Content uploaded by Barry Vipond
Author content
All content in this area was uploaded by Barry Vipond on Jun 09, 2016
Content may be subject to copyright.
A school outbreak of Norwalk-like virus: evidence for
airborne transmission
P. J. M ARKS
1
*, I. B. V I P OND
2
,F.M.REGAN
3
,K.WEDGWOOD
4
, R. E. FEY
4
AND E. O. C AUL
2
1
Division of Epidemiology and Public Health, University of Nottingham, Queens Medical Centre,
Nottingham NG7 2UH
2
Bristol Public Health Laboratory, Myrtle Road, Kingsdown, Bristol B52 8EL
3
Environmental Services Division, Derby City Council, Celtic House, Heritage Gate, Friary Street,
Derby DE1 1QK
4
Southern Derbyshire Health Authority, Derment Court, Stuart St, Derby DE1 2FZ
(Accepted 20 March 2003)
SUMMARY
An outbreak of gastroenteritis affected a school attended by children aged 4–11 years.
Epidemiological features suggested this was due to Norwalk-like virus (NLV) and this was
confirmed by polymerase chain reaction (PCR). Nucleotide sequence analysis of the PCR
amplicons revealed identical strains in all five positive stool samples. Pupils were significantly
more likely to become ill following an episode of vomiting within their classroom
(adjusted odds ratio 4.1, 95 % CI 1.8–9.3). The times from exposure to illness were consistent
with direct infection from aerosolized viral particles where exposure to vomiting was high.
Cleaning with quaternary ammonium preparations made no impact on the course of the
outbreak. However, the outbreak stopped after the school closed for 4 days and was cleaned
using chlorine-based agents. This study confirms the importance of vomiting in the transmission
of NLV and provides evidence that direct infection with aerosolized viral particles occurs.
INTRODUCTION
Norwalk-like viruses (NLVs) are a genetically diverse
group of highly infectious RNA viruses of the family
Caliciviridae [1] which were first reported following
an outbreak of gastroenteritis in Norwalk, Ohio in
1972 [2]. NLVs have the ability to cause outbreaks of
gastrointestinal infection characterized by vomiting,
which is often of sudden onset and projectile, and di-
arrhoea [3, 4]. NLV infection is the most important
cause of non-bacterial gastroenteritis worldwide [5].
The Infectious Intestinal Disease Study in England [6]
reported an incidence rate of 12.5 per 1000 patient
years in the community. However, that study did not
use polymerase chain reaction (PCR) techniques
to detect NLV and will therefore have significantly
underestimated the true incidence.
Outbreaks due to contaminated food [7], particu-
larly oysters [8–10], and water [11–13] have been
described. Faecal-oral spread of NLV infection is
important, but vomitus also represents a major source
of infection [14]. The production of viral aerosols and
airborne transmission following vomiting have been
suggested [15–17]. Widespread environmental con-
tamination may occur [18] and widespread cleaning
has been advocated [14]. However, the importance
of environmental cleaning in controlling the trans-
mission of NLV has not been proven.
In this report we describe an outbreak of NLV
gastroenteritis in a primary school during which
* Author for correspondence: Division of Epidemiology and
Public Health, D Floor, Medical School, Queens Medical Centre,
Nottingham NG7 2UH, UK.
Epidemiol. Infect. (2003), 131, 727–736. f2003 Cambridge University Press
DOI : 10.1017/S0950268803008689 Printed in the United Kingdom
vomiting occurred in some, but not all classrooms.
This resulted in a ‘ natural experiment ’ which enabled
us to investigate the importance of vomiting as a
mode of transmission of NLV, and the likelihood
that environmental contamination played a role in
the spread of the outbreak.
THE OUTBREAK
The outbreak occurred in a primary school and nur-
sery attended by children aged 4–11 years. The 15
classrooms were situated in two buildings : one for
younger children aged 4–6 years and the other for 7–11
year olds. Children did not move between classrooms
for different lessons. Children in the nursery attended
for either the morning or afternoon. Morning and
afternoon groups were taught in the same room. All
children, whether eating school meals or lunches pre-
pared at home ate in the same dining room. The school
roll at the onset of the outbreak was 492.
