PulseNet USA: a five-year update.

Abstract

PulseNet USA is the molecular surveillance network for foodborne infections in the United States. Since its inception in 1996, it has been instrumental in detection, investigation and control of numerous outbreaks caused by Shiga toxin-producing Escherichia coli O157:[H7] (STEC O157), Salmonella enterica, Listeria monocytogenes, Shigella spp., and Campylobacter. This paper describes the current status of the network, including the methodologies used and its future possibilities. The currently preferred subtyping method in the network is pulsed-field gel electrophoresis (PFGE), a proven highly discriminatory molecular subtyping method. New simpler sequencebased subtyping methods are under development and validation to complement and eventually replace PFGE. PulseNet is essentially a cluster detection network, but the data in the system will now also be used in attribution analyses of sporadic infections. The PulseNet platform will also be used as a primary tool in preparedness and response to acts of food bioterrorism.

Full text

Page 1

FOODBORNE PATHOGENS AND DISEASE
Volume 3, Number 1, 2006
© Mary Ann Liebert, Inc.
PulseNet USA: A Five-Year Update
P. GERNER-SMIDT,1K. HISE,1J. KINCAID,1S. HUNTER,1S. ROLANDO,2
E. HYYTIÄ-TREES,1E.M. RIBOT,1B. SWAMINATHAN,1and the PulseNet Taskforce3
ABSTRACT
PulseNet USA is the molecular surveillance network for foodborne infections in the United States. Since its in-
ception in 1996, it has been instrumental in detection, investigation and control of numerous outbreaks caused by
Shiga toxin–producing Escherichia coli O157:[H7] (STEC O157), Salmonella enterica, Listeria monocytogenes,
Shigella spp., and Campylobacter. This paper describes the current status of the network, including the method-
ologies used and its future possibilities. The currently preferred subtyping method in the network is pulsed-field
gel electrophoresis (PFGE), a proven highly discriminatory molecular subtyping method. New simpler sequence-
based subtyping methods are under development and validation to complement and eventually replace PFGE.
PulseNet is essentially a cluster detection network, but the data in the system will now also be used in attribu-
tion analyses of sporadic infections. The PulseNet platform will also be used as a primary tool in preparedness
and response to acts of food bioterrorism.
9
INTRODUCTION
F
lion annual cases in the United States in 1999
(Mead et al., 1999). In recent years, the inci-
dence of infections caused by Shiga toxin–pro-
ducing Escherichia coli (STEC) O157, thermotol-
erant Campylobacter, and Listeria monocytogenes
has declined, whereas infections caused by
non-typhoid Salmonella or Vibrio spp. seem to
occur with unchanged or higher incidence
(CDC, 2005). Most foodborne infections have
no known connection to each other; that is, they
are sporadic. However, outbreaks are not in-
frequent and knowledge about the most im-
portant vehicles of infection and transmission
of foodborne pathogens often come from out-
break investigations (CDC, 2005).
PulseNet USA is the national molecular sur-
veillance network for foodborne infections in
OODBORNE INFECTIONS are major threats to
the public health, with an estimated 76 mil-
the USA (Gerner-Smidt et al., 2005; Swami-
nathan et al., 2001). By rapidly detecting clus-
ters of foodborne infections, it is a key tool in
the recognition and investigation of outbreaks.
Clusters detected by the network must be con-
firmed by epidemiological means to determine
if an outbreak is occurring. Presently, the pre-
ferred subtyping method of PulseNet USA is
pulsed-field gel electrophoresis (PFGE). The
network was established in 1996 for the detec-
tion and investigation of outbreaks of food-
borne infections in the United States (Swami-
nathan et al., 2001). PulseNet has proved itself
as an extremely efficient tool for detection, in-
vestigation and subsequently the control of
outbreaks of foodborne infections in this coun-
try. Since 1996, the network has grown steadily
from five participating laboratories to 65 par-
ticipants in 2005. Full national participation
of all 50 states was achieved in 2001. Also the
number of patterns submitted to the national
1Foodborne and Diarrheal Diseases Branch, Centers for Disease Control and Prevention, Atlanta, Georgia.
2Association of Public Health Laboratories, Washington, D.C.
3M. Adams, M. Bird, P. Bordoni, K. Cooper, S. Van Duyne, M.A. Fair, N. Garrett, L.M. Graves, D. Jennings, M.
Joyner, B. McGlinchey, and N. Patel.

