Journal of Infectious Diseases Advance Access published March 27, 2012

MAJOR ARTICLE

The Etiology of Severe Acute Gastroenteritis

Among Adults Visiting Emergency Departments
in the United States
Joseph S. Bresee,1,a Ruthanne Marcus,2,a Richard A. Venezia,3,a William E. Keene,4 Dale Morse,5 Mark Thanassi,6,a
Patrick Brunett,7 Sandra Bulens,1,a R. Suzanne Beard,1 Leslie A. Dauphin,1,a Laurence Slutsker,8,a Cheryl Bopp,8
Mark Eberhard,9 Aron Hall,1 Jan Vinje,1 Stephan S. Monroe,1,a Roger I. Glass,1,a and the US Acute Gastroenteritis
Etiology (AGE) Study Team

Downloaded from http://jid.oxfordjournals.org/ by guest on May 26, 2016

1Division of Viral Diseases, National Center for Immunizations and Respiratory Diseases, 2Connecticut Emerging Infections Program and 3Albany
Medical Center and 4Oregon Public Health Division, Department of Human Services, and 5New York Department of Health, Albany, New York; and
6Department of Emergency Medicine, YaleNew Haven Hospital, New Haven, Connecticut; 7Oregon Health and Science University, Portland, 8Foodborne
and Diarrheal Diseases Branch, Division of Bacterial and Mycotic Diseases, and 9Office of the Director, Division of Parasitic Diseases, National Center
for Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia

Background. Acute gastroenteritis (AGE) remains a common cause of clinic visits and hospitalizations in the
United States, but the etiology is rarely determined.
Methods. We performed a prospective, multicenter emergency departmentbased study of adults with AGE.
Subjects were interviewed on presentation and 34 weeks later. Serum samples, rectal swab specimens, and/or whole
stool specimens were collected at presentation, and serum was collected 34 weeks later. Fecal specimens were tested
for a comprehensive panel of viral, bacterial, and parasitic pathogens; serum was tested for calicivirus antibodies.
Results. Pathogens were detected in 25% of 364 subjects, including 49% who provided a whole stool specimen.
The most commonly detected pathogens were norovirus (26%), rotavirus (18%), and Salmonella species (5.3%).
Pathogens were detected significantly more often from whole stool samples versus a rectal swab specimen alone.
Nine percent of subjects who provided whole stool samples had .1 pathogen identified.
Conclusions. Viruses, especially noroviruses, play a major role as agents of severe diarrhea in adults. Further
studies to confirm the unexpectedly high prevalence of rotaviruses and to explore the causes of illness among
patients from whom a pathogen cannot be determined are needed. Studies of enteric pathogens should require the
collection of whole stool samples.

Received 3 May 2011; accepted 14 November 2011.

a
Present affiliations: Influenza Division, National Center for Immunizations and
Respiratory Disease (J. S. B.), Division of Healthcare and Quality Promotion (S. B.),
Divisions of Preparedness and Emerging Infections (L. A. D.) and High Consequence
Pathogens (S. S. M.), National Center for Emerging and Zoonotic Infectious
Diseases, and Office of the Director, Center for Global Health (L. S.), Centers for
Disease Control and Prevention, Atlanta, Georgia; Yale School of Medicine, New
Haven, Connecticut (R. M.); Department of Pathology, University of Maryland
School of Medicine, Baltimore (R. A. V.), and Fogarty Center, National Institutes of
Health, Bethesda (R. I. G.), Maryland; and Kaiser Santa Clara Medical Center,
Santa Clara, and Stanford University, Palo Alto, California (M. T.).
Correspondence: Joseph S. Bresee, MD, Centers for Disease Control and
Prevention, MS A-20, 1600 Clifton Rd NE, Atlanta, GA30333 ([email protected]).
The Journal of Infectious Diseases
Published by Oxford University Press on behalf of the Infectious Diseases Society
of America 2012.
This is an Open Access article distributed under the terms of the Creative
Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0), which permits unrestricted non-commercial use, distribution, and
reproduction in any medium, provided the original work is properly cited.
DOI: 10.1093/infdis/jis206

Acute gastroenteritis (AGE) remains a common cause

of clinic visits and hospitalizations in the United States.
An estimated 179 million cases of AGE occur each year,
resulting in millions of clinic visits, nearly 500 000
hospitalizations, and .5000 deaths [1, 2]. Although the
burden of AGE in children has been described [3, 4],
data on AGE in adults remain sparse. Between 1979
and 1995, 1.5% of all hospital discharges among adults
had an International Classification of Diseases, Ninth
Revision, Clinical Modification code for gastroenteritis,
meaning that the lifetime risk of being discharged from
the hospital with a diagnosis of gastroenteritis is approximately 1 in 8 among US adults [5].
Over the past 40 years, .20 new agents of gastroenteritis have been discovered [6]. Nevertheless,
Severe Gastroenteritis in US Adults