The initial case was first absent from school on 25
June 2001 (day 1) and the local Environmental Health
Department was notified on day 11 that a number of
pupils had vomiting and diarrhoea. In total 186 pupils
had some absence from school with gastrointestinal
symptoms. Five members of staff were also ill. The
onset of vomiting was often sudden with a number
of children vomiting within classrooms. Vomiting
also occurred in corridors and lavatories, but not in
the dining room. The areas visibly contaminated by
vomitus were cleaned immediately.
Extensive environmental cleaning of the school
took place on days 13 and 14. Despite advice about
their potential lack of efficacy, concerns about the
health and safety implications of using chlorine-
releasing agents meant that a quaternary ammonium
compound was used for this cleaning. Cleaning
took place again on days 19 and 20, this time using
chlorine-based products. The school closed from
days 18–21 inclusive. After the second cleaning oper-
ation and closure no further school absences occurred,
although three pupils reported symptoms on day 22.
METHODS
Epidemiological investigation
Lists of pupils attending each class and sickness ab-
sence records were supplied by the school. The number
of pupils absent because of gastrointestinal symptoms
compatible with NLV infection (diarrhoea, vomiting
or abdominal pain) was recorded. A questionnaire
was designed asking about the pupil’s date of birth,
whether they had vomiting or diarrhoea, the date
of onset and cessation of symptoms, the number of
adults and children residing in the household, how
many of them had been ill and the dates of their ill-
nesses. The parents or guardians of each pupil were
asked to complete this on day 22 of the outbreak,
and return them to the pupil’s class teacher. Stool
samples were requested from 15 pupils who had
symptoms.
Cases were defined as follows :
$for those pupils who returned a questionnaire :
those who reported either diarrhoea or vomiting or
both from 25 June to 16 July 2001 inclusive,
$for those pupils who did not return a questionnaire:
those who were absent from school with symptoms
compatible with NLV infection from 25 June to 16
July 2001 inclusive.
Secondary cases were defined as other household
members reporting diarrhoea or vomiting on the
questionnaire after a pupil had been ill.
Data from the questionnaire and school absence
records were analysed using Microsoft Excel 97 and
SPSS for Windows, version 9. Attack rates were cal-
culated by sex, class and age group. Classes were
grouped according to the number of vomiting episodes
which occurred within the classroom and attack rates
calculated for each group. Chi square test for trend
was undertaken on these data. Logistic regression was
used to calculate odds ratios (OR) with exact 95 %
confidence intervals (CI) and two-sided Pvalues.
Odds ratios were adjusted for the child’s age and sex,
and the school building in which the child’s classroom
was located.
Comparisons were also made between those class-
rooms where a pupil vomited less than 24 h after the
first case becoming ill in that class and those classes
where nobody vomited in the classroom. Logistic re-
gression was used to calculate odds ratios with exact
95% CI and two-sided Pvalues. Odds ratios were
again adjusted as outlined above.
The date of onset of illness was taken to be the first
recorded day on which a pupil was absent from school.
For those pupils who were not absent the onset date
given in the questionnaire was used. Mean and median
times from exposure to illness were calculated for
those classes where episodes of vomiting within the
classroom occurred on one day only. Medians were
compared using the Mann–Whitney Utest.
728 P. J. Marks and others
Secondary household attack rates were calculated
from the questionnaire data.
Laboratory investigation
Faecal specimens were sent to Bristol Public Health
Laboratory and analysed by solid phase immune
electron microscopy (SPIEM), an in-house antigen
capture enzyme immuno assay (EIA) [19] specific
for Lordsdale strain, and reverse transcription-
polymerase chain reaction (RT–PCR) using broadly
reactive inosine containing primers targeting a
region of the polymerase gene [15]. Further PCR
investigations used alternative combinations of
broadly reactive PCR primers (YGDD [20] and Ni
[21]).
Nucleotide sequencing of purified amplicons was
carried out using the Beckman Dye Termination
Cycle Sequencing Kit and analysed on a CEQ2000XL
DNA Analysis System running software version 4.3.9.
