was the NV antibody prevalence Obtained in the Gabaro tribe of
Ecuador. This tribe lived in an extremely isolated village and had little if any contact with
neighboring tribes or Westerners prior to 1972, and very inﬁ'equent contact since then.
The inhabitants of the Gabaro village have had some hostile contact with other Indians and
outsiders, but this has not been of a sustained nature. None of the 16 adults or children
had an antibody response to NV indicating prior infection. The Gabaro are felt to be
closely related to the surrounding Indian tribes, so their lack of antibody was thought
unlikely to be due to a genetic inﬂuence. Apparently, NV had not circulated in this
isolated group in the 40 years prior to testing. Perhaps a small population cannot sustain
the virus. Greenberg et al 64) proposed that the virgin antibody status of this population
is evidence against a common animal vector for NV, although it is was not stated which
animals are “common” in remote Amazon Indian villages. A later study on isolated
Amazonian Indian tribes in Brazil using a more sensitive enzyme immunoassay showed
that seroprevalence ranged from 39% overall in one tribe to 100% in several others, two
of which ﬁrst contacted Westerners in 1970 (52). Reasons for this could be that host
differences actually exist, that the earlier methods were not sensitive enough to detect the
antibodies, or that other strains of NLV might have been present instead.

Poor sanitation conditions are likely contribute to earlier and more prevalent
infections, which may account for differences in the rates between developed and

developing countries. However, these studies show that NV infection is a common and

29

Table 2. Age distribution of healthy adults tested worldwide and percent testing
positive for anti-NV IgG antibodies.

 

Age (years)
Country 18-19 20-29 30-39 40-49 >50
South Africa 47.2% 65.9% 62.1% 62.1% -
Taylor (158) n=579
Singapore 63 65 - 64 -
Numata (129) n=50
England 74.9 81.4 85.5 91.6 87.9
Gray (59) n=l976
Japan 76 80 89 98 -
Numata (129) n=380
Sweden 88.2 80 87.5 89.4 94.7
Hinkula (77) n=122
Indonesia 90 100 - 92 -
Numata (129) n=90
Australian Aborigine 100 100 90.9 100 90.9
Parker (136) n=38
Papua New Guinea 100 100 100 100 -
Numata (129) n=50
Kuwait 100 100 100 100 100

Dimitrov (46) n=222

Note- to permit more accurate comparison, all studies cited above used ELISA methods.

Note- serum specimens cited here were tested only against prototype Norwalk Virus capsid antigen.

30

recurring condition throughout the world. The prevalence rate of _>_7 0% found in most
adult populations might reﬂect a steady-state condition in which reinfection with an
antibody response balances the disappearance of the antibody (64). Table 2 portrays the
seroprevalence of Norwalk virus infection among adults in several countries, as detected
using ELISA. Figure 2, a chart titled “Seroprevalence of NLVS” , highlights the
differences in prevalence of two strains of NLV in two developing countries, Indonesia
and Papua New Guinea, and Singapore, which enjoys a much greater level Of economic
development and sanitary infrastructure (86, 136).

As the genomes of other NLVS were sequenced in the late 1990’s, it became
possible to produce recombinant capsid antigen for other strains of NLV. Seroprevalence
studies were conducted using recombinant capsid antigens from Mexico virus (rMX),
Hawaiivirus (rHV), and Lordsdale virus (rLV), all belonging to GH ; and Southampton

virus (rSV) and rNV (GI strains), to see what relationships might exist. When seroassays
for NV and MxV were conducted on samples from 222 Kuwaiti adults, the prevalence
was 100% for both viruses in adults aged >20 (46). In Singapore, while the rate of NV
was very moderate at 63% among subjects aged 20-29 and 65% in those aged 30-39 yrs,
the rate of MxV antibody detection was 73% and 95% in the respective age groups (79,
129). Among Canadian subjects tested, HV antibody was predominant in subjects <20
years of age, when the prevalence of NV seropositivity became predominant (3 8, 39). In
Japan, MxV was more prevalent than NV in all age groups <50 years of age (79). In
Italy, LV antibodies were present in >93% of subjects of >5 years of age, while SV was
present in 26-3 8% subjects in age groups ranging from 1 year to>70 years (140). Studies

of subjects from London, England, and in UK showed that MxV infections occurred

31

Seroprevalence of NLVs
Singaxre (S), indmesia (I) quua (P)

0% ,
2039(3) 3039(3) 20290) 30390) 202903) 303903)
Ageinveaswamw

RaeNV Dieter/w

Seropositive
t

     

 

 

Figure 2 - Seroprevalence of NLVS in Singapore, Indonesia and Papua New Guinea

32

earlier in life, affecting children during the preschool years, whereas NV was seen to occur
later, among older children and adults (3 9, 136).

The predominance of virus strains may vary widely in different regions within the
same country (50,126, 127, 131,137). In Chile, a study was conducted comparing the
rates of seropositivity for MxV and NV in Punta Arenas and Santiago, cities separated by
hundreds of miles. The overall seropositivity for NV in Santiago was 83%, whereas in
geographically distant Punta Arenas 65% were seropositive for NV, a statistically
signiﬁcant difference. The rates of NV vs. MxV seropositivity also differed signiﬁcantly
between the two cities: 91% in Santiago compared to 76% in Punta Arenas (p<.001, x2,
Yates corrected). Note that study populations included similar proportions of subjects
from low, middle and high socioeconomic groups (131). Cubitt and Jiang (3 8) report that
strains of MxV and UK3 NLVS predominated in different areas of the UK during the same
year. In a study of Amazonian Indians of Ecuador, 4 isolated tribes had a NV
seroprevalence of 100% among adults, while 4 other isolated tribes had NV
seroprevalences of 50%, 55%, 30% and 67% (52). Figure 2 depicts the differences in
seroprevalence between two strains of NLV in three Asian countries. The varying levels
of exposure may be attributed to diﬂ‘erences n the predominant strain of NLV circulating
in that general time period. It appears that Singapore’s greater level of economic and
infrastructure development my be reﬂected in seroprevalence rates which are similar to
those in more developed Western countries.

The major diﬁ’crences in prevalence between countries lie in the age at which
infections and antibody are acquired. In the pediatric populations studied, children ﬁom

the developed world including the US, Japan, UK, and South African whites generally

33

Seroprexdence

Seroprevalence in Children
Norwalk Wrus (NV)

1CD°/o
93%
80%

70% ‘ “
.0. \
50°/oi~ s

40%)
30%-
”/09?-
10%1

Cl

 

 

 

 

dime-£09m 9311 121151018102324130303637'48 5yr > By
A96

+ Aborigine + Mexico

Figure 3 - Norwalk Vrrus Seroprevalence in Children

34

acquired antibody more slowly than did children from developing different age groups in
developing countries like Ecuador, Bangladesh, Mexico, and among South African blacks
(11, 59, 64, 86, 136, 157). Children are born with IgG antibody roughly equal to that of
the mother. The maternal antibody disappears by the fourth month of life (136).
Antibody levels then rise after the initial infection with NLV. Studies show that NV
infection appears to be relatively uncommon in early childhood in developed countries like
the United States and Japan, while in many developing countries, primary NV infection
occurs very early in life (12, 65). In Mexico and among the Cuna Indians ofPanama, less
than halfof the study children of less than 2 year of age had antibody to NV, but by age
ﬁve, 98% had antibody (86, 142). In these cases, there is no Obvious window between the
decay of maternal antibody and primary infection, depicted in Figure 3 as a lack of a steep
trough after maternal antibody is gone but before the primary infection occurs (157).
What proportion of these differences is due to geographical location, population densities,
sanitary facilities, cultural practices or other factors has not been clearly established.
Prevalence studies among children in the developing world conﬁrm that primary
infection tends to occur very early in life (47, 64). Figure 3 illustrates the dramatic decline
of maternal antrbody at about 6 months of age. In children from UK and Japan,
representing developed countries, the rate of seropositivity remains low for several
months, indicating that the rate of primary infection is lower. The less evident decline of
maternal antibody among Mexican and Aboriginal children indicates a high rate of primary
infection with NLV at a very young age. By the age of 6 years, the level of
seroprevalence to NLV in Mexican and Aboriginal (and interestingly, British) children

equals the level in adults. By contrast, seroprevalence in Japanese children does not reach

35

adult levels until the late teenage years (79, 129). Children in the US and Japan acquired
primary infection at a much slower rate, approaching adult levels much later, during
adolescence (59, 136). Paradoxically, a study of healthy children in Finland found that
73% were positive for NV IgG antibodies by the age of 23 months, compared to a study
in Sweden where seroprevalence was 25% in children of g 5 years of age (77,109).

The use of sera collected for a longitudinal study of infectious diseases and
nutrition in rural Bangladesh allowed Black (10) to conduct seroprevalence studies for
NV and also to describe the annual incidence of NV there. The prevalence of antibody to
NV in children 3-6 months was 7% (aﬁer the decline of maternal antibody), and increased
rapidly to 100% in four year-old children. The incidence of seroconversion (Z4-fold titer
increases) was highest in 1-2 year-olds and in children who had low or undetectable levels
of antibody. The annual incidence of NV seroconversion was 29% among children of less
than 50 months of age. Overall, 46% of the children developed NV infection during one
year. The infection was most common during the December-March period, representing
the cool, dry period in Bangladesh, although moderate rates of transmission were
observed year-round (10, 41).

Distinct strains of NLV have appeared to predominate by age group. Comparison
of the acquisition of antibody to erV and rNV antigens indicates that in UK, MxV
infections occur earlier in life affecting children during the preschool years, while NV
occurs later, predominantly among adolescents and adults (38, 135).

Seasonal prevalence was the ﬁrst epidemiological descriptor for NLV infection: it
was christened ‘Vvinter vomiting disease” by Zahorsky (1), who ﬁrst described the illness

in 1927. Longitudinal data show that outbreaks of NLV illness are most common from

36

October through May, and one strain of NLV usually predominates among strains co-
circulating in a geographic area (106, 110, 127). It has been observed that the strain
which is predominant during the high season was often completely absent during the
preceding high season. Lewis (110) reported that the detection of 4 isolates of MxV from
British children in the summer of 1993 heralded the major epidemic which occurred in the
fall of that year. A similar pattern was reported in the Netherlands during the sarrre time
period (165, 166). Volunteer studies have shown that immunity to NLVS is short-lived;
thus herd immunity following an epidemic may play a role in the dynamics ofNLV
prevalence (110). It appears that over the winter season, infections with the predominant
strain are common and only a subset of patients actually requires medical attention which
leads to establishing a deﬁnitive viral diagnosis. During this period immunity to the
speciﬁc genotype may develop throughout the population and the prevalence of infection
with that genotype then declines and is superceded by a different genotype (106).

Recently, Noel et a] (127) described the global identiﬁcation of a common strain of
NLV in the Lordsdale virus (LV) cluster. Prior to 1995, the nucleotide sequence in the
RNA polymerase region in NLV outbreak strains was so diverse that no two strains from
diﬁ’erent outbreaks in the US were identical. In 1995, this changed when strains from 60
out of 109 (55%) dispersed and apparently unrelated outbreaks were founds to be very
closely related; 39 of the 60 (65%) had identical sequences, and an additional 21 differed
by only 1-4 nucleotides in this region. Further inquiry revealed that the same strain had
been identiﬁed in UK in February 1995, in Brazil, the Netherands, Canada, and Australia
in late 1995; Australia, the Netherlands and China in 1996, and in Germany before

October 1997. It is not understood how this strain Spread to geographically distant

37

 

locations in the US and abroad with no obvious common source of exposure and what
modes of transmission or selective factors allowed the sudden emergence and rapid global
spread Ofthis strain (127, 150, 156, 165, 166).

Seroprevalence information must be interpreted with an eye to several poorly
described factors. Blacklow et a1 (1 l) have suggested that either a genetic factor is
involved in protective immunity or repeated exposures to the virus are necessary to elicit a
protective immune response. Future research is needed to clarify the nature of both of
these important aspects. More research is needed to clarify the amount of crossreactivity
possible between speciﬁc NLV antibodies and other phlyogenetically closely related NLV.
Varying levels of crossreactivity between NV and other NLVs from genogroups have been
shown in adult volunteer studies (160). Also, human antibodies vary in their afﬁnity to
bind, and antibody response to infection varies widely between individual subjects (83).

Nonetheless, it appears that NLV infection is ubiquitous, and although antibodies
are acquired early in life, the role of NLVS as pediatric pathogens has yet to be elucidated.
More research is needed to describe the mechanisms responsible for resistance or

susceptibility to NLV.

38

Chapter 5
MODES OF TRANSMISSION

Foodbome contamination with NLVS is an important mode of transmission.
Foods can become contaminated either at source (primary contamination) or at the time
and place of preparation (secondary contamination)(5). Molluscan shellﬁsh are thus far
the only foods which have been found to be contaminated with NLV at the source. Other
potential sources of primary viral contamination are the application of sewage-polluted
ground or surface water during irrigation and fertilization. The growing trend toward the
land application of municipal sewage sludge could also be a source of contamination to
ﬁuit and vegetable crops. It has also been suggested from experimental studies that
mammalian viruses might be taken up through the roots of plants (5). Viruses could
conceivably be harbored in stem scar and bud tissues. They also might conceivably be
drawn through the skin into ﬁ'uits and vegetables along with rinse water when a
temperature gradient between the water and the ﬁuit exists, as has been shown to happen
with coliform bacteria in tomatoes and other fruits. However, no outbreaks of NLV
illness have yet been traced to these sources. NLVS might also be borne to hand-harvest
food crops by the hands of farm laborers. While this source of contamination has been
demonstrated for bacterial and parasitic pathogens such as E. coli 015 7:H7 and
Cyclospora cayatenensis, the association has not yet been reported for NLV. In most
cases when food items such as lettuce were implicated in outbreaks of NLV illness,
infected food handlers were also implicated.

Other important modes of transmission are sewage-contarninated water and

person-to-person, including aerosols traveling through the air to contaminate surfaces,

39

which will be discussed in detail.

Contaminated Shellﬁsh

A signiﬁcant source of NLV illness is bivalve mollusks, including oysters, clams,
cockles, and mussels. Because they are ﬁlter feeders, they are able to concentrate any
viruses which may be present in waters of the shellﬁsh beds due to sewage contamination.
In Britain, 169 outbreaks of illness associated with mollusks contaminated at the source
occurred between 1965-88; of these, 138 were conﬁrmed or consistent with NLV viral
gastroenteritis (5).

The consumption of fecally contaminated shellﬁsh has long been associated with
outbreaks of illness caused by enteropathogens, including NLVs. Raw shellﬁsh have been
implicated in outbreaks of foodbome viral gastroenteritis in several countries, including
the US, Europe, Australia and Japan (35, 55, 69, 70, 155). Most of the commonly eaten
shellﬁsh, including oysters ( 23-26, 35, 49, 69, 70, 155) , clams (124), and cockles (49)
have been vehicles of transmission for viral gastroenteritis. Norwalk-like viruses have
been the most common agents identiﬁed in outbreaks of oyster-associated gastroenteritis.

A recent review of seafood-associated disease outbreaks in New York for the
period 1980-1994 oﬂ‘ers a perspective on the proportion of illnesses caused by NLV in
seafood (167). Outbreaks in which a Norwalk or Norwalk-like viral agent was suspected
based on epidenrioiogic proﬁle, but not conﬁrmed, were classiﬁed as caused by a
“gastrointestinal virus”. During this time period , the etiologic agent was conﬁrmed for
654 (3 6%) of 1802 foodbome outbreaks, and 148 (44%) of the 339 seafood-associated

outbreaks. Of the seafood-associated outbreaks, l4 (9%) were attributed to bacteria, 69

4o

(47%) to viruses, and 65 (44%) to chemical agents, such as ciguatoxin and scombrotixin.
Seafood vehicles accounted for 85% of foodbome outbreaks caused by Norwalk virus. In
the 192 outbreaks where an agent was not conﬁrmed, 129 (67%) suspected agents were
compatible with NLV. Shellﬁsh accounted for 64% of the seafood-related outbreaks, and
all 204 outbreaks caused by conﬁrmed or suspected viruses were associated with shellﬁsh.
Norwalk or NLVS were conﬁrmed or suspected in 94% of those shellﬁsh outbreaks. Of
the 216 shellﬁsh-associated outbreaks, 210 (97%) were attributed to raw or lightly cooked
foods (167).