JID

MATERIALS AND METHODS

Subjects

Subjects were enrolled from the EDs in 3 major medical centers

that also serve as Foodborne Diseases Active Surveillance
Network sites [16, 17] in the United States: YaleNew Haven
Hospital, New Haven, Connecticut; Albany Medical Center,
Albany, New York; and Oregon Health and Science University,
Portland, Oregon. YaleNew Haven Hospital is a 944-bed
tertiary referral center where staff in the adult-specific ED
treat approximately 60 000 patients per year. Albany Medical
Center is a 631-bed tertiary referral center with approximately
55 000 ED visits a year. Oregon Health and Science University
is a 420-bed urban university teaching hospital with an annual
ED census of 46 000 visits.
Subjects were enrolled 5 days per week at each site during
the following periods: at YaleNew Haven Hospital, between
2

JID

Bresee et al

1 September 1999 and 31 August 2001; at Albany Medical

Center, between 22 January 2000 and 20 April 2001; and at
Oregon Health and Science University, between 25 August 2000
and 24 August 2001. All persons aged $18 years who presented
to a participating ED with AGE were asked to participate.
AGE was defined as the occurrence of $1 episode of vomiting
and or $3 episodes of diarrhea within a 24-hour period. Subjects were excluded if (1) they had onset of AGE symptoms
$7 days prior to the ED visit, (2) their reason for care was
unrelated to treatment for AGE, or (3) they had a known
noninfectious or chronic cause of their symptoms, such as
a inflammatory bowel disease or medication overdose. All enrolled subjects gave written informed consent. The study was
approved by institutional review boards at the CDC and at
each of the participating institutions.
Data Collection

After subjects provided informed consent, they were administered a standardized questionnaire on their illness characteristics, medical history, and specific exposures. Some data
were extracted from the subjects ED records, including illness signs and treatment details. Each subject was contacted
36 weeks after the ED visit, to assess illness duration, outcome,
and possible secondary spread of illness.
Specimen Collection and Testing

Whole stool samples were collected from each subject during

the ED visit. If whole stool specimens were not available, 2 rectal
swab specimens at minimum were collected.
Viruses
Viral RNA was extracted from a 10% clarified stool or rectal
swab specimen suspension, using a NucliSens Extractor (BioMerieux, Durham, NC), and was amplified by conventional
RT-PCR for norovirus [18] and astrovirus [19]. Positive results
were confirmed by sequencing [20]. Stool samples were tested
for rotavirus, using Pathfinder Rotavirus Kit (BioRad) and
RT-PCR [21].
Bacteria
A swab of each whole stool specimen or the original rectal
swab specimen was examined for Shigella species, Salmonella
species, Yersinia enterocolitica, Vibrio species, Escherichia coli
O157:H7 and other Shiga toxinproducing E. coli, Campylobacter
species, and enterotoxigenic E. coli. Stool samples were tested
for Clostridium difficile toxins A and B by the diagnostic assays
routinely used in each hospital.
Parasites
Two aliquots of stool were fixed in 10% formalin (1 aliquot) or
polyvinyl alcohol (1 aliquot) and stored at room temperature.
Stool specimens preserved in 10% formalin were concentrated
using a formol-ethyl acetate procedure and examined by microscopy for Giardia species and Entamoeba histolytica (wet
mount), Cyclospora and Isospora species (wet mount with
epifluorescence, safranin staining), and Microsporidia species

Downloaded from http://jid.oxfordjournals.org/ by guest on May 26, 2016

an etiologic agent is rarely identified in AGE cases, either

because stool samples are infrequently collected or because
many laboratories have a limited ability to detect the full range
of pathogens, especially viruses [7, 8]. Less than 20% of AGE
cases in the United States, including those requiring hospitalization, are attributed to specific pathogens [2]. In a study of
.30 000 hospitalized adults with diarrhea, a bacterial agent was
identified in ,6% of cases [9]. Similarly, over half of foodborne
disease outbreaks that occurred during 20062007 and were
reported to the Centers for Disease Control and Prevention
(CDC) had no confirmed etiologic diagnosis [10, 11].
The availability of new, more sensitive assays for detection
of enteric pathogens may change this picture. Between 1993
and 1997, the proportion of all foodborne outbreaks reported
to the CDC that were confirmed as outbreaks of norovirus
infection increased from 0.3% to 27% because of wider availability of reverse-transcription polymerase chain reaction (RTPCR) in public health laboratories [11, 12]. Noroviruses are
now recognized as the leading cause of epidemic gastroenteritis
in all age groups, accounting for .90% of viral gastroenteritis
outbreaks and approximately 50% of all gastroenteritis outbreaks [13]. Data on norovirus prevalence among sporadic
AGE cases, particularly among adults, have remained sparse
because RT-PCR is generally not used for diagnostic purposes
in clinical settings. Past reviews have found that norovirus was
responsible for 12% of AGE cases among all age groups in
community and outpatient settings [14] and 4.4%8.7% of
AGE cases among adults and elderly individuals admitted to
the emergency department (ED) or hospitalized [15].
This study prospectively enrolled and tested adults with
AGE presenting to EDs to determine the frequency, characteristics, and etiology of infectious AGE. We chose EDs as a study
setting because the disease severity and economic burden
associated with these cases are likely to be significant.