PCR products were sequenced in both directions
using the same primers used to amplify the DNA.
Sequences were edited using the BioEdit [22] soft-
ware package and alignments and phylogenetic trees
were generated using the ClustalX [23] and TreeView
[24] software packages.
Table 1. Summary of school absences and reports of illness on
questionnaires
No diarrhoea
or vomiting
reported on
questionnaire
Diarrhoea
or vomiting
or both
reported on
questionnaire
No response to
questionnaire Total
Absent from school with
diarrhoea, vomiting
or abdominal pain
48 60 78 186
Not absent from school
with diarrhoea, vomiting
or abdominal pain
166 15 125 306
Total 214 75 203 492
Cells shaded grey indicate those pupils classified as ‘ill’ by the case definition used.
50
45
40
35
30
25
20
15
10
5
0
1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526
1st day absence
1st day reporte
d
ill
1st clean 2nd clean
& closure
Number of pupils with first day of absence from school and first
day of reported symptoms
Day of outbreak
Fig. 1. Epidemic curve for whole school.
Airborne transmission of NLV infection 729
RESULTS
Completed questionnaires were returned for 289
pupils (response rate 59 %). In total 153 pupils met
our case definition (Table 1), giving an attack rate of
31%. The epidemic curve for the outbreak is shown in
Figure 1. The mean duration of illness was 2.3 days
and the median 2 days (range <1 to 13 days).
The attack rates by age (Table 2) show an inverted
‘U’ shaped pattern, with the highest rate in children
aged between 6 and 7 years old.
Exposure to vomiting in classroom
Fifteen children vomited within ten classrooms.
Table 3 shows attack rates for classes grouped by the
number of episodes of vomiting occurring within the
relevant classrooms. There was a significant trend,
with attack rates increasing with the number of
vomiting episodes to which pupils were exposed ( x
2
for linear trend=37.8, P<0.00001). The results of
logistic regression analyses are also shown in Table 3.
These demonstrate increased odds ratios when vomit-
ing occurred within the classroom, with an adjusted
odds ratio of 14.6 for those classrooms where three
episodes of vomiting occurred.
In four classrooms (classrooms 2, 7, 9 and 10 – Figs
2–5) the vomiting occurred in the classroom less than
24 h after the first case in that class. In one of these
classes (class 9 – Fig. 4) a second episode of vomiting
occurred 9 days later. These four classrooms were
taken as those in which the children in the class were
exposed to another child vomiting. The unadjusted
odds ratio for having been exposed to another child
vomiting in the classroom was 3.9 (Table 4). When
adjusted for the child’s age and sex, and the building
in which the classroom was situated these odds ratios
remained similar at 4.1.
In three classrooms vomiting episodes occurred
on one day only. In two of these (classrooms 2 and
7 – Figs 2, 3) only one child vomited. In the third
(classroom 10 – Fig. 5) three episodes of vomiting
occurred on the same day. In classroom 10 the median
time from exposure to onset of illness was 1.0 day and
the mean 1.5 days (standard deviation (S.D.) 1.1 days).
In classroom 2 the median time was 14.0 days (mean
11.1, S.D.5
.4 days) and in classroom 7 it was 15.0 days
(mean 12.6, S.D.3
.3 days). The median time from
exposure to onset of illness was significantly shorter in
the class where three pupils vomited on the same day
than in the other two classrooms where vomiting
occurred only once (1.0vs.14
.0 days; Mann–Whitney
Utest=17.0, P<0.001).
Household secondary attack rates
Two hundred and fifty-six people (143 adults and 113
other children) were reported to be living in the
households of the 75 children who reported diarrhoea
and/or vomiting on the questionnaire. Of these, 24
adults (attack rate 17 %) and 52 children (attack rate
46%) became ill after the initial case in the household,
giving an overall attack rate of 30 %. Each ill school
child produced, on average, 1.01 household secondary
cases.