A series of major outbreaks associated with the consumption of raw oysters
occurring in the 19th and early 20th centuries, including one in 1855 that led to the deaths
of several “highly esteemed” citizens of New York from cholera, an outbreak of typhoid
fever at Wesleyan College in Connecticut in 1894, and a typhoid epiderrric in 1924-25, led
to the development of the National Shellﬁsh Sanitation Program (N S SP). This
cooperative program, involving FDA, State regulatory agencies and the shellﬁsh industry,
is charged with controlling the safety and quality of shellﬁsh shipped in interstate
commerce. Recommendations, such as limiting harvests to areas with clean water (<14
fecal coliforms/ 100ml water), depurating harvested shellﬁsh to reduce bacterial counts
below the market guidelines (230 coliforms/ 100 g meat) and requiring tags nanring the
location and date of harvest on all boxes of shellﬁsh sold to allow back-tracing of
contaminated lots and identiﬁcation of the contaminated shellﬁsh beds, have targeted the
continuing problem of shellﬁsh safety (49). NS SP standards require fecal coliform testing

of shellbed waters from various points at least once a month (25).

41

Oysters are ﬁlter feeders, and are believed to concentrate virus in the rnidgut gland
(155). It has been suggested that a period of depuration, wherein oysters are held in a
tank of disinfected sea water for a time to reduce the concentrations of harmful organisms
within the gut, might reduce overall levels of contamination (35). Experimental ﬁndings
suggest that oysters may be able to rid themselves of 99% of enteric viruses in 25 days
under laboratory depuration conditions (100). Unfortunately, this time period for
depuration is not economically feasible for the shellﬁsh industry. In Australia, the state of
New South Wales since 1979 has mandated a 36 hour period of depuration; the Australian
government is now evaluating this for national legislation. England and Wales have
adopted the EU Shellﬁsh Directive 79/923/EEC, which mandates depuration of oysters
grown in any but the most microbiologically pristine (Class A) waters.

Regardless, outbreaks of NLV illness have been associated with depurated oysters
(3 5, 55 69). Gill et a] (55) describe a large outbreak of gastroenteritis associated with
consumption of raw oysters which had been depurated for 72 hours in seawater. The
deputation tank held 5500 liters of seawater and about 8000 oysters. Water was
circulated continuously through a 30 watt ultraviolet light sterilization plant at 2730 liters
per hour; the average depuration rate was 0.3 5 liter/oyster/hour. The tank water was
recirculated through the ultraviolet light for 48 hours before the introduction of the
oysters. Depurated oysters had been independently tested by a hospital laboratory on
behalf of local authorities for 5 years prior to the outbreak; of the 78 samples tested at
regular intervals, only three had more than 5 fecal coliforms per ml and two of the three
samples had >15 fecal coliforms per ml. Studies of marine caliciviruses, more closely

related to human Sapporo virus than to NLVS, have shown that shellﬁsh exposed to

42

marine caliciviruses and held at less than 10° C in a continuous ﬂow of sterile sea water
still retained virus 60 days later, when samples were seen to grow in mammalian cell
culture. The marine caliciviruses remain viable for more that 14 days in 15° C sea water
(152). While this model may not fully represent the fate of NLVS in sea water, nonetheless
it appears that they can survive and persist in shellﬁsh beds for relatively long periods of
time.

The use of coliform counts as an indicator of quality of water in the harvesting
area has often failed to predict contamination of oysters with NLVS (25). In a large
outbreak involving several states stemming from Louisiana oysters, the oysters of
acceptable bacteriological quality were harvested from waters meeting NSSP standards of
quality (49). In an outbreak involving oysters from Appalachicola Bay, Florida, the
oysters were taken from areas with acceptable water quality; water was sampled from 39
monitoring sites in the bay three times in a two month period. No environmental source of
pollution was identiﬁed. Sanitation procedures at the oyster processing facilities where
seafood dealers purchased oysters also met the applicable sanitary standards (26).

Another incident, again involving Louisiana oysters, occurred in 1996-7.
Investigators interviewed 15 of 20 implicated oyster harvesters, and inspected 8 of their
boats. Seven of those boats had inadequate sewage collection and disposal systems, and
harvesters admitted to routinely discharging their sewage overboard. In a previous
outbreak it was found that harvesters ill with NLV gastroenteritis had disposed of sewage
overboard and contaminated a broad area of oyster bed (100). Sewage from Off-shore oil
rigs has also been the source of sewage contamination of shellﬁsh beds (27).

In the 1996-7 outbreak, the implicated shellﬁsh beds lay 12-15 miles from the

43

nearest community sewage outlet, recreational boating was inﬁequent in December,
commercial boating trafﬁc was inﬁequent because of the shallow depth of the water, and
all oil rigs were considered to have adequate sewage facilities. Molecular analysis of stool
samples from the six outbreak clusters identiﬁed three different strains of NLV which
were associated with three geographically separate harvest sites. Researchers suggested
that different harvesters who were infected during the same time period with the
genetically distinct strains of NLV, and each of whom dumped their sewage in different
waterways, possibly during favorable environmental conditions (low temperature and
lowered salinity due to an inﬂux of diverted fresh water), caused contamination of oysters
with NLV (27).

In an earlier Louisiana outbreak associated with oysters from Cabbage Reef, Kohn
et al (100) calculated that, if we assume an infectious dose of 10 virus particles and if we
assume that a 1 ml sample of stool from an ill person at 10’ dilution is still infectious, then
stool from an ill person contains at least 109 virus per liter. If we further assume that an ill
person’s output of stool is 1 L per day, that an oyster contains 25 m] of water, and that an
oyster can concentrate enteric viruses 50-fold (conservative estimate), then stool from a
single infected person over 1 day would be enough to contaminate an oyster bed
containing 2 X 108 liters of water. This is equivalent to a reef area 1 km long by 100 m
wide and 2mm deep, approximately half of the size of the entire Cabbage Reef harvest
area (100).

Rainfall, reduced salinity, total and fecal coliform counts, pH, turbidity and
sediment total /fecal coliform counts are all associated with the presence of viruses in

water. Other signiﬁcant factors such as currents, and geophysical factors such as bottom

44

topography, depth, inﬂow changes, and shoreline contours affect the rate and pattern of
the ﬂow of water from the source of contamination, making prediction of dispersal
difﬁcult. This complicates our understanding of the true relationship between virus
survival and the eﬁ‘ect of these environmental factors on the rate and range of distribution
of the virus in the environment (54, 118).

A recent perspective by Smith et a] (152) describes the emergence of caliciviruses
from ocean reservoirs causing mammalian disease. This is exempliﬁed by the emergence of
highly contagious vesicular exanthema of swine, caused by the feeding of infected raw ﬁsh
offal to swine, which caused devastation of the California swine industry in the 1930’s. In
another instance, shellﬁsh beds on US coasts were positive for caliciviruses of unknown
type when tested with a cDNA calicivirus group-speciﬁc hybridization probe from a
marine calicivirus (San Martin Sea Lion virus SMSV-S), a virus which can infect ﬁsh of
commercial value and has infected humans, causing blister-like lesions. This review raises
interesting questions on the role the ocean might play as a reservoir to other caliciviruses
which have been known to cause zoonoses (152). Might the ocean also retain NLVS for
indeﬁnite, protracted time periods, or indeed as act as a reservoir?

While most outbreaks are attributed to oysters eaten raw (“oysters naturelle”) (3 5,
49, 155), the survival ofNLVs in lightly cooked shellﬁsh has also been documented.
Several incidents have resulted from oysters served steamed (23, 25, 26, 49), and roasted
(26) and ﬁom steamed clams (124). In a study by Dowell et a] (49), the attack rate for
oyster eaters at a large festival was 54% for those who ate steamed oysters only, while it
was 56% for those who ate both raw and steamed oysters. In the widespread shellﬁsh-

related outbreaks in New York, attack rates of 26% occurred among those who ate only

45

steamed clams, and 56% among those eating both raw and cooked clams (124). Similarly,
at two Rochester, New York clambakes, attack rates for those eating baked clams was
18%; for raw clams 60%, and for raw and baked clams 80% (161).

This indicates that in actual ﬁeld situations, temperatures used to steam or cook
shellﬁsh are often insufﬁcient to inactivate NLVS. At the Rochester clambakes, the clams
were prepared ﬁ'om raw clams on the half shell, topped with a breaded “casino mix”, and
allegedly baked at 177° C for 20 minutes. The casino mix consisted of bread crumbs,
green peppers, pirnientos and other seasonings, a mixture which may have insulated the
oyster meat from the heat. In spite of this cooking time, some persons who ate the clams
considered them undercooked (161). In a Louisiana outbreak, 33% of persons who ate an
oyster stew became ill. Oysters for the stew were reportedly sauteed until “brown and
shriveled” before being added to the stew. Of 11 persons who ate only ﬁied oysters,
eating 3-24 oysters each, none became ill (100).

McDonnell et a] (116) investigated the effect of cooking in the large Appalachicola
outbreak. They interviewed persons involved in the outbreak as to the subjective level of
doneness of their cooked oysters. While temperature could not be ascertained, the levels
of doneness were classiﬁed as less done, which included oysters described as ‘Vvet and
slippery” and “moist and juicy” and more done, including oysters described as “ﬁrm and
dry” and “tough and dried out”. They report high attack rates in all persons who ate
cooked oysters, with attack rates not differing signiﬁcantly between less done (71%) and
more done (53%). Kirkland et a] (98) also found no relationship between attack rate and
the extent to which oysters were cooked in another outbreak investigation. While this

doneness measure is subjective, it suggests that the level of cooking that would be

46

required to inactivate NLVS might render the oysters unpalatable to consumers (3, 99,
116). Protection of the shellﬁsh beds from fecal contarrrination offers the best defense
against shellﬁsh-borne NLV.

Virus particles have been shown to survive the standard cooking practices because
adequate internal temperatures may not be achieved throughout the preparation of the
shellﬁsh dish. Viruses are known to be inactivated by heat, which causes the coagulation
and breakdown of the virus protein coat, or capsid. However, the medium in which
viruses are held has been shown to inﬂuence virus sensitivity to thermal inactivation (45).
Using poliovirus model to test various cooking methods, DiGirolamo (45) demonstrated
that 10% of poliovirus innoculum could be recovered from oysters after 8 nrinutes of
stewing in milk, with a ﬁnal internal temperature of 75° C; 8 minutes of ﬂying oysters in
vegetable oil to a internal temperature of 100° C left 13% virus surviving ; oysters baked
for 20 minutes to an internal temperature of 90° C lelt 12.7% virus surviving; and 7-13%
of poliovirus innoculum added to oysters survived 30 minutes of typical steaming with a
ﬁnal internal temperature of 94° C ( 45). Other studies reported that it took four to six
minutes of steaming for the internal temperature of soft-shell clams to reach 100° C, but
only about 60 seconds for their shells to Open, which is the standard often used to
determine whether shellﬁsh are ready to serve (99, 124). Using feline calicivirus (F CV)
as a model for NLV, Slomka and Appleton (149) showed that in experimentally
contaminated cockles, immersed in boiling water for 30 seconds, and at a mean internal
temperature of 62° C, FCV remained infective in tissue 4/4 cases, though the titer was
reduced IOO-fold. When cockles were boiled for 1 nrinute, to a mean internal

temperature of approximately 78° C, F CV survived in 0/7 cases. Norwalk-like agents

47

L‘l.

have been shown to be relatively heat-stable and can survive at a temperature of 60° C for
30 minutes (48, 161). Neither poliovirus, which is more heat sensitive than NLV, nor
FCV, appear to be accurate models for NLV heat inactivation studies (48, 149). More
work is needed to deﬁne the time/temperature requirements for inactivating NLV in foods.

Consumption of other foods, such as beer, crackers, and hot sauce which has an
acid pH, at the same meals with the contaminated shellﬁsh has consistently been shown to
have no protective eﬂ'ects (25, 49, 98, 116).

Dose-response relationships have been observed . Dowel] et a1 (49) report an

attack rate of 40% among those who ate 1-5 oysters, 68% among those who ate 6-17
oysters, and 77% among those who ate 18 or more, (p=.16 x2 for linear trend) among

oyster eaters at a large festival. A dose-response was also observed in an outbreak where
45 became ill after eating raw oysters. The attack rate was highest among those who had
consumed more that 5 dozen oysters (91%) and lowest among those who had consumed
less that 12 oysters (46%), ( x2 for trend=3.98; p=0.05) (25). Morse (124) presents a
similar ﬁnding; attack rates were 34% among persons consuming 1-3 raw clams, 50%
among those consuming 4-9 clams, and 59% among those consuming 10 or more ( x’ for
trend = 13.92, p<0.001).

Current standards for tagging and identiﬁcation of oyster lots are not always
diligently applied, and do not guarantee rapid tracing for the purposes of identifying and
destroying contaminated shipments. Oysters can be traced to their harvest beds because
of the regulation requiring sacks of oysters‘to carry a tag identifying the harvester’s ID
number, the harvest date and the general harvest areas from which they are taken, and the

tags must be retained for 90 days after sale. However, they are not sufﬁciently detailed to

48

allow recall of oysters from a speciﬁc site, and they can be lost during shucking (26).

Oysters pass through a complex commercial network, ﬁom the harvest area, to
distributors, to large packers and shippers, to wholesalers, retailers and consumers (49).
Outbreak-related tracebacks have also revealed that records from wholesalers often do not
agree with the information on the oyster sack tag, and that harvester ID numbers cannot
be consistently traced to harvesters (23, 26, 49). Further, there is no mechanism for
forward tracing of oyster shipments once the source of the contaminated oysters has been
identiﬁed. Dowel] et a] (49) report that in the 1993 Louisiana outbreak inspectors were
able to document which merchant had had implicated lots of oysters conﬁscated or
destroyed, but could not account for most of the oysters which had been shipped. Most
state shellﬁsh programs do not routinely collect information about the number of oysters
or boxes destroyed during a recall Or about the number of potentially contaminated boxes
received and distributed prior to notiﬁcation about the recall (49).

Since shellﬁsh can be shucked, frozen and sold at a later date, contaminated
oysters may be available to consumers for months following recognized outbreaks. In a
large outbreak of NLV gastroenteritis in Australia, two further clusters of illness occurred
6 months after the original outbreak was identiﬁed, and were associated with consumption
of frozen oysters ﬁ'om the same lot (49). These ﬁeld observations of the viability of
frozen virus are supported by laboratory research. Much earlier studies using the
poliovirus model demonstrated that 13% percent of virus innoculum still persisted after 30
days in reﬁigerated (5°C), but by that time decomposed, oysters. In oysters held frozen at
-17.5°C, 10% of innoculum survived 12 weeks of storage (45). In another outbreak, a

large batch of NLV-contaminated cookies was frozen, and then portions were

49

subsequently served at different times to groups of people, causing a series of related
outbreak clusters over a period of fourteen weeks (120).

Theoretically, screening oysters for viral contamination might help protect eaters
of raw oysters. Current techniques relying on fecal coliform testing are an insensitive
indicator of viral contamination (25, 49, 100, 118); direct detection of NLVS would be
preferable. Kohn (100) reported that in the Cabbage Reef outbreak, NLV were not
detected in implicated lots of oysters by RT-PCR. Current RT-PCR techniques in oysters
have a lower limit of detection of 50-500 NLV particles when oysters are seeded with
known quantities of virus in the laboratory (6,100). This lower limit of detection appears
to be much higher than the infectious dose of NLV, which is thought to be about 10-100
virus particles. Also, the detection of NLVS in oysters contaminated in the wild has not
been reported. Natural substances occurring in the oyster homogenate which is tested
may also interfere with the ability of RT-PCR to amplify the viral genome (100, 149).

In the absence of detection techniques, recommendationss to prevent ﬁrrther
shellﬁsh-home outbreaks of NLV might include: improved surveillance and reporting of
shellﬁsh-related outbreaks of gastroenteritis; embargo shellﬁsh implicated in disease
outbreaks; adopt strict state and federal laws to control the sanitary quality of shellﬁsh
(the new seafood HACCP requirement is a step in this direction); increase participation in
the Interstate Shellﬁsh Sanitation Conference; provide an sufﬁcient number of
enforcement oﬁicers; develop adequate water quality standards which address viral as well
as bacteriological parameters (using poliovirus or other cultivable, sewage-related virus, as
an indicator); mandate a manifest-type tagging system; strictly enforce wholesale and retail

tagging requirements; require depuration of shellﬁsh sold (though this has yielded mixed

50

results in Britain ); and increase educational outreach concerning the risk of consuming

raw shellﬁsh (35, 55, 72, 143, 163).

Infected Food Handlers

Although shellﬁsh have been important regional vehicles for outbreaks of food-
borne NLV-related illness, transmission ﬁom infected food handlers appears to be a very
common and widespread phenomenon.

The dynamic nature of transmission by food handlers was demonstrated by an
outbreak in a bakery in which a single food handler experienced the onset of symptoms on
the way to work and had at least 5 episodes of diarrhea and 2 of vomiting throughout his
6 hour work day. During this time, he made 76 liters of butter cream ﬁ'osting that was
used on at least 10,000 frosted bakery items that were sold to the public. Observation of
the frosting preparation procedure showed that the baker preparing the frosting often
submerged his bare arm up to the elbow in the frosting as it was being prepared to break
sugar lumps and scrape down the side of the vat. At least 3000 cases, arising at a
corporate picnic, a wedding reception, a graduation party, and among the general public,
were attributed to this outbreak (74, 104).