Data Handling and Statistical Analyses

All questionnaire and laboratory results were collated at each

participating site. At the CDC, data were merged and analyzed.
Proportional outcomes were compared using v2 tests, and
continuous variables were compared using nonparametric tests
(ie, the Wilcoxon rank-sum test). Stratified analyses were
performed using the Cochran-Mantel-Haenszel test.
RESULTS
During the study period, 389 subjects were enrolled. Of these,
25 did not meet eligibility criteria (24 had had symptoms
for .7 days before their ED visit, and 1 was ,18 years of age)
(Table 1). Of 364 subjects, 180 (49%) were from YaleNew
Haven Hospital, 138 (38%) were from Albany Medical Center,
and 46 (13%) were from Oregon Health and Science University.
Median age was 34 years (range, 1891 years). Fewer than 5%
of patients were $65 years of age. One-third (32%) reported
an underlying or chronic disease, and 6.4% lived in a chronic
care facility or nursing home. No differences in the characteristics of the subjects between the 3 study sites were observed.
Etiology From Testing of Stool Samples

Of 364 subjects, 330 (91%) provided stool specimens; 133 (40%)

had a whole stool specimen, 197 (60%) had a rectal swab
specimen, and 34 provided both (Table 2). Overall, norovirus
was detected in 16% of whole stool and rectal swab specimens
(42 of 264) combined, and rotavirus was detected in 14% (20 of
140). Although rotavirus was identified in all sites in each year

Table 1. Characteristics and Potential Risk Factors for Illness

Among Study Subjects, Multicenter Gastrointestinal Disease
Study, 19992001
Characteristic

Value (n 5 364)

Age, years
Median (range)
1835

34 (1891)
201 (55)

3664

145 (40)

6574

6 (1.6)
12 (3.3)

$75
Female sex
Chronic diseasesa
Illness duration before presentation, days,
median (range)

213 (59)
116 (32)
1 (17)

Possible risk factors for illness

Member of known outbreak

31 (9)

Exposed to ill person in household

Exposure to ill person outside household

55 (15)
46 (13)

Recent international travel

12 (3)

Resident of long-term care facility

Any possible risk factorb

23 (6)
110 (30)

Data are no. (%) of subjects, unless otherwise indicated.

a
Chronic diseases queried included diabetes, Crohn disease, hyperthyroidism,
ulcerative colitis, systemic lupus erythrematosis, history of bowel surgery,
irritable bowel syndrome, cancer, human immunodeficiency virus infection or
AIDS, history of organ transplantation, immunodeficient state, gastric ulcers,
renal disease, or other.
b

Subjects with any of the 5 potential risk factors listed in the table.

of the study, it was the predominant agent identified at Oregon

Health and Science University, where 13 of the 20 rotaviruspositive subjects were enrolled. All rotavirus-positive specimens
were confirmed by RT-PCR. Bacterial agents were detected in
29 of 316 cases (9%), with Salmonella being most common.
Parasitic agents were detected in 2 patients during the study.
Whole stool specimens were significantly more likely to yield
positive results than rectal swab specimens, both for viral and
bacterial agents (Table 2). Overall, a pathogen was detected in
25% of all patients, but the detection rate was higher (49%) for
subjects who provided whole stool samples, compared with subjects who had only rectal swab specimens (8.7%; P , .0001).
Detection rates for norovirus, rotavirus, and any bacterial
agent were 46-fold higher when testing was performed on
whole stool samples, compared with rectal swab specimens.
Mixed infections were detected in whole stool samples from
12 subjects (9%) but in none from rectal swab specimens
(Table 3). Norovirus was the most commonly detected pathogen in mixed infections. Samples from most (5 of 7) subjects
positive for C. difficile also tested positive for other pathogens.
Five of the 19 rotavirus-positive stool specimens (26%) also
tested positive for norovirus. No specific demographic characteristics or clinical risk factors were found to be associated
with having a mixed infection.
Severe Gastroenteritis in US Adults

JID

Downloaded from http://jid.oxfordjournals.org/ by guest on May 26, 2016

(chromotrope stain). Formol-ethyl acetatefixed specimens

were also tested for Cryptosporidium and Giardia species by
direct fluorescent antibody assay [22, 23]. Smears from some
polyvinyl alcoholpreserved stool specimens were examined
for E. histolytica and Blastocystis species. Formalin-fixed material from rectal swab specimens was tested on Alexon-Trends
Giardia intestinalis and Cryptosporidium parvum microplate
assay, and results were read visually. Polyvinyl alcohol
preserved slides were stained using Trends Trichrome Stain
Set and examined by light microscopy.
Serum specimens from the acute phase (during the first
5 days of symptoms) and the convalescent phase (during
36 weeks after resolution of symptoms) were collected from
each subject. Serum samples were tested for immunoglobulin G
antibodies to norovirus by use of a recombinant virus-like
particle (VLP)based enzyme immunosorbant assay [18, 24],
using the following VLPs: GI.1 Norwalk Virus, GII.1 Hawaii
Virus, GII.2 Chesterfield Virus, GII.3 Toronto Virus, GII.4
Burwash Landing Virus, GII.5 White River Virus, GII.6
Florida Virus, GII.7 Gwynedd Virus, and GII.8 Idaho Falls
Virus. Seroconversion for an individual patient was defined as
a $4-fold increase in antibody units between acute-phase and
convalescent-phase sera.