Page 2

database on an annual basis has increased al-
most exponentially from 196 STEC O157 in
1996 to almost 40,000 of seven organisms in
2004 (Fig. 1).
Every time a new pathogen is added to the
network more outbreaks caused by that or-
ganism have been detected as shown for out-
breaks of listeriosis after the introduction of Lis-
teria monocytogenes in the PulseNet system in
1998 in Figure 2.
The current paper presents an update of the
organization, a description of some illustrative
functions and features of, and future directions
for PulseNet USA.
PULSENET USA ORGANIZATION
AND PARTICIPANTS
PulseNet USA is a collaboration between
the Association of Public Health Laboratories
(APHL), The Centers for Disease Control and
Prevention (CDC), the Council of State and Ter-
ritorial Epidemiologists (CSTE), the Food and
Drug Administration’s Center for Food Safety
and Applied Nutrition (FDA-CFSAN), Office
of Regulatory Affairs (FDA-ORA) and Center
of Veterinary Medicine (FDA-CVM), and the
U.S. Department of Agriculture, Food and
Safety Inspection Service (USDA-FSIS), Agri-
cultural Research Service (USDA-ARS) and
Agricultural Marketing Service (USDA-AMS).
The network is coordinated from the Food-
borne and Diarrheal Diseases Branch (FDDB)
at CDC in Atlanta. It is governed by the
PulseNet USA steering committee, which has
members from all involved parties. The chair-
man of the committee is the chief of Epidemi-
ological Investigations and PulseNet unit at
CDC. The participants include public health
laboratories in all 50 states, four counties and
three cities and eight food safety regulatory
laboratories (USDA-FSIS, FDA-CFSAN/ORA/
CVM 7) (Fig. 3).
APHL is the administrative center of PulseNet.
All PulseNet meetings, training sessions and
workshops are organized and coordinated by
APHL. Method development projects are ad-
ministered by APHL under a cooperative
agreement with CDC. The participating public
health laboratories perform PFGE of foodborne
organisms from clinical specimens and some-
times also from food sources in their catchment
area in real-time; the patterns they generate are
compared with profiles in a local database
thereby enabling local cluster detection before
they are submitted or uploaded to the National
databases; local clusters/outbreaks are investi-
gated in a collaborative effort between the lo-
cal PulseNet laboratories and epidemiologists.
The participation in PulseNet of the public
health laboratories is partly funded by the
states and partly by grants from the Depart-
ment of Health and Human Services adminis-
tered by CDC. The participants in the USDA/
FSIS/ARS/AMS and the FDA-CFSAN/ORA/
CVM laboratories upload patterns of isolates
from cultures of food and animal samples ob-
tained during the legally required surveillance
of the food chain and during outbreak investi-
gations.
In order to increase the efficiency of the net-
work and promote local collaboration the coun-
try is divided into eight PulseNet areas (Fig. 3)
GERNER-SMIDT ET AL.
10
FIG. 1.
(PFGE) patterns submitted to the PulseNet national data-
base 1996–2004.
Number of pulsed-field gel electrophoresis
FIG. 2.
the United States, 1978–2002. The arrow points to the in-
clusion of Listeria monocytogenes in the PulseNet program.
Number of outbreaks of listeriosis detected in

Page 3

each with a designated PulseNet area labora-
tory. These laboratories provide participants in
their region laboratory support including trou-
bleshooting of the PulseNet methods and surge
subtyping capacity in the case a participant
temporarily is unable to perform PFGE in real-
time.
The PulseNet database team and the meth-
ods development laboratory are located cen-
trally at CDC.
The database team is responsible for main-
tenance of the database system including
naming and confirmation of the profiles sub-
mitted/uploaded to the National databases,
continuous central cluster detection and is the
central point of communication for the partic-
ipants and epidemiologists for clusters and
outbreaks. The CDC database team continu-
ously monitors patterns submitted to the Na-
tional database. At least once a week, the team
compares all patterns submitted to the central
database within the previous 60 days (120 for
Listeria). If the team detects a new local cluster,
it contacts the submitting laboratory to gather
more information and encourage the labora-
tory to contact the local epidemiologist to ini-
tiate an outbreak investigation. If the database
team detects a new cluster of patterns submit-
ted from more than one state, the information
is brought to the attention of the CDC epi-
demiologists. If the cluster is considered likely
to represent an outbreak, the state public health
laboratories, which have submitted isolates in
the cluster, are contacted and if there is no clear
explanation for the clustering, an outbreak in-
vestigation is initiated. What makes a cluster
worthy of consideration is currently done at the
discretion of the database team member who
takes into account the overall frequency of the
clustered pattern in the database and the time
span of the collection dates of the patterns to
the central database. The database team also
routinely compares patterns of food isolates
submitted by USDA/Food Safety Inspection
Service (FSIS) and FDA with all other patterns
in the national database. If a food product iso-
late pattern is indistinguishable from that of a
contemporaneous human isolate, the appro-
priate food regulatory agency is informed
about the match and an epidemiologic investi-
gation is initiated.
The methods development laboratory is the
PULSENET USA: A FIVE-YEAR UPDATE
11
FIG. 3.
Geographical organization of PulseNet USA.