In the majority of reported outbreaks that probably resulted from transmission by
infected food handlers, a food handler who was ill before or while handling the implicated
food item was identiﬁed. In 20 reported outbreaks reviewed here that probably resulted
from transmission by food handlers, a food handler who was ill before or while handling
the implicated food item was identiﬁed (see Table 3). Tossed salads, fruit salads or fruit

slices were implicated in 13 (65%). Cold foods, such as smoked trout, tuna and turkey

51

Table 3. Summary of Foodbome Outbreaks of NLV Gastroenteritis

 

Reference Number meeting Attack Rate Foods implicated NLV Illness status
case deﬁnition epidemiologically of foodhandler
Brondum (16) 29/58 50% Sandwiches Acute illness
Quiche squares
CDC (29) 99/835 12% Crumb cake Acute illness
pie, cinnamon
rolls, ice cream
Fleissner (51) 350/700 50% Baked beans, chicken Acute illness
26/87 30% potatoes, meatballs,
fruit salad. water
Gordon (56) 155/3 36 46% Shrimp newburg Asymptomatic
bisque
Grifﬁn (68) 38/41 92.7% Green salad Acute illness
25/31 80.6%
71/118 60%
Herwaldt (75) 217/527 41% Fresh fruit (Not stated NS)
Stuffed eggs
Iverson (81) 250/280 89% Melon, horseradish Asymptomatic
and
sauce, vermicelli probable post-
consomme’ symptomatic
Kilgore (97) 188/363 52% Salad Acute illness
Kuritsky (104) 129/248 52% Cake, frosting Acute illness
Lieb (111) 277/790 35% Tossed salad, ranch Asymptomatic
dressing, oil & vinegar
L0 (112) 195/N S NS Turkey salad sandwich Presymptomatic
Tuna salad, salad items
Parashar (133) 85/234 36% Sandwiches Postsymptomatic
and
asymptomatic
Patterson (139) 67/263 25% Ham, coronation Postsymptomatic
chicken
Reid (141) 23/31 85% Smoked trout Acute illness
7/19 31% soup, salads

52

salad, coleslaw, baked goods, sandwiches were implicated in 7 (3 5%).

In three of these outbreaks, chicken, potatoes, and shrimp entrees were implicated
as vehicles of NLV infection. Food cooking or holding temperatures were not reported.
It appears that while these foods were served “hot”, they were not held at temperatures
hot enough to inactivate the virus particles (51, 56,139).

In 3 outbreaks, asymptomatic food handlers prepared the implicated foods (51, 56,
103). In one of these, the food handler had a sick infant at home, and also had a
signiﬁcant rise in antibody titer to Norwalk virus, but did not experience symptoms (51).
This episode exempliﬁes asymptomatic virus excretion.

In other studies, asymptomatic persons associated with outbreaks have also
demonstrated signiﬁcant rises in anti-NV titers (76, 103, 133). Heun et a] (76) proposed
that their epiderniologica] data suggest that asymptomatic persons do not efﬁciently
transmit illness.

In a 1998 study using more sensitive laboratory methods of detection, Parashar et
al (130) detected NLV in the stools of an asymptomatic food handler who did not exhibit
a positive immune response to NLV. Further, in recent volunteer studies NV was
detected in stools of both ill and well volunteers. Viral shedding was detected in over
50% of well volunteers, and it persisted up to 2 weeks (130).

Graham et a] (157) conducted a volunteer study to measure serological responses
and viral shedding using new assays. Their work indicated an infection rate higher than
expected, with a high rate of asymptomatic infection and prolonged viral shedding. Of 28
sujects, 26 (82%) shed virus and developed an immune response following challenge.

One subject shed virus but did not seroconvert. Infection without any clinical symptoms

53

were seen in 8 (29 %) subjects, and mild symptoms were reported in 5 (18%). Thus, a
large percentage (47%) of NLV infections may occur without any or signiﬁcant
symptoms (5 7).

More problematic are four outbreaks in which food handlers became ill after
preparing or serving implicated food items. In an outbreak associated with green salads
served at a restaurant, the outbreak was initially recognized because of illness among
patrons of two luncheon banquets. Transmission occurred over a period of 6 days. One
of two workers who prepared lettuce for the tossed salads reported onset of illness the day
after preparing lettuce for the index banquets. This employee had a diagnostic rise in
antibody titer to NV (68). Two other outbreaks occurred among students and staff eating
lunches prepared in their school cafeterias. One of these was associated with sandwiches
which appear to have been contaminated by a cafeteria worker who placed the sandwiches
on plates without wearing gloves. This worker had a diagnostic rise in NV antibody titer
but did not become ill until 36 hours after serving the implicated sandwiches (74). The
second school-associated outbreak was associated with consumption of hamburgers and
french ﬁies handled 1-2 days before onset of diarrhea in two food servers who did not
wear gloves while serving. Further, their contact with the implicated food items was
reportedly restricted to handling pre-wrapped hamburgers and taking french ﬂies with a
scoop from a warming tray and placing them in plastic containers (74). Finally, an
outbreak of gastroenteritis in two hospitals due to a NLV was associated with a food
handler who became ill the day after preparing the implicated chicken sandwiches. This

food handler had two children at home who were ill at the time (74).

54

 

In these outbreaks the reported onsets of illness in food handlers overlapped the
distribution of illness onsets among patrons. Possible explanations for these observations
include inaccurate recall, or lying on the part of the food handler regarding the onset of
symptoms, transmission of the virus during the incubation period, or another unrecognized
source of viral contamination. From the standpoint of implementing public health control
measures, these scenarios have very different implications. In any outbreak investigation,
it is essential to keep an Open mind to the various possible modes of contamination, so as
to permit consideration of all possible means of prevention. Although food handlers may
seek to avoid blame for an outbreak and there is a natural tendency for investigators to be
suspicious of food handlers who deny being ill, biases Should be avoided. It is possible
that a food might be contaminated prior to its arrival at the restaurant; produce which is
chopped or shredded by the distributor could be contaminated. Such contamination at the
distribution level could be missed if a source within the establishment’s kitchen is assumed
at the expense of other possibilities (74).

Post-symptomatic contamination has been implicated in several outbreaks (74, 81,
130, 133, 139, 141, 169). In a 1982 outbreak stemming from a hotel kitchen, two salad
preparers who admitted to being ill were shown to have contaminated foods for up to 48
hours from the time their symptoms subsided. This was consistent with the shedding
pattern of Norwalk virus previously reported in a human volunteer study (159). In this
study, NV was visualized by immune electron microscopy in 2 of 11 stool specimens
Obtained 72 hours after onset of illness. Note that the sensitivity of IBM detection
methods is 10°-107 particles per gram; whereas the infectious dose of NLV is thought to

be 10-100 particles. In another outbreak, where deboned cooked chicken was implicated,

55

it was found that the person who deboned and prepared the chicken without wearing
gloves had been ill with NLV-like illness two days prior, during which time she stayed
home. When she was recovered and asymptomatic, she returned to work and prepared
the chicken, over 48 hours from the cessation of her symptoms (139). Reid et a] (141)
investigated a hotel outbreak lasting 8 days and found that the main vehicle of infection
was cold foods prepared by a food handler during and after a mild gastrointestinal illness.
The foodhandler was excreting NLV particles, detected by IBM, 48 hours after symptoms
had subsided (141). Most recently Parashar et a] (133) investigated a gastroenteritis
outbreak among employees of a manufacturing company and found as association between
disease and eating sandwiches prepared by 6 food handlers, 1 of whom reported
gastroenteritis which had subsided 4 days earlier. The sick food handler’s stool specimen
was obtained 6 days after the lunch, or a total of 10 days after recovery from illness. A
positive immune response was determined by IBM for sera from the sick food handler.
These reports are consistent with recent ﬁndings on postsymptomatic viral shedding in
human volunteer studies (130).

Early volunteer studies indicated that shedding of NLV occurs in <50% of ill
persons and does not persist beyond 100 hours after initial infection (159). Polymerase
chain reaction (PCR) analysis and new sensitive ELISA techniques indicate that viral
shedding is probably more prolonged than previously recognized. In a recent volunteer
study, NV was detected in both ill and well volunteers, by IBM, ELISA, and/or RT-PCR
(130). It was found more frequently in unformed than formed stools, was maximal within
the ﬁrst 72 hours after exposure, and was present up to 13 days after challenge. NV was

undetectable after 100 hours after challenge. Kaplan-Meier life table analysis showed that

56

after the ﬁrst challenge, ill subjects shed NV for a longer period than well subjects.
Subjects were challenged twice, 6 months apart and a third time 12 or 18 months after the
ﬁrst challenge. The duration of shedding was similar among groups after the second and
third challenges. Further, NV shedding occurs in >90% of ill volunteers, and in over 50%
of well volunteers, in whom in persisted up to 2 weeks (130). Similarly, Graham et a] (57)
found that virus particles are shed by infected symptomatic and asymptomatic subjects for
at least 7 days.

In several outbreaks, a policy of unpaid sick leave maintained by the food service
management has directly contributed to outbreaks, caused by workers who were sick and
felt “compelled” to work (74, 169). This “compulsion” can come in two forms: the
employee may not be able to afford to take time off, when pay would be lost; or the
management cannot afford to have an employee absent from work due to the tight labor
market and the lack of enough trained staff to ﬁll in for the sick worker.

Hedberg and Osterholrn (74) summarize the proﬁle of NLV gastroenteritis
outbreaks which were reported in Minnesota from 1981-1983. Salad items were
implicated in 6 of these outbreaks, and an ill food handler was identiﬁed in 5 of those 6
outbreaks (103). From 1984-1991, an additional 39 outbreaks were reported. Cold food
items were implicated as the vehicles in each outbreak; salad items were implicated in 12
outbreaks (3 5%). 111 food handlers were identiﬁed in 23 outbreaks (59%). In 6
outbreaks,(15%), the food handler who prepared the implicated food item was not ill but
had ill household members. Thus, in 74% of outbreaks of viral foodbome gastroenteritis
in Minnesota, a food handler source was identiﬁed. In 3 outbreaks (8%), food handlers

had onset of illness at the same time as patrons, and no ill food handlers or other sources

57

were identiﬁed in 7 outbreaks (18%). These cumulative results of food-borne disease
surveillance in Minnesota provide some perspective for evaluating the relative importance
of transmission routes in outbreaks of food-bome viral gastroenteritis.

Control of outbreaks of food borne illness arising from food handler transmission
requires removal of infected food handlers from contact with cold, lightly cooked, or
ready-to-eat foods and food preparation surfaces, cleaning and sanitizing of contaminated
surfaces and equipment, and disposal of contaminated food items. In several of
outbreaks, multiple ill food handlers were identiﬁed, some being violently and repeatedly
ill in the workplace. Of 18 restaurant-based outbreaks of food-bome gastroenteritis in
Minnesota from 1984-1991, multiple ill food handlers were identiﬁed in 14 outbreaks
(78%) (74). The frequency of transmission between food handlers within a kitchen,
transmission from asymptomatic persons, transmission by persons who have recovered
from illness, and apparent transmission from persons who are incubating infections make
control efforts more diﬁcult. The absence of paid sick leave beneﬁts for most food
handlers makes it economically disadvantageous for them to remain home when they are
ill. Removal of cold food items from the menu controlled one outbreak; however, hot
food items have been implicated in other outbreaks, and the potential for transmission
from infected servers has not been thoroughly evaluated (7 4). The State of Minnesota
recommends that ill food handlers be excluded from the workplace for 72 hours (74).
However, excluding food handlers from work for 48-72 hours after recovery from illness
may not always prevent transmission of virus, since virus may be shed for up to 10 days
after the resolution of symptoms (133). Exclusion may not always be feasible. The 1999

FDA Model Food Code, which is being adopted by many states presently, requires

58

workers to wear gloves when handling ready-tO-eat foods. While gloves can also be
abused and contaminated, this offers a feasible and generally acceptable barrier to viral
contamination, especially for unrecognized, asymptomatic cases.

When there is evidence of transmission of illness among the food handlers and
transmission to patrons over several days, it may be advisable to recommend voluntary
closure of the establishment for 72 hours to allow the virus to “run its course”, and allow
time for cleaning and sanitizing the equipment and facilities within the restaurant, and for
disposing of potentially contaminated food (74). In Minnesota , 6 of the 18 (33%) of
restaurants involved in outbreaks closed voluntarily due to the risk of ongoing disease

transmission to patrons (74).

Waterbome Transmission

Data reviewed by Kaplan et a1 (90) suggests that NLVS may responsible for 23%
of the waterborne outbreaks of acute gastroenteritis in the United States. The deﬁnition
of waterborne disease outbreaks used by the CDC for surveillance purposes is restricted to
illness that occurs after consumption or use of water intended for drinking (74). Classic
outbreaks of waterborne viral gastroenteritis have typically been associated with private
wells, small water systems, and community water systems. Several outbreaks arising from
cross connections, or physical links between potable water systems and non-potable or
waste water sources, have been reported (see Table 4). Taylor et a] (157) describes an
explosive outbreak of NLV gastroenteritis occurring at a school where 170 children were
absent on the ﬁrst day of the outbreak and 20-30 more were sent home sick. The

investigation focused on the water supply when sports teams from other areas who had

59

Table 4. Summary of Waterbome Outbreaks of NLV Gastroenteritis

Reference

Site Of Outbreak

Water Source

Details

 

Adler (10)

Bruglra (17)

CDC (29)

CDC (30)

Kaplan (92)

Kukkula (102)

Morens (123)

Taylor (157)

Public school

Large bakery

Campground

Rustic lodge

Commercial ice

Textile plant

Community water

Resort camp

Elementary school

School well
(presumptive)

Municipal water
system

Well

Stream

Well

Municipal water

Lake water

Spring water

Well

60

Coliform count negative
No cross connections found

Cross-connection between muni—
cipal supply and river water found
in nearby building High coliform
levels. Multiple strains of NLV
detected

Hydrant very near to septic tank
Dye test indicated septic leakage.
High coliform count, no chlorine.

Intermittent iodization of water,
Filter removed from puriﬁcation
system Stream water turbid
High coliform count.

Well had been ﬂooded by creek
following torrential rains.
Over 1000 cases of NLV illness.

Cross-connection with industrial
(surface water) supply.
Over 1500 cases community-wide

Sand filtration with substandard
levels of chlorine. High coliform
count. Water contaminated by
sewage outfall several miles away.
1500- 3000 eases of NLV illness.

Low chlorine levels. Spring was
Contaminated by nearby septic
tank; conﬁrmed by dye test. 418
cases of NLV illness.

Cross-connection between well and
septic tank

competed at the school, also became ill. It was highly probable that back-siphonage
through a cross connection was the cause. Well water was pumped through a pressure
tank through a ball-check release valve, with the desired pressure automatically maintained
by on-off cycling of the pump. As the pump shuts off and the water column in the well
shaft falls, a port in the valve opens to allow air into the system. As the pump turns on,
and the water column rises, the air is expelled through the valve. AS the valve closes,
some water is spilled. Because spills were frequent, custodians had placed a plastic hose
over the port and had run this to a ﬂoor drain. The ﬂoor drain had backed up near the
time of the outbreak, resulting in 6” of sewage backﬂow pooling on the ﬂoor and
submerging the end of the overﬂow hose. This resulted in the hose becoming a siphon
when the pump cycled on and off, and sewage containing NLVS was introduced into the
potable water system (157).

Contamination of potable water supplies by cross-connections with non-potable
supplies have occurred so often as to necessitate the promulgation of state and local
ordinances prohibiting cross-connections. Unfortunately, these ordinances often are not
vigorously enforced. In many cases, ﬁnancial stresses have reduced or eliminated cross-
connection inspection activities among regulatory agencies, usually local building
departments. In addition, cross-connections are often rigged after the ﬁnal construction
building inspection occurred, indicating that such inspections should be conducted on an
ongoing basis to be most effective.

Groundwater contamination of wells has occurred as a result of municipal lagoon
leakage, leakage from septic tanks, and ﬂooding after heavy rainfalls. Well construction

plays a role in the protection of groundwater; a properly constructed well with a pitless

61

adapter and grouted casing oﬁ’ers less possibility of allowing leakage to traverse the
vertical well shaft to reach the aquifer. Older, unprotected construction acts as a conduit
for surface pollution to reach groundwater, since water from the surface can ﬂow down
unto the earth unobstructed along the well pipe. The aquifer and overburden are also
important factors in groundwater source contamination. A solid bedrock aquifer is best
protected: clay overburdens are dense and add protection, while sandy soil or fractured
bedrock are quite permeable. Wells should also be isolated from known potential sources
of sewage, such as septic systems, both by distance and by higher elevation. Wells which
are not properly abandoned are a real danger to aquifer protection. Properly abandoning a
well involves removing the pump and backﬁlling the shaft with concrete or bentonite.
\Vrthout this, the Open well shaft is an unimpeded pipeline for pollutants to enter the
aquifer from the surface. Trends toward increasing land application of sewage sludge over
land which may contain abandoned farm wells could create major sewage contamination
of groundwater.