Table 2. Distribution of Pathogens by Type of Stool Specimen,

Multicenter Gastrointestinal Illness Study, 19992001
Whole Stool, % Rectal Swab, %
(No. Positive/
(No. Positive/
No. Tested)
No. Tested)

Pathogen

Subject
Age/Sex
P

Viral

25/F

Norovirus, rotavirus

27/M

Norovirus, rotavirus

26 (33/127)

6.6 (9/137)

.00002

25/M

Norovirus, rotavirus

Rotavirus

18 (19/106)

2.9 (1/34)

.04

48/M

Norovirus, rotavirus

47/F

Norovirus, rotavirus, Salmonella enterica serovar

Enteritidis

35/M

Norovirus, Clostridium difficile

20/M
23/F

Norovirus, C. difficile
Norovirus, C. difficile

1 (1/106)

Not tested

Total
Salmonellaspeciesb

17 (22/133)
5.3 (7/133)

3.8 (7/183)
2.2 (4/183)

Clostridium difficile

5.3 (7/133)

Not tested

Campylobacter
speciesc

3 (4/133)

0 (0/183)

Otherd

3 (4/133)

.003a
.21
.03

1.6 (3/183)

.46

Parasitic
3 (3/102)

0 (0/87)

.25

Giardia intestinalis
Blastocystis hominis

1 (1/102)
2 (2/102)

0 (0/87)
0 (0/87)

1.0
.25

Endolimax nana

1 (1/102)

0 (0/87)

1.0

Mixed infections

9 (12/133)
49 (65/133)

0 (0/197)
8.7 (17/197)

.00001
,.00001

The number of infections is greater than the number of subjects who tested
positive (12 subjects had mixed infections, and all pathogens are accounted for
in the table).
a
Comparison performed for all subjects with any bacterial pathogens detected
except Clostridium difficile, since only those persons with whole stool were
tested for this pathogen.
b

Includes S. enterica serovar Enteritidis (5), S. enterica serovar Typhimurium

(2), group B Salmonella organisms (1), S. enterica serotype Berta (1), S. enterica
serotype Newport (1), S. enterica serotype Thompson.
Includes C. jejuni (3) and C. coli (1).

Includes Vibrio parahaemolyticus (2); Shigella sonnei (2); enterotoxigenic

Escherichia coli O148:H28 LT, ST; Shiga toxinproducing E. coli O157:H7; and
Shiga toxinproducing E. coli O Rough:H34 stx1.
e
Denominators are total number of eligible subjects. Because not all subjects
specimens were tested for all pathogens, this proportion is assumed to be
a lower limit of the true proportion in this population.

Comparison of Norovirus Diagnosis From Stool and Sera

Samples

Serum pairs were obtained from 133 of the subjects with stool
specimens (Table 4). Subjects from whom paired sera were
obtained were similar to those without paired sera with respect to illness characteristics, sex, age, exposure history, and
outcomes. Evidence of acute norovirus infection was observed
in 29 (22%) of these subjects. The incidence of serologically
confirmed norovirus infection was similar among subjects
for whom rectal swab specimens versus whole stool specimens
were obtained (21% vs 22%; P 5 .96). While the sensitivity
of RT-PCR for norovirus detection was slightly better than
that for serology when whole stool specimens were available
(26% vs 22%), serology significantly increased the rate of
norovirus detection among subjects for whom only rectal swab
d

JID

Bresee et al

52/F

Norovirus, Campylobacter coli, Vibrio species

39/F

Shigella sonnei, C. difficile

26/M

S. enterica serotype Newport, C. difficile

38/F

Endolimax nana, Blastocystis hominis

specimens were available (21% vs 6.6%; P 5 .004). Subjects

with a longer duration of illness prior to ED visit and sample
collection had lower rates of norovirus detection in stool and
sera samples (Figure 1). Almost 90% of all norovirus-positive
subjects presented #3 days after illness onset. All noroviruspositive subjects presented #5 days after illness onset. In
samples collected #3 days after onset, norovirus was detected
in 35 of the 157 stool samples (23%), and seroconversion was
confirmed in 22 of the 83 serum pairs (27%), compared with
4 of the 70 stool samples (5.7%; P 5 .003) and 4 of 39 sera
(10%; P 5 .04), respectively, collected .3 days after onset
of illness. No significant differences in detection of bacterial
pathogens or rotavirus by the duration of symptoms prior
to presentation were observed.
Epidemiologic Features

No differences were observed in age distribution or presence of

underlying chronic disease, by pathogen type. Gastroenteritisassociated ED admissions and confirmed norovirus cases occurred throughout the year, and no clear seasonality was
noted. Seasonal peaks were difficult to discern for rotavirus and
bacterial pathogens because of the small numbers of cases.
The probable sources of the illnesses were usually not
known to the subject. Only 30% subjects had known prior
contact with another person with AGE, most commonly
a household member, during the week before their ED visit;
9% reported that, before their illness, they attended a group
event after which other people also became ill. Noroviruspositive subjects were more likely to have known exposures to
someone with AGE prior to their onset of illness, compared
with subjects without norovirus infection (46% vs 26%;
P 5 .006). Neither recent international travel nor residence
in a long-term care facility was associated with any particular
pathogen.