Laboratory investigations
Seven faecal specimens were analysed in the labora-
tory for viral agents of gastroenteritis. One specimen
was reactive in the EIA. The initial RT–PCR using
inosine containing primers targeting the region of the
Table 2. Attack rates by sex, age and class
Attack rate ( %) 95 % CI (%)
Sex
Male 79/260 (30.4) 25.1–36.2
Female 72/230 (31.3) 25.7–37.6
Age group
3to<4 yr 6/30 (20.0) 9.5–37.3
4to<5 yr 15/58 (25.9) 16.3–38.4
5to<6 yr 26/58 (44.8) 32.7–57.5
6to<7 yr 23/44 (52.3) 37.9–66.2
7to<8 yr 23/59 (39.0) 27.6–51.7
8to<9 yr 17/60 (28.3) 18.5–40.8
9to<10 yr 20/74 (27.0) 18.2–38.1
10 to <11 yr 14/63 (22.2) 13.7–33.9
11 to <12 yr 7/42 (16.7) 8.3–30.6
Class
1 19/86 (22.1) 14.6–31.9
2 15/36 (41.7) 27.1–57.8
3 6/25 (24.0) 11.5–43.4
4 14/29 (48.3) 31.4–65.6
5 5/32 (15.6) 6.9–31.8
6 3/27 (11.1) 3.9–28.1
7 8/28 (28.6) 15.3–47.1
8 6/33 (18.2) 8.6–34.4
9 13/36 (36.1) 22.5–52.4
10 17/24 (70.8) 50.8–85.1
11 12/28 (42.9) 26.5–60.9
12 10/29 (34.5) 19.96–52.7
13 3/30 (10.0) 3.5–25.6
14 10/29 (34.5) 19.9–52.7
15 12/20 (60.0) 38.7–78.1
730 P. J. Marks and others
polymerase gene were all negative. The use of alterna-
tive combinations of broadly reactive PCR primers
(YGDD and Ni) led to detection of viral RNA in
five of the clinical samples (including the EIA posi-
tive specimen). Nucleotide sequence analysis revealed
a single strain of NLV with identical nucleotide
sequence amongst the cases within the outbreak.
Phylogenetic analysis revealed that this strain was
closely related to viruses in the Melksham virus clus-
ter (Fig. 6).
DISCUSSION
This study provides convincing evidence that vomi-
ting is important in the transmission of NLV infec-
tion. However, there are certain limitations. Two
sources of case ascertainment were used : school sick-
ness absence records and a questionnaire to parents.
Both sources have shortcomings. The questionnaire
responses may be affected by recall bias, particularly
with regard to dates of illness, and non-responders
who have been symptomatic will not be recorded.
School sickness absences will fail to record those
children who were ill at weekends and those who were
symptomatic but did not take time off school. It
would appear that some children were absent from
school with reported symptoms compatible with NLV
infection, but reported no vomiting or diarrhoea
when completing the questionnaires. Thus, there were
some discrepancies between the two sources of infor-
mation. This would suggest that either some parents
kept their children off school at the height of the
outbreak as a precaution, or that they failed to recall
the illness correctly on the questionnaire.
The clinical features of the illness suggested NLV
infection and the outbreak met three of the four epi-
demiological criteria proposed by Kaplan et al. [25],
namely that all stools submitted were negative for
bacterial pathogens, the mean or median duration of
illness was 12–60 h and that vomiting occurred in at
least 50% of cases. Data to meet the fourth criterion,
that the mean or median incubation period was
24–48 h could not be obtained.
The clinical and epidemiological features of this
outbreak were supported by the identification of NLV
by PCR in 5 of 7 stool samples analysed. The presence
of a single strain of virus confirmed a common source
of infection. The failure to detect the virus with a
standard set of PCR primers indicated that the viral
genome had mutated at the primer binding site such
that the primer could no longer bind. The Lordsdale
specific EIA detected a virus in a single specimen,
showing that the outbreak strain is antigenically
related to Lordsdale virus. However, phylogenetic
Table 3. Attack rates and odds ratios by number of vomiting episodes
within classrooms
Number of
vomiting
episodes
within
classroom Attack rate (%)
Unadjusted
OR (95 % CI) P
Adjusted OR
(95% CI)#P
0* 23/147 (15.6) 1.0<0.001 1.0<0.001
1 78/236 (33.1) 2.7(1
.6–4.5) <0.001 5.1(2
.2–11.6) <0.001
2 23/65 (35.4) 3.0(1
.5–5.8) 0.002 3.9(1
.8–8.6) 0.001
3 29/44 (65.9) 10.4(4
.8–22.4) <0.001 14.6(5
.9–36.5) <0.001
* Baseline variable.