While groundwater is generally regarded as the purest and safest source of water,
sewage contamination of groundwater has resulted in outbreaks of NLV-related illness
(18, 29, 30, 74, 105, 114, 123). An outbreak ofNLV-like illness in Michigan inl970 led
to the ﬁrst recovery of virus particles from drinking water. Mack et al (114) concentrated
vaccine strain poliovirus by using an ultracentriﬁrge, proving that sewage contamination of
the well had occured. A large and interesting outbreak involving over 100 people
occurred at a large Arizona desert resort, which had a state-of-the-art, ﬁve branched
seepage sewage treatment system. When two of the leach ﬁelds became incapable Of

accepting eﬂluent, the others were overburdened, and the leach ﬁelds became saturated.

62

The undertreated sewage water then percolated down through silt and sandstone bedrock
into the water-ﬁlled caverns which were the source of the resort’s potable water. The
previously pure water in the deep bedrock caverns became contaminated and constant
pumping was not suﬁicient to remove the contamination. The previously pristine
groundwater could only be used for irrigation after this incident (105).

A sewage-contanrinated well was also the source of a large NLV outbreak
associated with ice. An outbreak involving hundreds of people in four outbreak groups
including a university football game and museum fundraiser, was traced to a large
commercial producer of ice (18). Prior to the outbreak, the manufacturer’s water wells
and septic tank were ﬂooded by water from an adjacent creek after a torrential rainfall.
Ice production was halted for 2 days and resumed after turbid water was pumped out of
the wells. The wells were not chlorinated before production was resumed. The ice
manufactured during this time period was associated with the massive outbreak. Several
families who obtained their water from private wells along the creek also reported
diarrheal illness that occurred a week after the ﬂooding (18). A relatively high attack rate
was present among those who ingested ice with alcoholic or carbonated beverages.
Neither the alcohol content of alcoholic beverages, the acid pH of carbonated beverages,
nor freezing affected the infectivity of the NLV (18). Indeed, ﬁeezing appears to have a
preservative effect on enteric viruses in water (42, 102).

The identiﬁcation of commercially manufactured ice as the source of this outbreak
highlights the need for regulation of ice and bottled water production. In the investigation
reported by Cannon et a] (18), jurisdiction over recalling potentially contaminated ice and

for ensuring the quality of ice production was unclear. Many containers did not carry

63

labels identifying the manufacturer or production date. Records of production and
distribution should be maintained for traceback purposes (18). No overriding federal laws
exist, and states vary widely in the regulation of the wells and the manufacture of ice and
bottled water (Richard Overmeyer, Michigan Department of Environmental Quality,
personal communication).

Surface water sources of drinking water, including springs and streams, have also
been contaminated by runoﬁ‘ associated with heavy rainfall or by sewage pollution.
Contributing factors in these outbreaks may have included the absence of ﬁltration and the
absence of or inadequate chlorination of the water supply (18, 29, 30). In Finland,
outbreaks were caused by consumption of sand-ﬁltered, chlorinated municipal water
drawn from a ﬁozen lake in which pollution incidents appeared to happen months apart,
and over 70 km away (102). Research has shown that enteric viruses persist for many
days in ﬁozen-over surface water, often surviving longer than the indicator fecal strep and
coliform bacteria which are commonly viewed as indicators of fecal pollution (42).

Studies have shown that NLVS are quite resistant to disinfectant in concentrations
commonly found in drinking water, and also in experimentally elevated concentrations
(95). In a volunteer study, dilute doses of NLV in distilled water were treated for 30
nrinutes with 3.75-6.25 mg/L sodium hypochlorite, yielding a ﬁ'ee chlorine residual of 0.5-
1.0 mg/L, as might be found in a drinking water distribution system; with 10 mg/L sodium
hypochlorite solution, which approximates the chlorine level present when super-
chlorinating a contaminated system; and with 0 mg/L. After the 30 minute treatment time
the chlorine was neutralized with sodium thiosulfate. The group ingesting the samples

with 0 chlorine had a 69% attack rate, those ingesting 3.75-6.25 mg/L had a 63% attack

64

rate, and those ingesting the 10 mg/L solutions had a 12.5% attack rate. The resistance of
NLV to chlorination may be enhanced by the natural tendency of the virus to aggregate
into clumps containing several viruses. Keswick reported that an outbreak of NLV-like
illness was associated with stored water with a residual 0.7-1.0 mg/L of iodine (95).

Thus, the routine disinfection treatment of NLV-contaminated source water has been
shown to be inadequate to inactivate NLV, consistent with the ﬁndings in the Finland
outbreak described by Kukkula (102)..

The role of water in outbreaks of viral gastroenteritis is broader than the CDC’S
surveillance deﬁnition implies. For example, an outbreak involving 1500 cadets and staff
at the US Air Force Academy was attributed to the consumption of chicken salad. No ill
food handlers were identiﬁed. However, celery used in the chicken salad had been washed
and soaked for an hour in water obtained from a hose that had been used previously to
unclog ﬂoor drains after sewage had backed up in to the kitchen (168). It was
hypothesized that virus particles were rinsed from the hose and taken up by the celery
during the wash and soak. Although this outbreak was food-borne, it appears that water
played the critical causal role.

A major class of outbreaks associated with water not intended for drinking is made
up of outbreaks associated with recreational waters. Outbreaks of NLV gastroenteritis
have been associated with swimming in lakes and swimming pools (8, 60, 94, 101, 131).

Swallowing water or immersing the head in contaminated recreational waters
increases the risk of illness (8, 60, 94, 101). These outbreaks typically occur during the
summer months in the United States. Infected individuals contaminating crowded

swimming areas can produce apparent point source outbreaks among exposed groups.

65

Potential sources of sewage contamination must also be investigated. Because of
potential for ongoing transmission, it may be necessary to close implicated swimming
areas for a minimum of 72 hours to prevent additional transmission.

An epidemiologic lead in deterrrrining whether the source of an outbreak is
foodbome or person-to-person, or is related to gross contamination of sewage is the
recovery of two or more strains belonging to NLV genogroups from victims. Sewage
systems concentrate wastes from large numbers of households within communities where
different genotypes of NLV are circulating concurrently, but rarely within the same
victims. Several outbreaks in which raw sewage contamination has been implicated have
resulted in concurrent infection with two or more NLV genogroups. This was
demonstrated among canoeists who ingested river water (60), in consumers of oysters
from contaminated waters (3, 100, 155), in potable water supply where heavy fecal
contamination was evidenced at a drinking fountain, and a community outbreak with
multiple genotypes of NLV resulted (17). Improving laboratory methods, mainly RT-PCR
techniques which can test for multiple strains at once, will assist in epidemiological
investigations by permitting the detection of mixed genogroups infections in victims, thus

enabling a ﬁner focus of investigational activities.

Person-to-Person Transmission

Person-to-person transmission, including direct contact, aerosol, and fomite
exposure, generally has been reported in outbreaks involving elder care settings, hospitals,
and cruise or military ships (33). In a study of 51 NLV outbreaks occurring in 1999 in

the US, the CDC found that 20% were due to person-tO-person transmission. About 10%

66

of reported NLV outbreaks in the US occur in extended care facilities for the elderly (56,
91). In Britain between 1992-1995, NLV accounted for 55% of all hospital and
residential facility outbreaks reported to the surveillance system. Of these, 97% occurred
in geriatric wards. The mode of transmission reported was mainly person-to-person in
93% of outbreaks in residential facilities, and a combination of foodbome and person-to-
person in and additional 2.6% (43).

A pathognomic symptom of NLV gastroenteritis is projectile vomiting. Vonritus,
in which virus particles have been identiﬁed by electron microscopy and RT-PCIL
represents a major source of infection (65, 97). Because 10° virus particles/m] must be
present for detection by electron microscopy, and if we assume that patients vomit a bolus
of at least 30 ml, then 3 X 107 virus particles, with an infectious dose of 10—100 particles,
will potentially be distributed into the environment (19). This is compounded as the
number of vomiting episodes and the number of patients in the space increases. The result
is the aerosolization of virus particles causing both airborne and enviromnental
contamination. Evidence for respiratory spread is lacking, since replication of NLVS in
respiratory mucosa] cells has not been documented. However, there is evidence of
airborne transmission via infectious aerosols, which impinge upon nasal and esophageal
surfaces and eventually are swallowed (19, 34, 65, 78, 97, 146).

Chadwick (34) describes a tour group traveling on an air conditioned coach,
wherein a lady vomited repeatedly on the outward leg of the trip, which included a 3 day
hotel stay. Two days later, 15 other members of the party, including the bus driver,
developed conﬁrmed NLV gastroenteritis. Over the next 6 days, 3 other members of the

bus group, 8 hotel guests, 10 hotel staff, and 2 environmental health ofﬁcers became ill;

67

these were ascribed to secondary spread. After interview and analysis, no food item was
implicated. Investigators concluded that illness was caused by aerosolized vomitus, which
might have been ﬁrrther distributed by the bus coach air conditioning recirculation system
(3 4). The predominant mechanism is practically impossible to discern; whether the
aerosol is breathed in and swallowed to cause illness, or if environmental surfaces are
blanketed with droplets which are then carried hand-to-mouth; probably both mechanisms
occur simultaneously.

Aerosol spread has been reported on cruise ships, in elder care facilities, military
facilities, and other densely occupied environments. Ho et a] (78) describes an outbreak of
NLV gastroenteritis aboard a cruise ship. There was no identiﬁable relation to food or
water consumption, but the risk of gastroenteritis among passengers who had shared toilet
facilities was twice that of those who had a private bathroom, and the rate of illness was
related to the number of passengers sharing a communal restroom (i.e. one or more
toilets). Contaminated bathrooms are an important vehicle for person-to-person spread of
NLV (33, 78, 148). In each cabin, index patients who had vomited in their cabins were
more likely to have cabin-mates who subsequently became ill than were patients who had
not vomited. Similar ﬁndings were reported in other shipboard outbreaks, one of which
continued through ﬁve consecutive voyages (71, 96). Many shipboard outbreak
investigations are inconclusive as to the mode of transmission, since the Shipboard
environment is such a crowded and interrelated system (32, 71, 96).

Sometimes pertinent facts relevant to person-to-person transmission may elude
investigators. After the conclusion of a major investigation of an outbreak of NLV aboard

a cruise ship (71, 75), it was revealed by housekeeping crew members to a ship’s Ofﬁcer

68

(author) that some cabin stewards would wipe down and polish the bathroom ﬁxtures and
drinking glasses with the same towel, since clean, sanitary replacement drinking glasses
were hard to get in a timely manner. This sort of incidental, hidden behavior could explain
how NLV environmental contamination can propagate illness from voyage to voyage after
all passengers and many crew have changed over ( 71, 75, 78).

Distinct winter-spring seasonal patterns of occurrence (43) and the lack of
seasonality (91 , 93) in institutional outbreaks have both been reported. Due to crowded
conditions, the infection can spread rapidly throughout the facility. Person-to-person
transmission may be identiﬁed by temporal clustering. In an investigation of an outbreak
in an elder care facility, Kaplan et a] (93) analyzed the data in various ways. By using the
attack rates and the number of rooms in the nursing home with 2, 3, or 4 residents
respectively, the numbers of rooms in which 0, 1, 2, 3, or 4 persons would be expected to
be ill given a random distribution of illness were calculated using the binonria] distribution,
yielding an analysis by geographic distribution. The observed numbers of double
occupancy rooms with 0, 1, or 2 persons ill were then compared with expected values; no
signiﬁcant differences were observed. When illness was examined as a function of
exposure to a roommate who had been ill one or two days earlier, by using a calculation of
person-susceptible days, the relative risk for becoming ill one or two days after a
roommate became ill versus becoming ill at other times was 3.74 (99% CI= 1.76, 7.96).
Among the staff, signiﬁcant diﬁ‘erences in risk of illness existed for employees who
reported daily physical contact with residents, compared with those who did not (x2:
8.64, p<.01). Kaplan cautioned regarding potential difﬁculties regarding the calculation

of relative risk. The inclusion of susceptible-person—days for individuals who are not

69

actually susceptible (due to genetic resistance or immunity), and the presence of
asymptomatic cases could distort the number of susceptibles. However, susceptible
person-days attributed to such individuals would probably occur randomly during the
outbreak period, and such random misclassiﬁcation should not affect the relative risk (93).
Other studies have supported these ﬁndings (56).

Among the elderly, NLV illness can have harmful consequences. Elderly people in
extended care facilities are known to be highly susceptible to outbreaks of NLV
gastroenteritis, probably due to decreased immunity and factors such as incontinence and
poor hygiene (38). Gastroenteritis has contributed to the deaths of elder care facility
residents by exacerbating heart conditions or diffuse athersclerosis. Hospitalizations have
occurred due to severe dehydration, and to injuries such as scalp lacerations, broken toes
and falls as elders attempt to make their way to toilet facilities during the acute illness (43,
56, 91, 93). NLV can thus hasten the death of an elderly debilitated person.

Actual airborne spread in controlled environments has been reported. Sawyer et a]
(146) report an explosive outbreak originating in a hospital emergency room where it was
shown that housekeeping staff who merely walked through the emergency area became ill
(relative risk=3.8, p=.029), as did ﬁiends and relatives of emergency patients who were
waiting in adjacent spaces. Recent reports of NLV illness investigations in Alaska
implicated airborne spread in three persons with tertiary cases who were exposed to
secondary cases who were vomiting (33).

Although airborne transmission has not been positively demonstrated in food
service settings, this failure is probably due to technical difﬁculties in conﬁrming that

transmission by aerosol occurred, rather than to the frequency of its occurrence (74).

70

Table 5. Summary of NLV Outbreaks Reporting Secondary (2°) Spread

 

Reference Site Primary scum 2° cases = % 2° Details
of NLV inf. Total cs. Cases
Adler (1) School Water supply 120/372 = 32.3% Household contacts.
Baron (8) Park Swimming in 21/111 = 19% Household contacts.
lake
Gordon (56) Retirement Foodbome 45/ 155 = 29% RR=6.5 for roommate
facility of 1° case.

Grifﬁn (68) Restaurant Foodhandler 12/38 = 31.2% Household contacts of
primary case.

Heun (76) School F oodborne 21/48 = 44% RR=5.5 for household
contacts of primary case.

110 (78) Cruise ship Not speciﬁed Total cases = 237 RR=2 for those using
common toilet facilities
vs. private toilet facilities.
OR=14 if roommate
vomited in cabin

Kaplan (93) Nursing home Not speciﬁed 15/45 = 33.3% RR=3.74 of 2° infection if
roommate is 1° ease.

Kappus (94) Municipal pool Fecal contam. 117/229 = 51% OR=12.1 for household of
index ease to have 2° case.

Khan (96) Cruise ship Ice 48/ 183 = 25% RR=1.6 for occupying a
cabin previously assigned
to 1° case.

Koopman (101) Municipal pool Feeal contam. 5/21 = 23.8% Household contacts.

Morens (123) Resort camp Well water 33/96 = 34% Household contacts.

Sawyer (146) Hospital Not speciﬁed 48/ 183 = 25% RR=3.8 for walking thru
contaminated area.

Sharp (148) Aircraft carrier Not speciﬁed 91/336 = 27% Common toilet rooms,
water use restrictions,
crowded conditions.

71

Secondary spread among household contacts is considered a feature of NLV
outbreaks: 8 out of 9 outbreaks reviewed by Kaplan et a] (91), and several other
outbreaks reviewed herein described secondary Spread (see Table 5 ). It is of note that
reporting secondary when secondary spread was not mentioned in the scope of the study;
“not reported” does not necessarily mean “not present”. In an investigation of an
outbreak of NLV in a school which speciﬁcally examined household secondary spread,
secondary cases were found to occur in 44% (21/48) of households in which there was at
least one primary case (76). The risk of secondary illness was 5.5 times higher in
households with sick school children or adults than in households with well school
children or adults in that outbreak. The secondary illness rate increased as the number of
primary cases in the household increased. The risk of secondary illness was related to age

of those susceptible: the preschool secondary illness rate was more than double the adult

secondary ilhress rate. This increased risk among younger children has been observed

elsewhere (157). Neither the age of the primary case nor the number of persons in the

household had any effect on secondary transmission (76).

The characteristic duration (5-9 days) of most outbreaks reviewed suggests that
outbreaks of Norwalk gastroenteritis terminate naturally in about one week. In a review
of several outbreaks in camps and cruise ships, continuing exposure to a common source
was responsible for longer outbreaks or repeated waves of illness when new susceptible
populations were introduced (91). Outbreaks of this nature have been interrupted by early
recognition and correction of deﬁciencies (such as cross connections), suggesting that
interruption of the longer outbreaks is possible. However, there is little evidence that the

course of a typical week-long epidemic has been altered by preventive measures (91).