Downloaded from http://jid.oxfordjournals.org/ by guest on May 26, 2016

Total

Any documented
enteric pathogene

Pathogens

Norovirus
Astrovirus
Bacterial

Table 3. Characteristics of Subjects With Mixed Infections,

Multicenter Gastrointestinal Infection Study, 19992001

Table 4. Comparison of Stool and Serum Testing for Norovirus

Infection, Multicenter Gastrointestinal Infection Study, 19992001

Test(s)

Stool Specimen Type, % Positive

(No. Positive/No. Tested)

Only stool RT-PCR

6.6 (9/137)

26 (33/127)

Only serology

21 (14/66)

22 (15/67)

Stool RT-PCR or serologya

24 (11/46)

27 (18/66)

Swab

Whole stool

Abbreviation: RT-PCR, reverse-transcription polymerase chain reaction.

a

From patients with both types of specimens.

Clinical Features

Treatment and Outcome

While few subjects had documented clinical signs of dehydration, 81% were treated with intravenous rehydration.
Overall, 45 of the 350 subjects (13%) for whom follow-up was
available were admitted to the hospital. No subject died. While
53 of 171 subjects (31%) continued to have gastrointestinal
symptoms 12 weeks following discharge and 20% of 108
subjects required medical follow-up after discharge because
of continuing symptoms, none were readmitted to the hospital during the follow-up period. No differences were identified in outcomes or treatments by pathogen type, except that
norovirus-positive subjects were less likely to report continued
symptoms 2 weeks following ED discharge, compared with
those who were norovirus negative (16% vs 38%; P 5 .02)
DISCUSSION
This study provides unique data on the infectious causes of
AGE in adults treated in EDs. An etiologic agent was identified
in almost 50% of cases when whole stool specimens were

collected and tested using all available laboratory assays for

known enteric pathogens. Norovirus was the most frequently
identified cause of ED visits for AGE, with detection in 27% of
subjects for whom stool and serum specimens were available.
While no other data are available on adults presenting to EDs,
studies of AGE etiology conducted in outpatient settings in
other industrialized countries also identified norovirus as the
leading cause [19, 2528]. A study of general practitioner visits
for AGE in England identified norovirus in 30% of adult cases
and 9% of adult controls, using a combination of electron microscopy and RT-PCR [25]. Noroviruses were detected in 16%
of cases and 3% of controls in outpatient clinics in Germany
across all age groups [26]. Other studies have generally detected
that #10% of stool specimens from adults with AGE were
positive for norovirus in community or outpatient settings
[19, 2729]. Hospital-based studies among adults in Ireland [30]
and South Africa [31] also found relatively lower rates of
norovirus infection (11% and 10%, respectively). The relatively
high rates of norovirus detection in our study may represent
the use of both serologic and RT-PCR methods, the appropriate collection and handling of specimens, and/or the exclusion of subjects with onset .1 week before presentation.
In a previous study in which serologic testing was used, evidence of norovirus infection was found in 33% of adults
followed for 1 year [32].
The finding of rotavirus in 18% of adult cases for which
whole stool specimens were available was surprising. While
most rotavirus-positive subjects were enrolled from one of
the 3 sites, cases were confirmed in all 3 sites during the study.
Rotaviruses are the most common cause of severe gastroenteritis among young children, but adult cases are thought to
be uncommon [31, 33, 34]. Adults who work in child-care
settings or who care for young children have been reported
to develop rotavirus gastroenteritis following exposure to sick
children [35]. Data on these exposures were not available in
our study, but the lack of rotavirus detection in elderly
Severe Gastroenteritis in US Adults

JID

Downloaded from http://jid.oxfordjournals.org/ by guest on May 26, 2016

Subjects reported becoming ill 1 day (interquartile range,

13 days) before presentation to the ED. Nausea (93%), vomiting
(81%) or diarrhea (89%), and abdominal pain (76%) were
reported by most subjects. Signs of moderate-to-severe dehydration, such as dry mucous membranes, decreased skin
turgor, or altered mental status, were present in ,10% of
subjects on examination. A temperature .37.8C at admission
(14%) and blood in stool (15%) was uncommon, and there
was no difference between subjects with confirmed viral
infection versus those with bacterial infection. Respiratory
symptoms, including sore throat, cough, and rhinorrhea,
were reported in approximately 10% of the subjects.
Few differences in clinical features were observed that could
distinguish subjects with AGE due to viral infection from those
with bacterial infection. Norovirus-positive subjects were
slightly more likely to present with vomiting (89%), compared
with subjects with either rotavirus infection (70%) or bacterial
infection (69%) (P 5 .06).

Figure 1. Cumulative proportion of pathogens detected, by duration

of illness prior to presentation. Abbreviations: ED, emergency department;
EIA, enzyme immunosorbant assay; NoV, norovirus; PCR, reverse-transcription
polymerase chain reaction; RV, rotavirus.