#Adjusted for sex, age and building in which classroom was situated.
Table 4. Attack rates and odds ratios for exposure to another child
vomiting in classroom
Attack rate (%)
Unadjusted
OR (95% CI) P
Adjusted OR
(95% CI)* P
Exposed 53/125 (42.4) 3.9(2
.2–7.0) <0.001 4.1(1
.8–9.3) 0.001
Not exposed 23/146 (15.8)
* Adjusted for sex, age and building in which classroom was situated.
Airborne transmission of NLV infection 731
analysis revealed that the outbreak strain was more
closely related to viruses in the Melksham virus cluster,
rather than Lordsdale-like strains (Fig. 6). Un-
fortunately, the primers used in this PCR generate a
smaller amplicon than the standard PCR, and so
contain less sequence information. The phylogenetic
trees based upon this smaller region were less reliable in
accurately describing genetic relationships of known
strains from the genetic database. Consequently, in-
terpretation of the phylogenetic tree should be treated
with caution.
Attack rates in males and females were similar,
but varied by age with a peak in children aged 6–7
years. Younger and older children had lower attack
rates. It may be that the attack rates were lower in
younger children because they were more intensively
supervised when hand washing, and in older chil-
dren because they have developed better standards
of hygiene.
In the classrooms where vomiting occurred, pupils
were significantly more likely to be ill than pupils in
classrooms where no vomiting occurred. There was
10
9
8
7
6
5
4
3
2
1
01234567891011121314151617181920
Day of outbreak
Vomiting episode
Number of pupils ill for first day
Fig. 3. Epidemic curve for class 7.
10
9
8
7
6
5
4
3
2
1
01234567891011121314151617181920
Day of outbreak
Vomiting episode
Number of pupils ill for first day
Fig. 2. Epidemic curve for class 2.
732 P. J. Marks and others
a highly significant trend. Significantly high odds
ratios persisted after adjustment for possible con-
founding factors. However, these analyses take no
account of temporal relationships. It may be argued
that increased numbers of ill children in a class make
it more likely that someone will vomit within the
classroom, rather than exposure to vomiting causing
the high attack rates.
In an attempt to overcome the issue of the time
sequence, comparisons were made between those
classrooms where vomiting occurred within 24 h of
the first case in that class and those where no vomit-
ing at all occurred during class. The significantly high
odds ratios obtained make it clear that the illnesses
followed the exposure. Taken together these analyses
support the hypothesis that exposure to vomiting is
a significant risk factor for development of infection
with NLV. As each area where a child vomited was
cleaned immediately after vomiting had occurred,
spread must have been by airborne transmission di-
rectly to susceptible individuals, or via contamination
of the wider environment.
12
8
6
4
2
01234567891011121314151617181920
Day of outbreak
Number of pupils ill for first day
21
3 vomiting
episodes
10
Fig. 5. Epidemic curve for class 10.
10
9
8
7
6
5
4
3
2
1
01234567891011121314151617181920
Day of outbreak
Vomiting episode
Number of pupils ill for first day
21
Vomiting episode
Fig. 4. Epidemic curve for class 9.
Airborne transmission of NLV infection 733
In one classroom where three episodes of vomiting
occurred on the same day, the median time from ex-
posure to illness was consistent with a point source of
infection and suggests that aerosolized viral particles
were inhaled and subsequently swallowed. This is a
similar pattern to that previously described when an
episode of vomiting occurred during a meal [15].
However, in the current outbreak, contamination of
food cannot be implicated. In addition, any increased
risk of eating food in the presence of aerosolized virus
can be excluded.