72

Dispersing the susceptible population has been seen to interrupt ongoing outbreaks
of NLV illness (146,148). When shore leave was granted to sailors aboard an aircraft
carrier which was in the midst of an explosive outbreak, crowding was reduced
signiﬁcantly and the number of new cases declined substantially. When the crew returned
after two days, the number of new cases spiked again, and the outbreak resumed. After a
second major shore leave at another port some 18 days later, the epidemic ceased (148).
In a hospital-based outbreak, the closure of the emergency room, which was the epicenter
of infection, may have facilitated control of the epidemic (146). While potentially
effective, as when the residents of London ﬂed from the cholera in John Snow’s day,
dispersing the susceptible population seldom represents a practical control measure (148).

Once the index case of NLV is identiﬁed in a residential facility, generally
presumptively by sudden projectile vomiting, immediate isolation of the case and the area
is essential to prevent spread. Enteric precautions and attention to deep environmental
cleansing of hard surfaces with hypochlorite (including high risk areas such as toilets and
toilet rooms) is essential to containment. Contaminated and potentially contaminated
linens from adjacent beds, and other removable fabrics should be rapidly removed and
laundered and exposed consumables such as ﬁuit should be discarded. Anti-emetics,
administered intra-muscularly, rapidly terminate the vomiting, lessen dehydration, and act
as a sedative, reducing patient mobility. Nursing staﬂ‘ are at risk of aerosol exposure, and
respiratory masks may be considered. Handwashing and the use of gloves must be
fastidious. Staff working in a ward where NLV illness is present should not work in other
wards. Because NLV outbreaks are so common, infection control policies should consider

environmental contamination as a routine program (19, 22, 141). In the home setting,

73

environmental decontamination, rigorous attention to hand washing, and segregating the

patient from other family members may curtail spread of illness.

74

Chapter 6

SURVEILLANCE

F oodbome diseases cause an estimated 76 million illnesses and 5000 deaths in the
United States each year (33, 117). Although foodbome illnesses are common, only a
fraction of these illnesses are routinely reported to the CDC because a complex chain of
events must occur before a foodbome infection is reported; a break at any point in the
chain will result in a case not being reported. In addition, most reported foodbome
illnesses are sporadic in nature; only a small number are identiﬁed as being part of an
outbreak and thus are reported through to CDC. Most outbreaks are never recognized,
and those that are recognized often go unreported. The likelihood that an outbreak is
brought to the attention of public health authorities depends on many factors, including
consumer and physician awareness, interest, and motivation to report the incident. The
thorouglmess of the investigation is conducted is inﬂuenced by the ﬁnancial and personnel
resources available to the agency conducting the investigation, and by the consultative and
surveillance support of other local, state and federal public health and environmental
agencies. Outbreaks that are most likely to be brought to the attention of public health
authorities include those that are large, interstate, or food-service related, or that cause
serious illness, hospitalization, or even death (33, 117).

Front-line surveillance for illness caused by NLV’s and the detection and
investigation of outbreaks in the US often falls to local health jurisdictions. Local health
departments may or may not collect fecal specimens during the course of the outbreak

investigation, and may only arrive at a presumption of NLV causation after routine

75

laboratory analysis has eliminated bacterial pathogens as the source of the illness. Many
areas in the US are not served by laboratories with the capacity for conducting PCR
analysis for conﬁrmation of the presence of NLVs; at this time, only about 17 State and
Regional government laboratories perform these tests. (Steve Michalik, Michigan
Department of Community Health, personal communication, 33). Thus, the reporting of
NLV related illness is greatly hampered by the relatively mild symptoms which seldom
require medical treatment, the lack of recognition of an outbreak event, lack of resources
and/or expertise in local health departments to conduct thorough epidemiological
investigation, and lack of access to necessary laboratory services.

Presently, data in the US is collected through several systems administered by
CDC. The Public Health Laboratory Information System collects passive national
surveillance data for a wide range of diseases reported by physicians and laboratories, and
the Foodbome Disease Surveillance System receives data from all states on recognized
foodbome illness outbreaks. The National Health Care Survey, including the National
Hospital Discharge Survey, the National Ambulatory Medical Care Survey, and the
National Hospital Ambulatory Medical Care Survey, measure health care use in various
clinical settings, and collect information on patient characteristics, patient symptoms or
reasons for visit, provider diagnosis, and other information. Up to three symptoms are
recorded, and diagnoses are recorded using International Classiﬁcations of Diseases, 9th
revision (ICD-9-CM). Norwalk virus is coded as 0008.63, and other small round
structured viruses are coded as 008.64 under ICD-9-CM. Information on food related
deaths is collected by the national Vital Statistics System (117). The F oodbome Diseases

Active Surveillance Network (F oodNet), established in 1996, conducts active surveillance

76

for seven bacterial and two parasitic foodbome diseases within a deﬁned population of
20.5 million Americans (170). Additional surveys conducted within the F oodNet
catchment area provide information on the frequency of diarrhea in the general population,
the proportion of ill persons seeking care, and the frequency of stool culturing by
physicians and laboratories for selected foodbome pathogens. NLV gastroenteritis is not
included in this surveillance. It has been proposed within CDC to establish a similar
“CaliciNet” active surveillance network to more accurately assess the true burden of
Caliciviral disease, including NLV gastroenteritis (22, S. Monroe, CDC, personal
communication).

In a recent summary of foodbome diseases ﬁom 1993-1997, the Centers for
Disease Control reported that of the 2751 reported outbreaks, only 9 were caused by
laboratory-conﬁrmed NLVs. The number of outbreaks which were conﬁrmed generally
increased by the year during that time period. It must noted that 68% of the total
outbreaks reported were of unknown etiology, and approximately 50% of those had an
incubation period of 3 15 hours (33).

Surveillance of waterborne diseases is similar to that of foodbome diseases, and
suﬂ‘ers from the same shortcomings. Outbreaks which are recognized are reported to the
CDC, and summaries of waterborne diseases are released periodically by the Centers for
Disease Control. For the period of 1991-96, three summaries were published. During
1995-1996, one outbreak of NLV gastroenteritis was associated with drinking water
wells, and one with contaminated recreational waters (173). During the same time period,
8 out of the 30 reported outbreaks (36.4%) were of unknown etiology. In the two

previous reports, covering 1991-92 and 1993-94, no outbreaks were attributed to NLVS,

77

and 23/34 and 5/34 outbreaks respectively, were of unknown origin (171, 172).
Improving methods of detection and laboratory funding may improve the detection of

NLV-associated outbreaks, and hopefully reduce the number of outbreaks of unknown

etiology.

78

Chapter 7

BURDEN OF DISEASE

Early on, researchers recognized that the burden of disease caused by NLV was
likely to be greatly underestimated. Kaplan et a] (90, 91) developed criteria based on
epidemiological and clinical characteristics of NLV illness: stool cultures negative for
bacterial pathogens, and mean/median duration of illness 12-60 hr, and vomiting in 50% of
cases, and an incubation period of 24—48 hrs. Using these criteria to analyze CDC data,
Kaplan et al estimated that up to 45% of non-bacterial gastroenteritis was caused by NLV,
a ﬁgure which greatly exceeded with CDC’s estimate of 4% (90, 91). Kuritsky et a]
(103) analyzed data collected within the State of Minnesota reported from January 1981
through December 1983 using Kaplan’s criteria, ﬁnding that 35% of outbreaks were
consistent with NLV etiology. With the advent of novel and sensitive laboratory methods,
researchers were able to offer better estimates of the disease burden. Fankhauser et a]
(50) analyzed fecal and emesis specimens from 90 outbreaks of non-bacterial
gastroenteritis of undetermined etiology reported to CDC between January 1996 and June
1997. NLVS were detected by RT-PCR in 86 (96%) of the 90 outbreaks. Using similar
laboratory methods, Vinje et a] (165) attributed 91% of nonbacterial gastroenteritis
outbreaks in The Netherlands to NLVS.

According to the Centers for Disease Control, NLVS are probably a much more
important cause Of outbreaks that is currently recognized. Using information from
Foodbome Diseases Active Surveillance Network, the national Notiﬁable Disease

Surveillance System, the Public Health Laboratory Information System, and many other

79

surveillance networks, Mead et a] (117) formulated estimates of incidence for important
pathogens causing gastroenteritis, including NLVs. Known pathogens account for an
estimated 38.6 million illnesses each year, with the large majority, 30.9 million (80%)
attributed to NLV, rotaviruses, astroviruses, and HV A. Of the 30.9 million cases
attributed to these viruses, 23 million cases (73.3%) are attributed to NLVs. Mead et al
estimate that 46.1% of all gastroenteritis caused by common pathogens is foodbome; 40%
of all NLV cases are estimated to be foodbome (33, 117).

Thus, NLVS are estimated to cause 66.6% of all foodbome illness. Illness caused
by NLVS are estimated to cause 50,000 (32.9 5%) of hospitalizations, and 310 (6.9%) of
deaths related to foodbome illness (117).

Mead et a] (117) used the FoodNet population survey, and self reported, age-
adjusted data from the Tecumseh and Cleveland population studies to estimate the
frequency of acute gastroenteritis in the general population. They estimated that in the
US, there are 0.79 episodes of acute gastroenteritis per person per year. Extrapolated to a
population of 267.7 million persons, (the 1997 US resident population), this rate is
equivalent to 211 million episodes of gastroenteritis each year in the US. Using Mead’s
estimates to go one step ﬁrrther, if 59.5% of all gastroenteritis is caused by NLV, there
may be 125.5 million episodes due to NLV per year (117). If only one day of work is lost
per episode, the economic impact is substantial.

A proﬁle of the use of the medical care system, loss of productivity, and economic
impact of an outbreak of NLV was described by Kukkula et al (102). Investigating a
waterborne outbreak of NLV illness, they surveyed the entire population of a region, some

5000 people, regarding their health during the time period of the outbreak. From this

80

data, they estimated that 1700-3100 persons became ill. Because of the abrupt onset and
short recovery time, only about 50 people (1.6%-3%) sought medical care during the
outbreak. They examined records of absences ﬁom school and work, and found that
about 800 working days were lost; that is, 25-47% of those who were ill. The total costs,
including medical care, were estimated to be about US$3 00,000, for the outbreak caused
by this rather harmless virus (102). This comes to an average cost of $375 per person-
work day. Using this ﬁgure, multiplied by 125.5 million one-day episodes in the US,
multiplied by the lower ﬁgure of 25% of victims who miss work, means the cost of NLV

illness may exceed $25 billion per year.

81

Chapter 8

SIGNIFICANCE OF NLVS

The signiﬁcance of NLV illness lies in its role as a common cause of
gastroenteritis. Traveler’s diarrhea is one of the most common complaints among US
military personnel who participate in exercises conducted in foreign countries, and it can
affect enough personnel to compromise military effectiveness, especially during times of
military mobilization (15, 80). Infection is often acquired during off—duty time when
servicemen eat in local restaurants and have the same potential contact with endemic
pathogens as other travelers. The development of a Norwalk-speciﬁc ELISA made it
possible to assess the role of NLV in troop morbidity during the Persian Gulf War in 1990.
In a study of 883 Marines deployed to Saudi Arabia and Kuwait, Hyams et a1 (80) found
that 65% of troops reported at least one episode of gastroenteritis while in the Persian
Gulf area. The overall rate of NLV infection during a 5 month deployment was 6/ 100
personnel (99% CI 4-9%). There was no difference in the rate of NLV infection by age,
race or ethnicity, rank, birth location, or whether troops had previously been deployed
outside the US. NLV infections were found to be a signiﬁcant and widespread cause of
acute gastroenteritis among US troops during the Gulf War. While the speciﬁc sources of
infection could not be determined, person-to-person spread was important because of the
crowded living conditions, and rugged living conditions in the desert in which latrines and
communal bathing facilities were used (80)

Several outbreaks aboard US Navy ships have been reported, involving from

dozens to hundreds of sailors (36, 132, 148). Corwin et a] (36) describe a large outbreak

82

of NLV involving 450 personnel aboard an aircraft carrier, where the mean duration of
illness was 37 hours, ranging from 3-96 hrs. The attack rate was 44%, and loss of work
was reported by 39% of the ill population. In another large outbreak, involving 585/4500
crew members (13%), 8% of these sought medical attention and missed 1 or more days of
work (161). Extreme demands were placed on the medical department, and all supplies of
intravenous ﬂuids aboard ship were depleted by the end of the epidemic. Close living
conditions, limitations in the use of water for hygienic purposes, and common toilet
facilities were again thought to contribute to the outbreak. In both outbreaks, recurrences
in later months occurred aboard the ships, suggesting possible shipboard persistence of
NLV over time, despite periodic ship-wide disinfection efforts (3 6, 132, 148). Fleet
readiness was impacted by each of these outbreaks.

Among civilians, travelers’ diarrhea has the potential for wrecking a meticulously
planned business or vacation trip. Annually, 35 million people travel ﬁ'om industrialized
countries to a developing country. Among them, the incidence of travelers’ diarrhea ﬁ'om
all causes was 20-50% per 2 weeks’ stay in 1979-1981 (154). The increasing travel has
resulted in dramatic changes in the tourism industry and put pressure on the local
inﬁ'astructure that might inﬂuence the epidemiology of diarrheal illness. For example, in
Jamaica, the number of tourists increased from 0.4 to 1.2 million per year during the 1980-
1996 period (154)..

In a study done in Jamaica by Steffen et a] (154), over 30,000 tourists were
interviewed at the airport, and 322 travelers were enrolled at hotels to participate in the
study, which lasted 15 months. Those at the hotels who became ill with travelers’ diarrhea

(TD) were invited to submit stool specimens in return for medical care. While NLV was

83

not among pathogens tested for, the data regarding days lost, quality of life assessments,
and costs is applicable to NLV illness, which must surely be endemic in Jamaica.

The mean duration of visitors’ stay in Jamaica was 4-7 days. The overall attack
rate for any diarrhea was 24%, but only 12% had “classic” TD, deﬁned as 2 3 unformed
stools in 24 ours and 31 accompanying symptoms. Teenagers and young adults were
most at risk, with rates at least twice as high as those aged > 55 yrs. British visitors had a
higher attack rates (24%) than visitors from Canada (15%), US (11%), Germany (7.5%),
Italy (4%), Japan (3%), and Latin America (1%). The British susceptibility was not
explained, but lack of previous exposure or “so far undetected dietetic differences” were
suggested (154).

Steffen also included “quality of life” dimensions, such as the degree of incapacity
to function (in terms of leisure activities, sexual activity, and general well being). Ahnost
halfthe patients with “classic” TD were incapacitated, compared with 9% of those with
“mild” TD. The mean time of incapacitation was 12-17 hours (154).

Pathogen detection was 31.4%, when tested for a variety of bacteria] and parasitic
pathogens, rotavirus and adenovirus. Seasonal variations were Observed, with higher rates
(26-3 0%) of TD reported between May and October, 15% in December, staying below
20% until March. Vrruses and Campylobacter spp. predominated in winter. Based on
US foodbome illness estimates and rates of seropositivity world-wide, one may surmise
that a large proportion of the cases of unknown etiology may have been caused by NLV.
Few sufferers needed professional attention, and only two needed hospital admission, but
the average cost for medication and missed activities was estimated to be $116.50 per

patient per sick day (154). This does not include the irretrievably lost vacation time,

84

worry and ﬁustration which is entailed in illness abroad.

85

Chapter 9

RISK FACTORS

Woridwide, diarrhea is a common and debilitating disease either directly or
indirectly contributing to morbidity and mortality, particularly among children. Repeated
episodes of acute diarrhea and vomiting or persistent symptoms can not only decrease
quality of life, but also contribute to dehydration and malnutrition, and have adverse
developments on development and survival. Children under 5 years Of age in developing
countries experience the highest rates of illness and death due to diarrhea, with the
majority of disease occurring in infants under 1 year of age (137). While illness caused by
NLV is relatively mild, it may contribute to the adverse effects of superimposed infections.

Knowledge about the role of NLVS in pediatric gastroenteritis in developing
countries has been limited by the lack of available, sensitive and speciﬁc diagnostic
techniques, and has probably been underestimated. Several recent studies document a
high, age-dependent seroprevalence in young children, suggesting early and repeated
exposure (46, 59, 86, 135, 151). In an urban Brazilian shantytown, Parks et a] (137)
isolated 7 unique strains, from both GI and GII genogroups of NLV from a cohort of 186
children over a 16 month period. The high number of strains detected in a short time
suggest that there were multiple foci of infection in the community. Also, they found that
infection with one strain offered no immunological protection against infection with other
strains.

NLV diarrhea is more common in individuals of lower socio-econonric status.

Risk factors for these groups that favor the occurrence of diarrhea include overcrowding,

86

lower hygiene standards, and lower maternal education level, all of which favor oral-fecal
transmission. A higher number of beds in the home (an indirect measure of crowding),
younger maternal age, and consumption of seafood are additional risk factors (131, 137).