JID

Bresee et al

likely to have an infectious pathogen, thus making testing

more efficient and directed. Most of the subjects were young
and healthy and had no obvious risk factors for illness. While
previous studies found higher rates of norovirus among elderly
individuals [25] or during winter and spring seasons [26, 28, 43],
our study failed to confirm these findings. While subjects
with norovirus and rotavirus infections were more likely to
report exposure to other ill people, most positive subjects did
not report any exposure. Similarly, while subjects with viral
infections were more likely to report exposure to a known
outbreak of AGE, this exposure was noted in ,15% of all
subjects. From these data, we were unable to determine variables that would help clinicians or public health investigators
target certain people for testing.
This study has several important limitations. The number
of sites and subjects, weekly period of enrollment, and the duration of the study were necessarily limited. Therefore, these
data may not be representative of all cases presenting to EDs
in the United States each year. Even so, this study provides
the most comprehensive data to date on the etiology of AGE
in adults, and the finding that the distribution of pathogens
was similar in both years and in all 3 geographically distinct sites
provides some reassurance that the findings are reasonably
representative. Because of the complexity of the specimen handling and testing algorithm and the variable amount of stool
available for testing, some patients were not tested for all
pathogens. This may have biased the results; however, no demographic or clinical differences were observed for the various
subsets of patients for whom testing was performed, nor for
subjects who were not tested at all because of ineligibility or lack
of any specimens. Finally, the lack of healthy controls limits
conclusions about the relative importance of each pathogen.
The high prevalence of viral agents and the superiority of
whole stool specimens collected early in illness for detection of
bacterial and viral pathogens may have important implications
for development of simple, efficient algorithms for testing patients with AGE. Timely and sensitive detection of viral pathogens may help avoid unnecessary antibiotic use and further
diagnostic testing. Parasitic agents were rare in these subjects,
possibly as a result of the case definition requiring that the
illness duration was ,7 days before the ED visit. Parasitic tests
might be limited to patients with consistent clinical and epidemiologic characteristics. Vaccines against rotavirus are
used widely in infants in the United States and other countries
[44, 45], and the impact of noted disease reductions among
infants may result in fewer infections among adults [46].
Vaccines against noroviruses are in early development stages
[47, 48] but offer the promise of preventing the most common
cause of outbreaks and sporadic gastroenteritis among adults.
Finally, simpler, economical diagnostics for the variety of
enteric pathogens are needed. New diagnostics that would
allow for detection of multiple enteric pathogens could support

Downloaded from http://jid.oxfordjournals.org/ by guest on May 26, 2016

subjects may indicate that these infections are more common

among those more likely to be exposed to young children (eg,
parents and child-care workers). This study was conducted
prior to the introduction of universal childhood rotavirus
vaccination. Rotavirus vaccine appears to confer herd protection in older children, and it may do so in parents of
vaccinated children as well [36].
Salmonella and Campylobacter species were the most common bacterial pathogens identified. C. difficile was identified
relatively frequently in the small number of samples available
for testing but in most cases was detected in association with
another pathogen. Of the 2 subjects for whom C. difficile was
detected as the sole pathogen, neither reported use of antibiotics in the month prior to their illness.
An important finding of this study was the high rate of
pathogen detection in whole stool specimens, compared with
rectal swab specimens. Historically, rectal swab specimens
placed in transport media have been the recommended specimens for diagnosis of bacterial infection. This study clearly
indicates that testing only rectal swab specimen significantly
reduces the chance of establishing an etiologic diagnosis, for
both viral and bacterial infections. For noroviruses, the addition of serologic testing of subjects for whom only rectal swab
specimens were available increased the detection rate to levels
equal to those of testing whole stool samples by RT-PCR.
Since testing of paired sera is impractical for clinical diagnosis, whole stool specimens should be collected for viral
testing. Few studies have compared bacterial detection rates
for whole stool specimens and rectal swab specimen in the
same study, and those that have performed such comparison
produced mixed results [3740]. Even so, use of rectal swab
specimens alone for diagnosis may significantly reduce sensitivity for detection and is not supported by this study.
Finally, while only 12% of subjects presented with vomiting
without diarrhea, these persons were less likely to have
a whole stool sample collected than were subjects with diarrhea. Pathogen detection in these patients may therefore
be a particular challenge.
The likelihood of detecting norovirus depended on the
duration of symptoms that were observed. While one previous study found no such differences [41], subjects in our
study who presented earlier in their illness were more likely
to test positive for norovirus by either seroconversion or by
RT-PCR than were those who presented later. This finding
supports CDC recommendations to collect specimens early
during illness from patients with AGE [42]. However, no apparent decline in rates of detection for rotavirus or bacterial
agents during this window was observed, so this variable
alone might not dissuade a clinician from testing a person
with AGE.
One objective of this study was to identify epidemiologic or
clinical variables that could be used to identify patients most

better understanding of the true impact of these agents, enhance syndromic surveillance systems, and spur interest in
development of novel therapeutics and vaccines for these
agents. Progress in treatment, prevention, and diagnosis, however, will always rely on the appropriate collection, testing, and
interpretation of clinical samples.
Notes