The incubation period for NLV is widely accepted
as being 24–48 h from exposure [4]. Therefore cases
in the two classrooms with extended times from
exposure to onset could not be due to direct infec-
tion. Transmission must have occurred by person-
to-person spread or through exposure to a contami-
nated environment. Given that opportunities for
person-to-person spread are likely to be similar in
children within the same age group and that the dif-
ferences in attack rates at different levels of exposure
persist after adjustment for age, sex and the building
in which the child’s classroom was situated, it is poss-
ible that environmental contamination could have ac-
counted for the increased attack rates in classrooms
where vomiting had occurred. As NLV cannot be
cultured, the length of time that it can survive in the
environment is difficult to assess. However, a closely
related cultivable virus, feline calicivirus, has been
shown to survive for between 21 and 28 days in a
dried state at room temperature [26]. If the survival
of NLV is similar, this could account for the extended
time between a vomiting incident and onset of illness
in some classrooms.
During the outbreak widespread environmental
cleaning was undertaken on two occasions. Because of
Bovine
DSV395
DSV275
SV
KY89
NV
984
995
903
689
816
325
286
689
996
962
515
HV
SMA
LV
BV
MRV
CV
MV
625
1000
Case 1
Case 5
Case 3
Case 4
Case 2
0·1
Fig. 6. Dendrogram showing the genetic relationships of the polymerase region amplicons of NLV strains identified among
pupils and staff in this outbreak, with the equivalent region of known strains from the GenBank database. Accession numbers
for strains include: SV: Southampton (L07418) ; KY89 : KY-89/89/JPN (L23828); NV: Norwalk (M87661) ; DSV395 :
Desert Shield 395 (U04469); DSV275: Desert Shield 275 (U04538); TV24 : Toronto (U02030) ; MV : Melksham (X81879);
SMA : Snow Mountain (L23831); HV : Hawaii (U07611) ; CV : Camberwell (U46500); MRV: Maryland (U07612) ; LV :
Lordsdale (X86557) ; BV: Bristol (X76716). The tree was rooted using the equivalent region of the bovine strain of NLV
‘Jena’ (AJ011099). The length of the abscissa to the connecting node is proportional to genetic distance between sequences.
The scale bar represents nucleotide substitutions per site. Numbers at the nodes of the tree indicate bootstrap values from
1000 replicates.
734 P. J. Marks and others
concerns about the potential adverse effects of chlor-
ine-releasing agents, the first of these used quaternary
ammonium products and anionic surface cleaners.
This cleaning intervention clearly made no impact
on the course of the outbreak. Although it has
been reported that a combination of quaternary
ammonium compounds and sodium carbonate at
twice the recommended concentration was effective
in removing feline calicivirus from surfaces [27],
other researchers have found that quaternary am-
monium compounds at twice the manufacturer’s re-
commended concentration failed to inactivate feline
calicivirus [26]. The experience during this outbreak
supports the view that cleaning with quaternary
ammonium compounds is unlikely to be effective at
inactivating NLV.
Following the second round of cleaning which in-
volved the use of chlorine-based products three pupils
reported an onset of illness on the first day the school
reopened but none of these stayed away from school.
These cases almost certainly occurred from person-
to-person transmission in the community whilst the
school was closed. This second cleaning was under-
taken when the school was closed for 4 days and it
is not possible to ascertain whether the dramatic
decrease in cases when the school reopened was due
to the closure reducing opportunities for person-
to-person spread or the removal of environmental
contamination by cleaning.
It could be argued that the outbreak would have
burnt itself out at this stage without cleaning or clos-
ure. Against this possibility is the fact that reported
attack rates in other outbreaks have been greater
(42–62 % [28–31]) and a study of human volunteers
using sensitive diagnostic techniques suggested that
around 70% of young adults are susceptible to NLV
infection [32]. This implies that there would have been
a pool of susceptible individuals in the school to allow
the outbreak to continue.
The level of secondary household cases in this study
(30%) is similar to that reported by Taylor et al.
(32%) [33]. The basic reproductive number greater
than one demonstrates the ability of NLV infection
to spread in the community, even when awareness of
the problem is high. Simple counts of those affected
within an institution where an outbreak occurs will
significantly underestimate the burden of the disease.
When outbreaks occur in institutions it is important
to reinforce advice about hygiene not only in the
institution at the centre of the outbreak, but also with
other people who come into contact with it.