O’Ryan et a] (131) studied differences between populations in two Chilean cities,
and found that certain risk factors were signiﬁcant in one city and not in the other. Child
care center attendance and consumption of seafood were signiﬁcant risk factors in Punta
Arenas, yet not in Santiago. Thus, the predominance of a source of infection appears to
depend on the habits of the population. They also found that in individuals with lower
socioeconomic (SE) level, hygiene-related and food preparation habits are probably most
important, while among members of higher SE groups, acquisition of infection is probably
related to externally contaminated foods or water. This hypothesis is sustained by
observations that swimming, and seafood and vegetable consumption were more
signiﬁcantly associated with seropositivity in individuals of higher SE groups (131).
Despite the provision of universally potable water to all neighborhoods in Santiago, NLV
infections were common, demonstrating that this measure is not sufﬁcient for preventing
NLV infections. Knowledge of food safety and general hygiene must be improved in
order to intervene in the transmission of NLV to children at a very early age. The early
use of supplementary foods, such as baby formula, into the diets of infants, is also
associated with diarrheal illness (131, 137).

With regard to risk factors among travelers , Steffen et a] (154) found that the old
British colonial advice to “boil it, cook it, peel it, or forget it” was generally ignored by
vacationers. Approximately 80% consumed dairy products and tap water, and more that

55% are ice cream, hamburgers, and incompletely cooked chicken, lobster or shrimp.

87

Less than 3% reported to have avoided all potentially contaminated food and beverage
items. People aged 36-55 were slightly more negligent with regard to contaminated foods
compared to other age groups. Those with TD scored 11.2 (out of 24) in the risky food
score compared with 10.9 in those who did not have TD (P< .012). Tourists staying at all
inclusive hotels, wherein all meals are provided, had a higher probability of diarrhea (OR,
1.58; 95% CI 1.26,1.96; P<.001). Despite the colonial rules of dietary discipline,
consumption of high-risk “potentially hazardous” food was a weak predictor for the
incidence of all diarrhea, and was no predictor for classic TD (154). This is consistent
with the cold and ready-to-eat types of foods which are frequently associated with
outbreaks of NLV foodbome illness.

ImmunO-compromised persons are not only at risk for primary NLV infection, but
also chronic infection may be established and has been documented among AIDS patients.
In one study, an AIDS patient was found to shed the same genotype of NLV for an 8-
month period. This Shedding has been observed in chronic infections by other
gastroenteritis viruses (106). Chronic infection with NLV may contribute to the
deterioration of the general health of irnmunocompromised patients.

While the effects of normal cooking temperatures on NLV have been explored,
research has been conducted as to the effect of novel pasteurization methods such as
irradiation, electron beam, and ultra-high pressure pasteurization, which appear promising
in the reduction or elimination of bacterial and parasitic pathogens contaminating the food
supply. Unfortunately, none of these novel methods has yet been shown to inactivate
viruses (L. Jaykus, CDC, personal communication).

The American military establishment is very interested in the development of a

88

 

vaccine to prevent outbreaks of gastroenteritis which have been seen to affect the battle-
readiness of troops (7, 15, 36, 80, 132, 148). To date, the immune status ofNLV-infected
individuals has not been well deﬁned and constituents of protective immunity are not

known. At least one vaccine is currently being developed for testing in volunteers (7).

89

Chapter 10

SUMMARY

Norwalk-like virus infection is the epidemiologic prototype for outbreaks of
foodbome and waterborne gastroenteritis. Worldwide, NLVS appear to be major causes
of foodbome and waterborne illness. Until very recently, surveillance of NLVS
throughout much of the world has been hampered by the lack of sensitive and speciﬁc
means of laboratory identiﬁcation, and today, many areas still lack facilities, equipment
and trained personnel to conduct even passive surveillance. Assessment of the overall
signiﬁcance of NLV to the epidemiology of foodbome and waterborne illness depends
upon the availability of routine laboratory services to conﬁrm the viral etiology.

Outbreaks of NLV gastroenteritis have been propagated by contamination of water
supplies, raw foods, and ill food handlers. Controlling an outbreak depends on identifying
and eliminating the source of contamination. The occurrence of person-to-person
transmission and the transmission of NLV via aerosols make it necessary to evaluate the
potential for transmission by food handlers and servers in every outbreak, regardless of the
original source.

The proﬁle of an outbreak of NLV illness at the beginning of an outbreak may
initially be indistinguishable from Staphylococcus aureus or emetic Bacillus cereus
foodbome illness; thus all must be considered in the early hours of the investigation.
Kaplan’s criteria (90) for the diagnosis of gastrointestinal illness likely to be associated
with NLV can be used to evaluate the outbreak proﬁle. Kaplan’s criteria are:

A mean or median duration of illness of 12-60 hours; a mean or median incubation period

90

of 24-48 hours; vomiting in at least 50% of cases; and stools negative for bacterial
pathogens. Hedberg and Osterholrn (74) also noted that having more cases present with
vomiting than fever may be used as a further epidemiological criterion for NLV outbreaks.

In order to conﬁrm the presence of NLV in foodhandlers or other suspected index
cases, it is recommended that fecal specimens be collected from them at the outset, and
held under secured reﬁigeration during the investigation until bacterial pathogens have
been ruled out. By collecting the specimens at the outset, the chances of detecting NLV
are greater, since days can elapse during the time needed for the laboratory to test
specimens for bacterial pathogens. In this span of time, viral numbers in foodhandler’s
stools could decline to become undetectable, and the opportunity to conﬁrm the source of
the outbreak is lost. During an investigation, it is also important to remember that not
only potentially hazardous foods, but any food which has been handled with bare bands
can transmit NLV. From the very beginning, it is essential to coordinate the collection,
holding, transport, and subrrrittal of all food and fecal specimens with the laboratory, to
prevent mistakes which can cripple the investigation.

While NLV is not the among the most virulent causes of gastroenteritis worldwide,
it is indeed a very common one. Its ubiquity, the lack of protective immunity, and the ease
and speed with which it spread make it an illness which is likely to be prevalent for some

time to come.

91

BIBLIOGRAPHY

92

10.

11.

Bibliography
Adler JL, Zickl R. Winter vomiting disease. J Infect Dis 1969; 119:668-673.

Agus SG, Dolin R, Wyatt G, Tousirnis AH, Northrup RS. Acute infectious
nonbacterial gastroenteritis: intestinal histopathology: Histologic and

enzymatic alterations during illness produced by the Norwalk agent in man.
Ann Int Med 1973; 79: 18-25.

Ando T, Jin Q, Gentsch JR, Monroe SS, Noel J S, Dowel] SF, Cicirello HG,
Kohn MA, Glass RI. Epidenriologic application of novel molecular methods
to detect and differentiate small round structured viruses (N orwalk-like
viruses). J Med Virol 1995; 47: 145-152 .

Ando T, Monroe SS, Gentsch JR, Jin Q, Lewis DC, Glass RI. Detection and
diﬂ’erentiation of antigenically distinct small round-structured viruses. J Clin
Microbiol 1995; 33(1): 64-71.

Appleton J. Foodbome viruses. Lancet 1990;336:1362-1364.

Atmar RL, Metcalf TG, Neill FH, Estes MK. Detection of enteric viruses in
oysters by using the polymerase chain reaction. Appl Env Micribiol 1993; 59(2):
63 1-635.

Ball JM, Hardy ME, Atmar RL, Conner ME, Estes MK. Oral immunization
with recombinant Norwalk virus-like particles induces a systemic and mucosa]
immune response in mice. J Virol 1998; 72(2): 11345-1353.

Baron RC, Murphy FD, Greenberg HB, Davis CE, Bregrnan DJ, Gary GW,
Hughes JM, Schonberger LB. Norwalk gastrointestinal illness: an outbreak
associated with swimrrring in a recreational lake and secondary
person-to-person spread. Am J Epidemiol 1982; 115(2): 163-172.

Benenson A, Editor. Control of communicable diseases in man, 16th edition,
1995. American Public Health Association.

Black RE, Greenberg HB, Kapikian AZ, Brown KH, Becker S. Acquisition of
serum antibody to Norwalk virus and Rotavirus and relation to diarrhea in a
longitudinal study of young children in rural Bangladesh. J Infect Dis 1982;
145(4): 483-489.

Blacklow NR, Cukor G, Bedigan MK, Echeverria P, Greenberg HB, Schreiber
DS, Trier J S. Immune response and prevalence of antibody to Norwalk
enteritis virus as determined by radioirnmunoassay. J Clin Microbiol 1979;
10(6): 903—909.

93

12.

l3.

14.

15.

16.

17.

18.

19.

20.

21.

22.

Blacklow R, Greenberg H. Viral gastroenteritis (review). N Eng J Med 1991;
325(4): 252-261.

Blacklow NR, Dolin R, Fedson DS, DuPont H, Northrup RS, Hornick RB,
Chanock RM. Acute infectious nonbacterial gastroenteritis: etiology and
pathogenesis. Ann Int Med 1972; 76: 993-1008.

Blacklow NR, Herrmann JE, Cubbitt WD. Irnmunobiology of Norwalk virus.
Ciba Found Symp 1987; 128: 144-161.

Bourgeois AL, Gardiner CH, Thornton SA, Batchelor RA, Burr DH, Escarnilla
J, Echeverria P, Blacklow NR, Herrmann JE, Hyams KC. Etiology of acute
diarrhea among United States military personnel deployed to South America
and West Aﬁica. Am J Trop Med Hyg 1993; 48(2): 243-248.

Brondum J, Spitalny KC, Vogt RL, Godlewski K, Madore HP, Dolin R. Snow
Mountain agent associated with an outbreak of gastroenteritis in Vermont. J
Infect Dis 1985; 152: 834-837.

Brugha R, Vipond IB, Evans MR, Sandifer QD, Roberts R], Salmon RL, Caul
EO, Mukerjee AK. A community outbreak of food-home small
round-structured virus gastroenteritis caused by a contaminated water supply.
Epidenriol Infect 1999; 122: 145-154.

Cannon RO, Poliner JR, Hirschhorn RB, Rodeheaver DC, Silverman PR,
Brown EA, Talbot GH, Stine SE, Monroe SS, Dennis DT, Glass RI.

A multistate outbreak of Norwalk virus gastroenteritis associated with
consumption of commercial ice. J Infect Dis 1991; 164: 860-3.

Caul EO. Small round structured viruses: airborne transmission and hospital
control. Lancet 1994; 343: 1240-1.

Caul EO. Viral gastroenteritis: small round structured viruses, caliciviruses,
and astroviruses. Part I. The clinical and diagnostic perspective. J Clin
Pathol 1996; 48; 874-880.

Caul EO. Viral gastroenteritis: small round structured viruses, caliciviruses
and astroviruses. Part H. The epidemiological perspective. J Clin Pathol
1996; 49: 959-964.

Centers for Disease Control. Viral agents of gastroenteritis public health
importance and outbreak management. MMWR 1990; 39 (RR-5); 1-24.

94

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

Centers for Disease Control. Gastroenteritis associated with consumption of
raw shell-ﬁsh-Hawaii, 1991. MMWR 1991; 40(18): 303-305.

Centers for Disease Control. Multistate outbreak of viral gastroenteritis
related to consumption of oysters--Louisiana, Maryland, Mississippi, and
North Carolina, 1993. MMWR 1993; 42(49): 945-948.

Centers for Disease Control. Viral gastroenteritis associated with consumption
of raw oysters--Florida, 1993. MMWR 1994; 43(24):446-449.

Centers for Disease Control. Epidenriologic notes and reports multistate
outbreak of viral gastroenteritis associated with consumption of oysters -
Apalachicola Bay, Florida, December 1994- January 1995. MMWR 1995;
44(2);3 7-39.

Centers for Disease Control. Vira] gastroenteritis associated with eating
oystersnLousiana, December 1996-January 1997. MMWR 1997; 46(47): 1109-
1112.

Centers for Disease Control. Norwalk-like viral gastroenteritis in US Army
trainees-Texas, 1998. MMWR 1999; 48(11):225-227.

Centers for Disease Control. Epidemiologic notes and reports of viral
gastroenteritis--South Dakota and New Mexico. MMWR 1988;37(5):69-71.

Centers for Disease Control. Epidemiologic notes and reports outbreak of
viral gastroenteriﬁs--Pennsylvania and Delaware. MMWR 1987; 36(43): 709-
71 1.

Centers for Disease Control. Recommendations for collection of laboratory
specimens associated with outbreaks of gastroenteritis. MMWR 1990; 39 (RR-
14): 1-13.

Centers for Disease Control. Epidenriologic notes and reports gastroenteritis
outbreaks on two Caribbean cruise ships. MMWR 1986; 35(23); 383-384.

Centers for Disease Control. Surveillance for foodbome disease
outbreaks-United States, 1993-1997. MMWR 2000; 49(SS01); 1-51.

Chadwick PR Airborne transmission of a small round structured virus.
Lancet 1994; 343: 171

Chalmers JWT, McMillan JH. An outbreak of viral gastroenteritis associated
with adequately prepared oysters. Epidemiol Infect 1995, 115: 163-167.

95

36.

37.

38.

39.

40.

41.

42.

43.

45.

46.

Corwin AL, Soderquist R, Edwards M, White A, Beecham J, Mills P, Larasati
RP, Subekti D, Ansari T, Burans J, Oyofo B. Shipboard impact of a probable
Norwalk virus outbreak ﬁ'om coastal Japan. Am J Trop Med Hyg 1999; 61(6):
898-903.

Cubitt WD, Jiang X], Wang J, Estes MK. Sequence similarity of human
caliciviruses and small round structured viruses. J Med Virol 1994; 43:
252-258.

Cubitt WD, and Jiang X. Study on occurrence of human calicivirus (Mexico
strain) as cause of sporadic cases and outbreaks of calicivirus-associated
diarrhoea in the United Kingdom, 1983-1995. J Med Virol 1996; 48: 273-277.

Cubitt WD, Green KY, Payment P. Prevalence of antibodies to the Hawaii
strain of human calicivirus as measured by a recombinant protein based
immunoassay. J Med Virol 1998; 54: 135-139.

Cubitt WD, Blacklow NIL Herrmann JE, Nowak NA, Nakata S, Chiba S.
Antigenic relationships between human calicivirus and Norwalk virus. J Infect
Dis 1987; 156(5): 806-813.

Cukor G, Blacklow NR, Echeverria P, Bedigian MK, Puruggan H,
Basaca-Sevilla V. Comparative study of the acquisition of antibody to
Norwalk virus in pediatric populations. Infect Irnmun 1980; 29(2): 822-823.

Dahling DR, Safferman RS. Survival of enteric viruses under natural
conditions in a subarctic river. Appl Environ Microbiol 1979; 38(6):
1103-1110.

Dedman D, Laurichesse H, Caul EO, Wall PG. Surveillance of small round
structured virus (SRSV) infection in England and Wales, 1990-5. Epidenriol
Infect 1998; 121: 139-149.

Delves PJ, and Riott IM, Ed. Encyclopedia of Immunology, 2nd Ed., 1998.
Academic Press, Inc., San Diego, CA.

DiGirolamo R, Liston J, Matches JR. Survival Of virus in chilled, frozen and
processed oysters. Appl Microbiol 1970; 20(1): 58-63.

Dimitrov DH, Dashti SHA, Ball JM, Bishbishi E, Alsaeid K, Jiang X, Estes
MK. Prevalence of antibodies to human caliciviruses (HuCVs) in Kuwait
established by ELISA using baculovirus-expressed capsid antigens representing
two genogroups of HuCVs. J Med Virol 1997; 51: 115-118.

96

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

Dolin R, Treanor JJ, Madore HP. Novel agents of viral enteritis in humans. J
Infect Dis 1987; 155(3): 365-376.

Dolin R, Blacklow NR, DuPont H, Buscho RF, Wyatt RG, Kasel JA, Horrrick
R, Chanock RM. Biological properties of Norwalk agent of acute infectious
nonbacterial gastroenteritis (36508). Proc Soc Exp Biol Med 1972; 140: 578-
83.

Dowell SF, Groves C, Kirkland KB, Cicirello HG, Ando T, Jin Q, Gentsch JR,
Monroe SS, Hunphrey CD, Slemp C, Dwyer DM, Meriwether RA, Glass RI.
A multistate outbreak of oyster-associated gastroenteritis: Implications for
interstate tracing of contaminated shellﬁsh. J Infect Dis 1995; 171: 1497-1503.

Fankhauser RL, Noel J S, Monroe SS, Ando T, Glass RI. Molecular
epidemiology of “Norwalk-like viruses” in outbreaks of gastroenteritis in the

United States. J Infect Dis 1998; 178: 1571-8.