References
1. Jones TF, McMillian MB, Scallan E, et al. A population-based estimate
of the substantial burden of diarrhoeal disease in the United States;
FoodNet, 19962003. Epidemiol Infect 2007; 135:293301.
2. Scallan E, Griffin PM, Angulo FJ, Tauxe RV, Hoekstra RM. Foodborne
illness acquired in the United Statesunspecified agents. Emerg Infect
Dis 2011; 17:1622.
3. Malek MA, Curns AT, Holman RC, et al. Diarrhea- and rotavirusassociated hospitalizations among children less than 5 years of
age: United States, 1997 and 2000. Pediatrics 2006; 117:188792.
4. Fischer TK, Viboud C, Parashar U, et al. Hospitalizations and deaths
from diarrhea and rotavirus among children ,5 years of age in the
United States, 19932003. J Infect Dis 2007; 195:111725.
5. Mounts AW, Holman RC, Clarke MJ, Bresee JS, Glass RI. Trends in
hospitalizations associated with gastroenteritis among adults in the
United States, 19791995. Epidemiol Infect 1999; 123:18.
6. Glass RI, Ando T, Noel J, et al. The human enteric caliciviruses: an
expanded role for an old virus. In: Scheld WM, Craig WA, Hughes JM,
eds. Emerging infections. 4th ed. Washington, DC: ASM Press, 2000:
3344.
7. Jones TF, Bulens SN, Gettner S, et al. Use of stool collection kits delivered to patients can improve confirmation of etiology in foodborne
disease outbreaks. Clin Infect Dis 2004; 39:14549.
8. Carpenter LR, Pont SJ, Cooper WO, et al. Stool cultures and antimicrobial prescriptions related to infectious diarrhea. J Infect Dis 2008;
197:170912.
9. Slutsker L, Ries AA, Greene KD, Wells JG, Hutwagner L, Griffin PM.
Escherichia coli 0157:H7 diarrhea in the United States: clinical and
epidemiologic features. Ann Intern Med 1997; 126:50513.
10. Boore A, Herman KM, Perez AS, et al. Surveillance for foodborne
disease outbreaksdUnited States, 2007. MMWR Morb Mortal Wkly
Rep 2010; 59:9739.
11. Ayers LT, Williams IT, Gray S. Surveillance for foodborne disease
outbreaksdUnited States, 2006. MMWR Morb Mortal Wkly Rep
2009; 58:60915.
12. Olsen SJ, MacKinnon LC, Goulding JS, Bean NH, Slutsker L. Surveillance for foodborne-disease outbreaksdUnited States, 19931997.
MMWR 2000; 49 Suppl 1:162.
13. Patel MM, Hall AJ, Vinje J, Parashar UD. Noroviruses: a comprehensive review. J Clin Virol 2009; 44:18.

Severe Gastroenteritis in US Adults

JID

Downloaded from http://jid.oxfordjournals.org/ by guest on May 26, 2016

Acknowledgments. The US AGE Etiology Study Team: Jan Keithly,

Jill Ennis, and Mary Marchewka (Parasitology Laboratory, Wadsworth
Center, Albany, NY); Nando Chaterjee (New York State Department of
Health, Albany, NY); and John Jui (Oregon Health and Science University,
Portland, OR).
Disclaimer. The findings and conclusions of this report are those of
the authors and do not necessarily represent the views of the funding
agency.
Financial support. This work was supported by the National Center
for Infectious Diseases, CDC.
Potential conflicts of interest. All authors: No reported conflicts.
All authors have submitted the ICMJE Form for Disclosure of Potential
Conflicts of Interest. Conflicts that the editors consider relevant to the
content of the manuscript have been disclosed.

14. Patel MM, Widdowson MA, Glass RI, Akazawa K, Vinje J, Parashar UD.
Systematic literature review of role of noroviruses in sporadic gastroenteritis. Emerg Infect Dis 2008; 14:122431.
15. Haustein T, Harris JP, Pebody R, Lopman BA. Hospital admissions
due to norovirus in adult and elderly patients in England. Clin Infect
Dis 2009; 49:18902.
16. Allos BM, Moore MR, Griffin PM, Tauxe RV. Surveillance for sporadic
foodborne disease in the 21st century: the FoodNet perspective. Clin
Infect Dis 2004; 38(Suppl 3):S11520.
17. Scallan E. Activities, achievements, and lessons learned during the
first 10 years of the Foodborne Diseases Active Surveillance Network:
19962005. Clin Infect Dis 2007; 44:71825.
18. Monroe SS, Stine SE, Jiang XI, Estes MK, Glass RI. Detection of antibody to recombinant Norwalk virus antigen in specimens from outbreaks of gastroenteritis. J Clin Microbiol 1993; 31:286672.
19. de Wit MA, Koopmans MP, Kortbeek LM, et al. Sensor, a populationbased cohort study on gastroenteritis in the Netherlands: incidence
and etiology. Am J Epidemiol 2001; 154:66674.
20. Fankhauser RL, Monroe SS, Noel JS, et al. Epidemiologic and molecular trends of Norwalk-like viruses associated with outbreaks of
gastroenteritis in the United States. J Infect Dis 2002; 186:17.
21. Grinde B, Jonassen TO, Ushijima H. Sensitive detection of group A
rotaviruses by immunomagnetic separation and reverse transcription
polymerase chain reaction. J Virol Methods 1995; 55:32738.
22. Garcia LS, Shimizu RY. Evaluation of nine immunoassay kits (enzyme
immunoassay and direct fluorescence) for detection of Giardia lamblia
and Cryptosporidium parvum in human fecal specimens. J Clin Microbiol
1997; 35:15269.
23. Johnston SP, Ballard MM, Beach MJ, Causer L, Wilkins PP.
Evaluation of three commercial assays for detection of Giardia and
Cryptosporidium organisms in fecal specimens. J Clin Microbiol 2003;
41:6236.
24. Noel JS, Ando T, Leite JP, et al. Correlation of patient immune responses with genetically characterized small round-structured viruses
involved in outbreaks of nonbacterial acute gastroenteritis in the
United States, 1990 to 1995. J Med Virol 1997; 53:37283.
25. Amar CF, East CL, Gray J, Iturriza-Gomara M, Maclure EA,
McLauchlin J. Detection by PCR of eight groups of enteric pathogens
in 4,627 faecal samples: re-examination of the English case-control
Infectious Intestinal Disease Study (19931996). Eur J Clin Microbiol
Infect Dis 2007; 26:31123.
26. Karsten C, Baumgarte S, Friedrich AW, et al. Incidence and risk
factors for community-acquired acute gastroenteritis in north-west
Germany in 2004. Eur J Clin Microbiol Infect Dis 2009; 28:93543.
27. de Wit MA, Koopmans MP, Kortbeek LM, van Leeuwen NJ, Vinje J,
van Duynhoven YT. Etiology of gastroenteritis in sentinel general
practices in The Netherlands. Clin Infect Dis 2001; 33:2808.
28. Huhulescu S, Kiss R, Brettlecker M, et al. Etiology of acute gastroenteritis in three sentinel general practices, Austria 2007. Infection 2009;
37:1038.
29. Rockx B, De Wit M, Vennema H, et al. Natural history of human
calicivirus infection: a prospective cohort study. Clin Infect Dis 2002;
35:24653.
30. Foley B, OMahony J, Morgan SM, Hill C, Morgan JG. Detection of
sporadic cases of Norwalk-like virus (NLV) and astrovirus infection in
a single Irish hospital from 1996 to 1998. J Clin Virol 2000; 17:10917.
31. Wolfaardt M, Taylor MB, Booysen HF, Engelbrecht L, Grabow WO,
Jiang X. Incidence of human calicivirus and rotavirus infection in patients with gastroenteritis in South Africa. J Med Virol 1997; 51:2906.
32. Payment P, Franco E, Fout GS. Incidence of Norwalk virus infections
during a prospective epidemiological study of drinking water-related
gastrointestinal illness. Can J Microbiol 1994; 40:8059.
33. Staat MA, Azimi P, Berke T, et al. Clinical presentations of rotavirus
infection among hospitalized children. Pediatr Infect Dis J 2002; 21:2217.
34. de Wit MA, Koopmans MP, Kortbeek LM, van Leeuwen NJ, Bartelds AI,
van Duynhoven YT. Gastroenteritis in sentinel general practices, The
Netherlands. Emerg Infect Dis 2001; 7:8291.