Outbreaks of NLV infection within institutions
cause considerable disruption and have significant
economic implications [5]. This study has produced
firm evidence that vomiting is important in the trans-
mission of the disease. It also produced further evi-
dence to suggest that aerosolization of virus
particles can lead to direct infection. Transmission via
environmental contamination may also account for
some of the increased risk following exposure to
vomiting episodes. Further research is needed to
elucidate the optimal control measures for insti-
tutional outbreaks of NLV, but this study supports
laboratory evidence that cleaning with quaternary
ammonium products is unlikely to alter the course of
an outbreak.
ACKNOWLEDGEMENTS
The authors would like to thank the staff of the school
involved in this outbreak for their help with collection
of data for this study. We would also like to thank
Drs David Bullock and Selina Hoque and the staff of
the microbiology laboratories of the Southern Der-
byshire Acute Hospitals Trust for undertaking bac-
teriological examination of the stool samples. We also
thank Dr Keith Neal for comments on previous drafts
of this paper.
REFERENCES
1. Green KY, Ando T, Balayan MS, et al. Taxonomy of
the caliciviruses. J Infect Dis 2000 ; 181: S322–330.
2. Kapikian AZ, Wyatt RG, Dolin R, Thornhill TS,
Kalica RA, Channock RM. Visualisation by immune
electron microscopy of a 27 nm particle associated with
acute non-bacterial gastroenteritis. J Virol 1972 ; 10:
1075–1081.
3. Caul EO, Sellwood N, Brown D, et al. Outbreaks of
gastroenteritis associated with SRSVs. PHLS Microbiol
Dig 1993 ; 10: 2–8.
4. Caul EO. Viral gastroenteritis : small round structured
viruses, caliciviruses and astroviruses. Part II. The
epidemiological perspective. J Clin Pathol 1996 ; 49 :
959–964.
5. Caul EO. Viral gastroenteritis : small round structured
viruses, caliciviruses and astroviruses. Part I. The clini-
cal and diagnostic perspective. J Clin Pathol 1996 ; 49:
874–880.
6. Anonymous. A report of the study of infectious intesti-
nal disease in England. London : The Stationary Office,
2000.
7. Patterson W, Haswell P, Fryers PT, Green J. Outbreak
of small round structured virus gastroenteritis arose
Airborne transmission of NLV infection 735
after kitchen assistant vomited. CDR Rev 1997 ; 7:
R101–103.
8. Berg DE, Kohn MA, Farley TA, McFarland LM.
Multi-state outbreaks of acute gastroenteritis traced
to fecal-contaminated oysters harvested in Louisiana.
J Infect Dis 2000 ; 181: S381–386.
9. Ang LH. An outbreak of viral gastroenteritis associated
with eating raw oysters. Commun Dis Public Health
1998; 1: 38–40.
10. Stafford R, Strain D, Heymer M, Smith C, Trent M,
Beard J. An outbreak of Norwalk virus gastroenteritis
following consumption of oysters. Commun Dis Intell
1997; 21: 317–320.
11. Hafliger D, Hubner P, Luthy J. Outbreak of viral gastro-
enteritis due to sewage-contaminated drinking water.
Int J Food Microbiol 2000 ; 54: 123–126.
12. Kukkula M, Maunula L, Silvennoinen E, von Bons-
dorff CH. Outbreak of viral gastroenteritis due to drink-
ing water contaminated by Norwalk-like viruses. J Infect
Dis 1999; 180: 1771–1776.
13. Beller M, Ellis A, Lee SH, et al. Outbreak of viral gastro-
enteritis due to a contaminated well. International
consequences. JAMA 1997; 278: 563–568.
14. Caul EO. Small round structured viruses : airborne
transmission and hospital control. Lancet 1994 ; 343:
1240–1242.
15. Marks PJ, Vipond IB, Carlisle D, Deakin D, Fey RE,
Caul EO. Evidence for airborne transmission of
Norwalk-like virus (NLV) in a hotel restaurant. Epi-
demiol Infect 2000; 124: 481–487.