Fleissner ML, Herrmann JE, Booth JW, Blacklow NR, Nowak NA. Role of
Norwalk virus in two foodbome outbreaks of gastroenteritis: deﬁnitive virus
association. Am J Epidemiol 1989; 129: 165-172

Gabbay YB, Glass RI, Monroe SS, Carcamo C, Estes MK, Mascarenhas JDP,
Linhares AC. Prevalence of antibodies to Norwalk virus among Amerindiarrs
in isolated Amazonian communities. Am J Epidemiol 1994; 139(7): 728-733.

Gary GW, Anderson LJ, Keswick BI-I, Johnson PC, DuPont HL, Stine SE,
Bartlett AV. Norwalk virus antigen and antibody response in an adult
volunteer study. J Clin Microbiol 1987; 25(10): 2001-2003.

Gerba CP, Goya] SM, LaBelle RL, Cech I, Bogdan GF. Failure of indicator
bacteria to reﬂect the occurrence of enteroviruses in marine waters. Am J
Public Health 1979, 69(11): 1116-1119.

Gill ON, Cubitt WD, McSwiggan DA, Watney BM, Bartlett CLR. Epidemic
of gastroenteritis caused by oysters contaminated with small round structured
viruses. Brit Med J 1983; 287: 1532-4.

Gordon SM, Oshiro LS, Jarvis WR, Donenfeld D, Ho M, Taylor F, Greenberg
I-IB, Glass R, Madore I-IP, Dolin R, Tablan O. Foodbome Snow Mountain
agent gastroenteritis with secondary person-to-person spread in a retirement
community. Am J Epidenriol 1990; 131(4): 702-10.

97

57.

58.

59.

60.

61.

62.

63.

65.

66.

Graham DY, Trang X, Tanaka T, Opekun AR, Madore HP, Estes MK.
Norwalk virus infection of volunteers: new insights based on improved assays.
Infect Dis 1994; 170: 34-43.

Gray JJ, Cunliffe C, Ball J, Graham DY, Desserberger U, Estes MK.
Detection of irnmunoglobulin M (IgM), IgA, and IgG Norwalk virus-speciﬁc
antibodies by indirect enzyme-linked immunosorbent assay with
baculovirus-expressed Norwalk virus capsid antigen in adult volunteers
challenged with Norwalk virus. J Clin Microbiol 1994; 32 (12): 3059-3063.

Gray JJ, Jiang X, Morgan-Capner P, Desselberger U, Estes MK. Prevalence of
antibodies to Norwalk virus in England: detection by enzyme-linked
immunosorbent assay using baculovirus-expressed Norwalk virus capsid
antigen. J Clin Microbiol 1993; 31(4): 1022-1-25.

Gray JJ, Green J, Cunliffe C, Gallimore C, Lee JV, Neal K, Brown DWG.
Mixed genogroup SRSV infections among a party of canoeists exposed to
contaminated recreational water. J Med Virol 1997; 52: 425-429.

Green J, Gallimore CI, Norcott JP, Lewis D, Brown DG. Broadly reactive
reverse transcriptase polymerase chain reaction for the diagnosis of
SRSV-associated gastroenteritis. J Med Vrrol 1995; 47: 392-398.

Green KY, Lew JF, Jiang X, Kapikian AZ, Estes MK. Comparison of the
reactivities of baculovirus-expressed recombinant Norwalk virus capsid antigen
with those of native Norwalk virus antigen in serologic assays and some
epidemiologic observations. J Clin Microbiol 1993; 31(8): 2185-2191.

Green SM, Larnbden PR, Deng Y, Lowes JA, Lineham S, Bushell J, Rogers J,
Caul EO, Ashley CR, Clarke IN. Polymerase chain reaction detection of small
round-structured viruses from two related hospital outbreaks of gastroenteritis
using inosine-containing primers. J Med Vrrol 1995;45:197-202.

Greenberg HB, Valdesuso J, Kapikian AZ, Chanock RM, Wyatt RG,
Szrnuness W, Larrick J, Kaplan J, Gihnan RH, Sack DA. Prevalence of
antibody to the Norwalk virus in various countries. Infect Irnmun 1979; 26(1):
270-273.

Greenberg HB, Wyatt RG, Kapikian AZ. Norwalk virus in vonritus. Lancet
1979; i: 55.

Greenberg HB, Wyatt RG, Valdesuso J, Kalica AR, London WT, Chanock
RM, Kapikian AZ. Solid-phase microtiter radioirnmunoassay for detection of
the Norwalk strain of acute nonbacterial, epidemic gastroenteritis virus and its
antibodies. J Med Virol 1978; 2: 97-108.

98

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

Greenberg HB, Valdesuso J, Yolken RH, Gangarosa E, Gary W, Wyatt RG,
Konno T, Suzuki H, Chanock RM, Kapikian AZ. Role of Norwalk virus in
outbreaks of non-bacterial gastroenteritis. J Inf Dis 1979; 139: 564-568.

Griffin MR, Surowiec JJ, McCloskey KI, Capuano B, Pierzynski B, Quinn M,
Woynarski R, Parkin WE, Greenberg H, Gary CW. Foodbome Norwalk virus.
Am J Epidemiol 1982; 115: 178-184.

Grohmann GS, Greenberg HB, Welch BM, Murphy Am. Oyster-associated
gastroenteritis in Australia: the detection of Norwalk virus by immune electron
microscopy and radioirnmunoassay. J Med Virol 1980; 6: 11-19.

Gunn RA, Janowski HT, Lieb S, Prather EC, Greenberg HB. Norwalk virus
gastroenteritis following raw oyster consumption. Am J Epiderrriol 1982; 115
(3): 348-51.

Gunn RA, Terranova WA, Greenberg HB, Yashuk J, Gary GW, Wells JG,
Taylor PR, F eldman RA. Norwalk virus gastroenteritis aboard a cruise ship:

an outbreak on ﬁve consecutive cruises. Am J Epidemiol 1980; 112(6): 820-
827.

Guzewich JJ, Morse DL. Sources of shellﬁsh in outbreaks of probable viral
gastroenteritis: implications for control. J Food Prot 1986; 49(5): 389-394.

Hale AD, Lewis DC, Jiang X, Brown DWG. Homotypic and heterotypic IgG and
IgM antibody responses in adults infected with small round structured
viruses. J Med Virol 1998; 54: 305-312.

Hedberg CW, Osterholm MT. Outbreaks of food-home and waterborne viral
gastroenteritis. Clin Microbiol Rev 1993; 6 (3): 199-210.

Herwaldt BL, Lew JF, Moe CL, Lewis DC, Humphrey CD, Monroe SS, Pon
EW, Glass RI. Characterization of a variant strain of Norwalk virus from a
foodbome outbreak of gastroenteritis on a cruise ship in Hawaii. J Clin
Microbiol 1994; 32(4): 861-866.

Heun EM, Vogt RL, Hudson PJ, Parren S, Gary GW. Risk factors for
secondary transmission in households after a common-source outbreak of
Norwalk gastroenteritis. Am J Epiderrriol 1987; 126(6): 1181-1186.

Hinkula J, Ball JM, LOfgren S, Estes MK, Svensson L. Antibody prevalence

and irnmunoglobulin IgG subclass pattern to Norwalk virus in Sweden. J Med
Virol 1995; 47: 52-57.

99

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

Ho M, Monroe SS, Stine S, Cubitt D, Glass RI, Madore HP, Pinsky P, Ashley
C, Caul EO. Viral gastroenteritis aboard a cruise ship. Lancet 1989; Oct 21:
961-965.

Honma S, Nakata S, Numata K, Kogawa K, Yarnashita T, Oseto M, Jiang X,
Chiba S. Epidemiological study of prevalence of genogroup H human

calicivirus (Mexico virus) infections in Japan and Southeast Asia as determined
by enzyme-linked immunosorbent assays. J Clin Microbiol 1998; 36(9): 2481-
2484.

Hyams KC, Malone JD, Kapikian AZ, Estes MK, Jiang X, Bourgeois AL,
Paparello S, Hawkins RE, Green KY. Norwalk virus infection among Desert
Storm troops. J Infect Dis 1993; 167: 986-987.

Iverson AM, Gill M , Bartlett CLR, Cubitt WD, McSwiggan DA Two
outbreaks of foodbome gastroenteritis caused by a small round structured
virus: evidence of prolonged infectivity in a food handler. Lancet 1987 ;
ii:556-558

Jakus LA. Epidemiology and detection as options for control of viral and
parasitic foodbome disease. EH) 1997; 3(4): 529-537.

Jiang X, Wang J, Graham DY, Estes MK. Detection ofNorwalk virus in stool
by polymerase chain reaction. J Clin Microbiol 1992; 30 (10): 2529-2534.

Jiang X, Wang M, Wang K, Estes MK. Sequence and genomic organization of
Norwalk virus. Virology 1993; 195:51-61.

Jiang X, Turf E, Hu J, Barrett E, Dai XM, Monroe S, Humphrey C, Pickering
LK, Matson DO. Outbreaks of gastroenteritis in elderly nursing homes and
retirement facilities associated with human caliciviruses. J Med Virol 1996; 50:
355-341.

Jiang X, Matson DO, Velasquez FIL Calva JJ, Zhong WM, Hu J,
Ruiz-Palacios GM, Pickering LK. Study of Norwalk-related viruses in
Mexican children. J Med Virol 1995; 47:309-316.

Jiang X Wang M, Graham JY, Estes MK. Expression, self-assembly, and
antigenicity ofNorwalk virus capsid protein. J Vrrol 1992; 66: 6527-6532.

Johnson PC, Mathewson JJ, DuPont HL, Greenberg HB. Multiple-challenge

study of host susceptibility to Norwalk gastroenteritis in US adults. J Infect
Dis 1990; 161: 18-21.

100

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

Kapikian AZ, Wyatt RG, Dolin 1L Thomhill TS, Kalica AR, Chanock RM.
Visualization by immune electron microscopy of a 27-nm particle associated
with acute infectious nonbacterial gastroenteritis. J Virol 1972; 10(5):

107 5-108 1.

Kaplan JE, Feldman R, Campbell DS, Lookabaugh C, Gary GW. The
frequency of a Norwalk-like pattern of illness in outbreaks of acute
gastroenteritis. Am J Public Health 1982; 72: 1329-1332.

Kaplan JE, Gary G, Baron RC, Singh N, Schonberger LB, Feldman R,
Greenberg HB. Epidemiology of Norwalk gastroenteritis and the role of
Norwalk virus in outbreaks of acute nonbacterial gastroenteritis. Ann Int Med
1982; 96 (part 1): 56-761.

Kaplan JE, Goodman RA, Schonberger LB, Lippy EC, Gary GW. 33
Gastroenteritis due to Norwalk virus: an outbreak associated with a municipal
water system. J Infect Dis 1982; 146(2): 190-197.

Kaplan JE, Schonberger LB, Varano G, Jackrnan N, Bied J, Gary GW. An
outbreak of acute nonbacterial gastroenteritis in a nursing home. Am J
Epidenriol 1982; 116(6): 940-948.

Kappus KD, Marks J S, Holman RC, Bryant JK, Baker C, Gary GW,
Greenberg HB. An outbreak of Norwalk gastroenteritis associated with
swimming in a pool and secondary person-to-person transmission. Am J
Epidenriol 1982; 116 (5): 834-839.

Keswick BH, Satterwhite TK, Johnson PC, DuPont HL, Secor SL, Bitsura J A,
Gary GW, Hoff JC. Inactivation ofNorwalk virus in drinking water by
chlorine. Appl Environ Microbiol 1985; 50(2): 261-264.

Khan AS, Moe CL, Glass R1, Monroe SS, Estes MK, Chapman LE, Jiang X,
Humphrey C, Pon E, Iskander JK, Schonberger LB. Norwalk virus-associated
gastroenteritis traced to ice consumption aboard a cruise ship in Hawaii:

Comparison and application of molecular method-based assays. J Clin
Microbiol 1994; 32(2): 318-322.

Kilgore PE, Belay ED, Hamlin DM, Noel J S, Humphrey CD, Gary HE, Ando
T, Monroe SS, Kludt PE, Rosenthal DS, Freeman J, Glass RI. A university
outbreak of gastroenteritis due to a small round-structured virus: application
of molecular diagnostics to identify the etiologic agent and patterns of
transmission. J Infect Dis 1996; 173: 787-93.

Kirkland KB, Meriwether RA, Leiss JK, MacKenzie WR. Steaming oysters

does not prevent Norwalk-like gastroenteritis. Pub Health Rep 1996; 111:
527-530.

101

99.

100.

101.

102.

103.

104.

105.

106.

107.

108.

Koff RS, Sear HS. Internal temperature of steamed clams. N Eng J Med
1967; 276 (13): 737-739.

Kohn MA, Farley TA, Ando T, Curtis M, Wilson SA, Tm Q, Monroe SS,
Baron RC, McFarland LM, Glass RI. An outbreak ofNorwalk virus
gastroenteritis associated with eating raw oysters: implications for maintaining
safe oyster beds. JAMA 1995; 273 (6): 466-471.

Koopman J S, Eckert Ea, Greenberg HB, Strohm BC, Isaacson RE,, Monto
AS. Norwalk virus enteric illness acquired by swimming exposure. Am J
Epidenriol 1982; 115: 173-177.

Kukkula M, Maunula L, Silvennoinen E, von Bonsdorff C-H. Outbreak of
viral gastroenteritis due to drinking water contaminated by Norwalk-like
viruses. J Infect Dis 1999; 180: 1771-6.

Kuritsky JN, Osterholm MT, Korlath JA, White KE, Kaplan JE. A statewide
assessment of the role of Norwalk virus in outbreaks of food-home
gastroenteritis. (letter). J Infect Dis 1985; 151(3): 568.

Kuritsky JN, Osterhohn MT, Greenberg HB, Korlath J A, Godes JR, Hedberg
CW, Forfang JC, Kapikian AZ, McCVullough JC, White KE. Norwalk
gastroenteritis: a community outbreak associated with bakery product
consumption. Ann Int Med 1984; 100: 519-521.

Lawson HW, Braun MM, Glass RIM, Stine SE, Monroe SS, Atrash HK, Lee
LE, Englender SJ. Waterbome outbreak of Norwalk virus gastroenteritis at a
southwest US resort: role of geological formation in contamination of well
water. Lancet 1991; 337: 1200-04.

Levett PN, Gu M, Luan B, F earon M, Stubberﬁeld J, Jarnieson F, Petric M.
Longitudinal study of molecular epidemiology of small round structured
viruses in a pediatric population. J Clin Microbiol 1996; 43 (6): 1497-1501.

Lew JF, Kapikian AZ, Jiane X, Estes MK, Green K. Molecular
characterization and expression of the capsid protein of Norwalk-like virus
recovered from a Desert Shield troop with gastroenteritis. Virology 1994;
200: 319-325.

Lew JF, Kapikian AZ, Valdesuso J, Green KY. Molecular characterization of

Hawaii virus and other Norwalk-like viruses: evidence for genetic
polymorphism among human caliciviruses. J Infect Dis 1994; 170:535-42.

102

109.

110.

111.

112.

113.

114.

115.

116.

117.

118.

119.

120.

Lew JF, Valdesuso J, Vesikari T, Kapikian AZ, Jiang X, Estes MK, Green KY.
Detection of Norwalk virus or Norwalk-like virus infections in Finnish infants
and young children. J Infect Dis.

Lewis DC, Hale A, Jiang X, Eglin R, Brown DWG. Epidemiology of Mexico
virus, a small round-structured virus in Yorkshire, United Kingdom, between
January 1992 and March 1995. J Infect Dis 1997; 175:951-954.

Lieb S, Gunn RA, Medina R, Singh N, May RD, Janowski HT, Woodward
WE. Norwalk virus gastroenteritis: an outbreak associated with a cafeteria at
a college. Am J Epiderrriol 1985; 121: 259-268.

Lo SV, Connolly AM, Palmer SR, Wright D, Thomas PD, Johnson D. The
role of the pre-symptomatic food handler in a common source outbreak of
food-home SRSV gastroenteritis in a group of hospitals. Epidemiol Infect
1994; 113: 513-521.

Madore HP, Treanor JJ, Buja 1L Dolin R Antigenic relatedness among the
Norwalk-like agents by serum antibody rises. J Med Vrrol 1990; 32: 96-101.

Mack WN, Lu YS, Coohon DB. Isolation of poliomyelitis virus from a
contaminated well. Health Services Rep 1972; 87(3): 271-274.

Matson DO, Zhong WM, Nakata S, Numata K, Jiang X, Pickering LK, Chiba
S, Estes MK. Molecular characterization of a human calicivirus with sequence
relationships closer to animal caliciviruses than other known human
caliciviruses. J Med Virol 1995; 45: 215-222.

McDonnell S, Kirkland KB, Hlady G, Aristegieta C, Hopkins RS, Monroe SS,
Glass RI. Failure of cooking to prevent shellﬁsh-associated viral
gastroenteritis. Arch Int Med 1997; 157: 111-116.

Mead PS, Slutsker L, Dietz V, MeCraig LF, Bresee J S, Shapiro C, Grifﬁn PM,
Tauxe RV. F cod-related illness and death in the United States. EH) 1999;
(5)5: 607-625.