35. Koopman JS, Monto AS, Longini IM Jr. The Tecumseh Study. XVI:
Family and community sources of rotavirus infection. Am J Epidemiol
1989; 130:7608.
36. Cortese MM, Tate JE, Simonsen L, Edelman L, Parashar UD. Reduction
in gastroenteritis in United States children and correlation with early
rotavirus vaccine uptake from national medical claims databases.
Pediatr Infect Dis J 2010; 29:48994.
37. McCall CE, Martin WT, Boring JR. Efficiency of cultures of rectal
swabs and faecal specimens in detecting salmonella carriers: correlation with numbers of salmonellas excreted. J Hyg (Lond) 1966; 64:
2619.
38. Stuart RD. Transport medium for specimens in public health bacteriology. Public Health Rep 1959; 74:4318.
39. Kaplan RL, Goodman LJ, Barrett JE, Trenholme GM, Landau W.
Comparison of rectal swabs and stool cultures for detecting Campylobacter fetus subsp. jejuni. J Clin Microbiol 1982; 15:95960.
40. Hardy AV, Mackel D, Frazier D, Hamerick D. The relative efficacy of
cultures for shigella. U S Armed Forces Med J 1953; 4:3934.
41. Marshall JA, Hellard ME, Sinclair MI, et al. Incidence and characteristics of endemic Norwalk-like virus-associated gastroenteritis. J Med
Virol 2003; 69:56878.

42. Centers for Disease Control and Prevention. Norwalk-like viruses:

public health consequences and outbreak management. MMWR 2001;
50:118.
43. Lopman B, Armstrong B, Atchison C, Gray JJ. Host, weather and
virological factors drive norovirus epidemiology: time-series analysis
of laboratory surveillance data in England and Wales. PLoS One 2009;
4:e6671.
44. Heaton PM, Goveia M, Miler JM, Offit P, Clark HF. Development of
pentavalent rotavirus vaccine against prevalent serotypes of rotavirus
gastroenteritis. J Infect Dis 2005; 192:S1721.
45. DeVos B, Vesikari T, Linhares AC, et al. A rotavirus vaccine for
prophylaxis of infants against rotavirus gastroenteritis. Pediatr Infect
Dis J 2004; 10:S17982.
46. Lopman B, Curns AT, Yen C, Parashar U. Infant rotavirus vaccination
may provide indirect protection to older children and adults in the
United States. J Infect Dis 2011; 204:9806.
47. Tacket CO, Sztein MB, Losonsky G, Wasserman SS, Estes MK.
Humoral, mucosal and cellular immune responses to oral norwalk-like
virus particles in volunteers. Clin Immunol 2003; 108:2417.
48. Herbst-Kralovetz M, Mason HS, Chen Q. Norwalk virus-like particles
as vaccines. Expert Rev Vaccines 2010; 9:299307.

Downloaded from http://jid.oxfordjournals.org/ by guest on May 26, 2016

JID

Bresee et al