16. Ho MS, Glass RI, Monroe SS, et al. Viral gastroenter-
itis aboard a cruise ship. Lancet 1989; ii : 961–965.
17. Sawyer LA, Murphy JJ, Kaplan JE, et al. 25–30 nm vi-
rus particle associated with a hospital outbreak of acute
gastroenteritis with evidence for airborne transmission.
Am J Epidemiol 1988 ; 127: 1261–1271.
18. Cheesbrough JS, Green J, Gallimore CI, Wright PA,
Brown DWG. Widespread environmental contami-
nation with Norwalk-like viruses (NLV) detected in a
prolonged hotel outbreak of gastroenteritis. Epidemiol
Infect 2000; 125: 93–98.
19. Vipond IB, Pelosi E, Williams J, et al. A diagnostic EIA
for detection of the prevalent SRSV strain in United
Kingdom outbreaks of gastroenteritis. J Med Virol
2000; 61: 132–137.
20. Green SM, Lambden PR, Deng Y, et al. Polymerase
chain reaction detection of small round-structured
viruses from two related hospital outbreaks of gastro-
enteritis using inosine-containing primers. J Med Virol
1995; 45: 197–202.
21. Green J, Gallimore CI, Norcott JP, Lewis D,
Brown DW. Broadly reactive reverse transcriptase poly-
merase chain reaction for the diagnosis of SRSV-asso-
ciated gastroenteritis. J Med Virol 1995 ; 47: 392–398.
22. Hall TA. BioEdit : a user friendly biological sequence
alignment editor and analysis programme for Windows
95/98/NT. Nucl Acids Symp Ser 1999 ; 41: 95–98.
23. Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F,
Higgins DG. The CLUSTAL-X Windows interface :
flexible strategies for multiple sequence alignment aided
by quality analysis tools. Nucleic Acids Res 1997; 25:
4876–4882.
24. Page RD. TreeView : an application to display phylo-
genetic trees on personal computers. Comput Appl
Biosci 1996; 12: 357–358.
25. Kaplan JE, Feldman R, Campbell DS, Lookabaugh C,
Gary WG. The frequency of a Norwalk-like pattern of
illness in outbreaks of acute gastroenteritis. Am J Public
Health 1982; 72: 1329–1332.
26. Doultree JC, Druce JD, Birch CJ, Bowden DS, Mar-
shall JA. Inactivation of feline calicivirus, a Norwalk
virus surrogate. J Hosp Infect 1999 ; 41: 51–57.
27. Gulati BR, Allwood PB, Hedberg CW, Goyal SM.
Efficacy of commonly used disinfectants for the inacti-
vation of calicivirus on strawberry, lettuce, and a food-
contact surface. J Food Prot 2001 ; 64: 1430–1434.
28. Kobayashi S, Morishita T, Yamashita T, et al. A large
outbreak of gastroenteritis associated with a small
round structured virus among schoolchildren and tea-
chers in Japan. Epidemiol Infect 1991 ; 107: 81–86.
29. Green J, Wright PA, Gallimore CI, Mitchell O,
Morgan-Capner P, Brown DW. The role of environ-
mental contamination with small round structured
viruses in a hospital outbreak investigated by reverse-
transcriptase polymerase chain reaction assay. J Hosp
Infect 1998; 39: 39–45.
30. Gellert GA, Waterman SH, Ewert D, et al. An outbreak
of acute gastroenteritis caused by a small round struc-
tured virus in a geriatric convalescent facility. Infect
Control Hosp Epidemiol 1990; 11: 459–464.
31. Anonymous. Outbreaks of Norwalk-like viral gastro-
enteritis – Alaska and Wisconsin, 1999. MMWR Morb
Mortal Wkly Rep 2000 ; 49: 207–211.
32. Graham GY, Jiang X, Tanaka T, Opekun AR, Madore
HP, Estes MK. Norwalk virus infection of volunteers:
new insights based on improved assays. J Infect Dis
1994; 170: 34–43.
33. Taylor JW, Gary GW Jr, Greenberg HB. Norwalk-
related viral gastroenteritis due to contaminated drink-
ing water. Am J Epidemiol 1981 ; 114: 584–592.
736 P. J. Marks and others