Metcalf TG. Indication of viruses in shellﬁsh growing waters. Am J Pub
Health 1979; 69(11): 1093-1094.

Michigan Department of Community Health Bureau of Infectious Diseases.

Polymerase chain reaction for detection of Norwalk-like viruses.
Procedure #MB-S, 1999.

Michigan Department of Public Health. A cluster of outbreaks of Norwalk

virus-related illness at an environmental education camp - West Michigan,
Winter 1993/ 1994.

103

121.

122.

123.

124.

125.

126.

127.

128.

129.

130.

Moe CL, Gentsch J, Ando T, Grohmann G, Monroe SS, Jiang X, Wang J,
Estes MK, Seto Y, Humphrey C, Stine S, Glass RI. Application of PCR to
detect Norwalk virus in fecal specimens ﬁ'om outbreaks of gastroenteritis. J
Clin Microbiol 1994; 32(3): 642-648.

Monroe SS, Stine SE, Jiang X, Estes MK, Glass RI. Detection of antibody to
recombinant Norwalk virus antigen in specimens from outbreaks of
gastroenteritis. J Clin Microbiol 1993; 31 (11): 2866-2872.

Morens DM, Zweighaft RM, Vernon TM, Gary GW, Eslien JJ, Wood BT,
Hohnan RC, Dolin R. A waterborne outbreak of gastroenteritis with
secondary person-to-person spread. Lancet 1979; i:964-966.

Morse DL, Guzewich JJ, Hanrahan JP, Stricof R, Shayegani M, Deibel R
Grabau JC, Nowak NA, Herrmann JE, Cukor G, Blacklow NR Widespread
outbreaks of clarn- and oyster-associated gastroenteritis. N Eng J Med 1986;
314(11): 678-81.

Nakata S, Chiba S, Terashirna H, Yokoyarna T, Nakao T. Humoral immunity
in infants with gastroenteritis caused by human calicivirus. J Infect Dis
1985;152(2):274-9.

Nakata S, Honma S, Numata K, Kogawa S, Ukae S, Adachi N, Jiang X, Estes
MK, Gatheru Z, Tukei PM, Chiba S. Prevalence of human calicivirus
infections in Kenya as determined by enzyme irnmunoassays for three
genogroups of the virus. J Clin Microbiol 1998; 36(11): 3160-3163.

Noel J S, Fankhauser RL, Ando T, Monroe SS, Glass RI. Identiﬁcation of a
distinct common strain of “Norwalk-like viruses” having a global distribution.
J Infect Dis 1999; 179(6):]334-1344.

Noel J S, Ando T, Leite JP, Green KY, Dingle KE, Estes MK, Seto Y, Monroe
SS, Glass RI. Correlation of patient immune responses with genetically
characterized small round-structured viruses involved in outbreaks of
nonbacterial acute gastroenteritis in the United States, 1990-1995. J Med
Vrrol 1997; 53: 372-383.

Numata K, Nakata S, Trang X, Estes MK, Chiba S. Epidemiological study of
Norwalk virus infections in Japan and Southeast Asia by enzyme-linked
immunosorbent assays with Norwalk virus capsid protein produced by the
baculovirus expression system. J Clin Microbiol 1994; 32(1): 121-126.

Okhuysen PC, Jiang X, Ye L, Johnson PC, Estes MK. Vrra] shedding and
fecal IgA response after Norwalk virus infection. J Infect Dis 1995; 171: 566-9.

104

131.

132.

133.

134.

135.

136.

137.

138.

139.

140.

O’Ryan ML, Vial P, Mamani N, Jiang X, Estes MK, Ferrecio C, Lakkis H,
Matson DO. Seroprevalence of Norwalk virus and Mexico virus in Chilean
individuals: Assessment of independent risk factors for antibody acquisition.
Clin InfDis 1998; 27: 789-95.

Oyofo BA, Soderquist R Lesmana M, Subekti D, Tjaniadi P, Fryauﬁ’ DJ,
Corwin AL, Richie E, Lebron C. Norwalk-like virus and bacterial pathogens
associated with cases of gastroenteritis onboard a US Navy ship. Am J Trop
Med Hyg 1999; 61(6): 904-908.

Parashar UD, Kow L, Fankhauser RL, Humphrey CD, Miller J, Ando T,
Williams KS, Eddy CR, Noel J S, Ingram T, Bresee J S, Monroe SS, Glass RI.
An outbreak of viral gastroenteritis associated with consumption of

sandwiches: implications for the control of transmission by food handlers.
Epidemic] Infect 1998; 121: 615-621.

Parker SP and Cubitt WD. Measurement of IgA responses following Norwalk
infection and other human caliciviruses using a recombinant Norwalk virus
protein EIA Epidenriol Infect 1994; 113: 143-152.

Parker SP, Cubitt WD, Jiang X. Enzyme immunoassay using
baculovirus-expressed human calicivirus (Mexico) for the measurement of IgG
responses and determining its seroprevalence in London, UK. J Med Virol
1995; 46: 194-200.

Parker SP, Cubitt WD, Jiang XD, Estes MK. Seroprevalence studies using a
recombinant Norwalk virus protein enzyme immunoassay. J Med Vrrol
1994;42: 146-150.

Parks CG, Moe CL, Rhodes D, Lima A, Barrett L, Tseng F, Baric R, Talal A,
Guerrant R Genomic diversity of “Norwalk-like viruses” (NLVS): pediatric
infections in a Brazilian shantytown. J Med Virol 1999; 58: 426-434.

Parrino TA, Schreiber DS, Trier J S, Kapikian AZ, Blacklow NR. Clinical
immunity in acute gastroenteritis caused by Norwalk agent. N Eng J Med
1977; 297(2): 86-89.

Patterson T, Hutchings P, Palmer S. Outbreak of SRSV gastroenteritis at an
international conference traced to food handled by a post-symptomatic caterer.
Epiderrriol Infect 1993; 111: 157-162.

Pelosi E, Larnbden PR, Caul EO, Liu B, Dingle K, Deng Y, Clarke IN. The
seroepidemiology of genogroup 1 and genogroup 2 Norwalk-like viruses in
Italy. J Med Virol 1999; 58: 93-99.

105

141.

142.

143.

144.

145.

146.

147.

148.

149.

150.

151.

Reid JA, White DG, Caul EO, Palmer SR. Role of infected food handler in
hotel outbreak of Norwalk-like viral gastroenteritis: implications for control.
Lancet 1988; Aug 6: 321-323.

Ryder RW, Singh N, Reeves WC, Kapikian AZ, Greenberg HB, Sack RB.
Evidence of immunity induced by naturally acquired rotavirus and Norwalk
virus infection on two remote Panamanian islands. J InfDisl985; 151: 99-105.

Richards GP. Outbreaks of shellﬁsh-associated enteric virus illness in the
United States: requisite for development of viral guidelines. J Food Prot
1985; 48(9): 815-823.

Richards GP. Limitation of molecular biological techniques for assessing the
virological safety of foods. J Food Prot 1999; 62(6): 691-697.

Rosenﬁeld SR J aykus L-A. A multiplex reverse transcription polymerase
chain reaction method for the detection of foodbome viruses. J Food Prot
1999; 62(10): 1210-1214.

Sawyer LA, Murphy JJ, Kaplan JE, Pinsky PF, Chacon D, Walmsley S,
Schonberger LB, Phillips A, Forward K, Goldman C, Brunton J, Fralick RA,
Carter AO, Gary WG Jr., Glass RI, Low DE. 25-30 nm virus particle
associated with a hospital outbreak of acute gastroenteritis with evidence for
airborne transmission. Am J Epidemiol 1988; 127(6): 1261-71.

Schreiber DS, Blacklow NR, Trier J S. The small intestinal lesion induced by
Hawaii agent acute infectious nonbacterial gastroenteritis. J Infect Dis 1974;
129 (6):705-708. -

Sharp TW, Hyams KC, Watts D, Trofa A, Martin GJ, Kapikian AZ, Green
KY, Jiang S, Estes MK, Waack M, Savarino SJ. Epidemiology of Norwalk
virus during an outbreak of acute gastroenteritis aboard a US aircraft carrier.
J Med Virol 1995; 45:61-67.

Slomka MJ, Appleton H. Feline calicivirus as a model system for heat
inactivation studies of small round structured viruses in shellﬁsh. Epidemic]
Infect 1998; 121: 401-407.

Small round structured viruses (SRSV): numbers are increasing. CDR Weekly
1996; 5(23): 1.

Srrrit TK, Steele AD, Peenze I, Jiang X, Estes MK. Study ofNorwalk virus
and Mexico virus infections at Ga-Rankua Hospital, Ga-Rankua, South Aﬁica.
J Clin Microbiol 1997; 35 (9): 2381-2385.

106

152.

153.

154.

155.

156.

157.

158.

159.

160.

161.

162.

Smith AW, Skilling DE, Cherry N, Mead JH, Matson DO. Calicivirus
emergence from ocean reservoirs: zoonotic and interspecies movements. EID
1998; 4(1): 13-20.

SRSV genera] outbreaks: England and Wales, by place of outbreak,
1992-1998. On the internet at: www.phls.cc.uk/facts/srsv-t04.htm .

Steffen R Collard F, Tomieporth N, Campbell-Forrester S, Ashley D,
Thompson S, Mathewson JJ, Macs E, Stephenson B, DuPont HL, von
Sonnenberg F. Epidemiology, etiology, and impact of travelers’ diarrhea in
Jamaica. JAMA 1999; 281(9): 811-817.

Sugeida M, Nakajirna K, Nakajirna S. Outbreaks of Norwalk-like
virus-associated gastroenteritis traced to shellﬁsh: coexistence of two
genotypes in one specimen. Epidenricl Infect 1996; 116: 339-346.

Surveillance of outbreaks of small round structured virus infection. CDR
Weekly 1996; 6(10):1.

Taylor JW, Gary GW J r., Greenberg HB. Norwalk-related viral gastroenteritis
due to contaminated drinking water. Am J Epidemiol 1981; 114(4): 584-592.

Taylor MB, Parker S, Grabow WOK, Cubitt WD. An epidenriolcgical
investigation of Norwalk virus infection in South Aﬁica. Epidemic] Infect
1996; 116: 203-206.

Thomhill TS, Kalica AR, Wyatt RG, Kapikian AZ, Chanock RM. Pattern of
shedding of the Norwalk particle in stools during experimentally induced
gastroenteritis in volunteers as determined by immune electron microscopy. J
Infect Dis 1975; 132(1): 28-34.

Treanor JJ, Irang X, Madore HP, Estes MK. Subclass-speciﬁc serum antibody
responses to recombinant Norwalk virus capsid antigen (rNV) in adults
infected with Norwalk, Snow Mountain, or Hawaii virus. J Clin Microbiol
1993; 31(6): 1630-1634.

Truman BI, Mador HP, Menegus MA, Nitzkin JL, Dolin R. Snow Mountain
agent gastroenteritis from clams. Am J Epidenricl 1987; 126(3): 516-525.

US. Public Health Service, Food and Drug Administration. Model Food
Code, 1999. Washington DC. ISBN 0-16-050028-1.

107

163.

164.

165.

166.

167.

168.

169.

170.

171.

172.

173.

U. S. Public Health Service, Food and Drug Administration. Procedures for the
safe and sanitary processing and importing of ﬁsh and ﬁshery products, ﬁnal
rule. Federal Register, 1995; 60(242): 65095-65202.

Van der Poel WHM, Vinje J, van der Heide R, Herrera M-I, Vivo A,
Koopmans MPG. Norwalk-like calicivirus genes in farm animals. EID 2000;
6(1):36-41.

Vinje J and Koopmans MPG. Molecular detection and epidemiology of small
round-structured viruses in outbreaks of gastroenteritis in the Netherlands. J
Infect Dis 1996; 174:610-5.

ije J, Altena SA, Koopmans MPG. The incidence and genetic variability of
small round-structured viruses in outbreaks of gastroenteritis in the
Netherlands. J Infect Dis 1997; 176: 1374-8.

Wallace BJ, Guzewich JJ, Cambridge M, Altekruse S, Morse DL.
Seafood-associated disease outbreaks in New York, 1980-1994. Am J Prev
Med 1999; 17(1): 48-54.

Warner RD, Carr RW, McCleskey FK, Johnson PC, Elmer LMG, Davison VE.
A large nontypica] outbreak of Norwalk virus: gastroenteritis associated with
exposing celery to nonpctable water and with Citrobacterﬁeundii. Arch
Intern Med 1991; 151: 2419-2424..

White KE, Osterholm MT, Mariotti J A, Korlath JA, Lawrence DH, Ristinen
TL, Greenberg HB. A foodbome outbreak of Norwalk virus gastroenteritis.
Am J Epidemiol 1986; 124(1): 120-126.

Yang S. F oodNet and Enter-net: Emerging surveillance programs for
foodbome diseases. EH) 1998; 4(3): 457-458.

Centers for Disease Control. Surveillance for waterborne disease
outbreaks-49914992. MMWR 1993; 42(SS-5): 1.

Centers for Disease Control. Surveillance for waterborne disease outbreaks--
1993-1994. MMWR 1996; 45(SS-1): 1.

Centers for Disease Control. Surveillance for waterborne disease outbreaks--
1995-1996; 47(SS-5): 1-7.

108

INDEX

A

Aquifer ...................... 63
Asymptomatic ..... 6, 8, 9, 54, 55, 57,

58,60,6l,72,81

Attack rate . . 46, 48, 49, 65, 66, 85, 86
B

Biotin-avidin immunoassay . 15, 17, 29
C

CaliciNet ..................... 79
Caliciviridae ................. 3, 25
Collecting specimens ............ 14
Cross connection ............... 62
D

Depuration ........... 23, 42, 43, 52
Detection methods .............. 57
Disinfection ................ 66, 85
Duration ..... 4, 58, 74, 81, 85, 86, 93
E

Eccnonric impact ............... 83
ELISA ...... 18-20, 24, 32, 58, 84, 99
Enzyme immunoassay ........... 31
Ethnicity ..................... 84
F

Food handlers. . 40, 53-57, 59-61, 67, 92
Food specimens ................ 22
Foodbome Disease

Surveillance System ............. 78
G

109

Genogroup .............. 25, 27-29
H
Histopathology ................ 95
I
ICD-9-CM ................... 79
Immune electron microscopy 13, 14, 57
Irnplicated food items ........... 55
Inactivation ............. 23, 48, 49
Incidence ............. 36, 82, 85, 90
Incubation period ....... 56, 80, 81, 93
Infectious dose ...... 3, 45, 52, 57, 69
L
Lordsdale virus ........ 20, 25, 32, 37
M
Maternal antibody ............ 34-36
Mexico virus ............ 20, 25, 32
N
Nomenclature ................ 25
O
Outbreaks
Aerosol ....... 68, 70, 73, 75
Control ............... 60
Droplets ............... 70
Foodbome. . . . 1, 27, 39-41,59,
68, 69, 77-80, 82, 87, 90,
92, 93

Investigation. . . . 19, 23, 48, 56,

62,65,71,73,77,78,93
Oysters ..... 22, 27, 40-52, 68
Person-to-person
transmission. . . . 68, 69, 71, 92
Waterbome. .1, 13, 61, 80, 83, 92

Oysters

HACCP ................ 52
Steamed ............. 46, 47
Tagging ............. 50, 52

P
Pasteurization ................. 91
Electron beam ........... 91
Irradiation .............. 91
Ultra-high pressure ........ 91
Phylogenetic analysis ............ 25
Porcine ...................... 29
Prevalence .................. 30-37
Productivity ................... 83

R
Radioirnmunoassay .......... 15, 30
RIA .................... 15-18, 30
Risk factors ................. 88-90
RT-PCR ....... 20, 22, 23, 25, 27, 28,
51,52,58,68,69,81

S
San Martin Sea Lion Virus ........ 46
Sanitation ............ 32, 42, 44, 52
Seasonal prevalence ............. 36
Seasonality ................... 71
Secondary spread ........... 70, 73
Serological response ............. 8
Seroprevalence . . 30, 32, 34-36, 38, 88
Ships ................ 25, 70, 71, 85
Small round structured viruses . . . 3, 79
Southampton virus ........... 20, 25
Surveillance ..... 1, 28, 52, 59, 61, 67,
69, 77-80, 82, 92
Symptoms ..... 4, 6, 15, 16, 53-57, 60,

110

78,79,86,88
T
Temporal clustering ............. 71
Thermal inactivation ............ 48
Traceback .................... 65
V
Vaccine ................ 11, 64, 9]
Vrral shedding .......... 6, 54, 55, 58
Volunteer. . . .10-13, 30, 37, 38, 54, 57,
58, 66
Vomitus .................. 69, 70
W
Water ....... 23, 39, 40, 42-45, 49,
61-68,70, 80, 85, 89, 90,
92

lljllllill?ill/lllllllllllll

293 02092 8929