Page 4

central reference laboratory responsible for de-
velopment of new standardized PFGE proto-
cols and the evaluation and validation of new
methods supplementing PFGE for the network;
it also helps the PulseNet area laboratories in
providing troubleshooting of the laboratory
methods and extra subtyping capacity for the
participants.
PULSENET PROCEDURES, DATA
ANALYSIS, AND COMMUNICATION
PFGE protocols
PFGE is by nature a comparative subtyping
method, but by rigorously standardizing the
subtyping protocols, the analysis tools and
training of the participants, the patterns gener-
ated in the network may be compared between
laboratories.
Currently, PulseNet USA has standardized
PFGE protocols for seven organisms: STEC
O157, Salmonella enterica, Shigella spp., Listeria
monocytogenes,
thermotolerant Campylobacter
spp. (?www.cdc.gov/pulsenet/protocols.htm?),
Clostridium perfringens (Maslanka et al., 1999),
and Vibrio cholerae (Cooper, 2006). For each or-
ganism the protocol has been developed with at
least two enzymes; the primary enzyme is used
for screening of relatedness and the secondary
for confirmation of the results obtained by the
first enzyme or supplementary discrimination.
For the protocols where a third enzyme has been
devised, this is also used for confirmatory pur-
poses, for example, the primary, secondary, and
tertiary restriction enzymes for E. coli are XbaI,
BlnI (AvrII), and SpeI, respectively. In order to
simplify the work with the different enzymes,
the electrophoretic running conditions are the
same for the primary and secondary (and where
applicable the tertiary) enzyme for all organisms
except for Campylobacter. However, all organ-
isms have running conditions that are unique to
that organism except for E. coli and Shigella,
which have identical protocols.
Another essential feature of the PulseNet
system is the use of a universal standard by all
participants, which is run in every fourth to
fifth lane in all gels thus allowing for reliable
normalization of the patterns of the test isolates
in the adjoining lanes. This standard, com-
prised of XbaI restriction fragments of Salmo-
nella Braenderup strain H9812 (ATCC BAA-
664) is used as a reference DNA fragment size
standard for all pathogens under surveillance
in PulseNet. The standard contains ?18 bands
in the range between approximately 20.5 and
1,135 kb, of which 15 are sufficiently evenly dis-
tributed to allow for a consistent and repro-
ducible normalization of all PFGE images run
according to any of the PulseNet protocols
(Hunter et al., 2005).
PulseNet data analysis
The image analysis and databasing software,
currently the BioNumerics Client-Server soft-
ware (Applied Maths BVBA, Sint-Martens-
Latem, Belgium), is another key tool for
PulseNet. It is a client/server software with the
client versions installed at the participants and
the server version at CDC. It enables the par-
ticipants to compare patterns within their own
laboratories and on-line with patterns in the
National database at CDC.
The software is customized for PulseNet
through the use of scripts (macros) to adapt it to
the needs of the PulseNet participants. Entering
of demographic data is standardized. Other
scripts automate frequently used functions like
opening of forms, performing queries, and sub-
mission of PFGE patterns and accompanying in-
formation to a National server database.
A special set of scripts has been developed
for the server databases. These have the same
features as the client scripts with some addi-
tions that are specific for the server database as
described below.
Automatic pattern naming
Pattern names are essential for communica-
tion and enable the users to get a quick
overview of which patterns are common, rare,
or tend to be associated with outbreaks, a spe-
cific source, or a specific part of the country. A
PulseNet standardized pattern name consists
of 11 characters in the format: XXXYYY.####.
The first three characters (X) represent the or-
ganism (e.g., EXH is the code for STEC O157);
the next three characters (Y) represent the en-
zyme that was used to cut the DNA (e.g., X01
GERNER-SMIDT ET AL.
12

Page 5

is the code that represents the enzyme XbaI);
the four digits to the right of the decimal (#) are
consecutive numbers assigned to new profiles
as they are detected. This number ascends from
0001 and in no way indicates any kind of re-
latedness between the different types. The des-
ignation EXHX01.0001 is the first STEC O157
XbaI pattern that was named. In the first years
of PulseNet, pattern naming was done manu-
ally but the number of patterns submitted to
the National databases rapidly exceeded the ca-
pacity of the database team to name them all
in a timely manner. Therefore a script was de-
veloped to help with this function. It greatly re-
duces the workload for the database team al-
though all automatically named patterns still
need to be confirmed visually.
Outbreak code
All communication between the PulseNet
participants, the epidemiologists and other stake-
holders in outbreak investigations about pat-
terns must be precise. PulseNet database man-
agers therefore assign a unique code to all
clusters/outbreaks investigated. The basic code
is as follows: ?YY??MM??LabID??organism code?-
?number of cluster in month?. The first four dig-
its of the outbreak code indicate the year and
month in which the outbreak or cluster was de-
tected. The LabID is the two- to four-letter
PulseNet Lab identifier for the laboratory that
initially recognized the cluster or outbreak; if the
outbreak/cluster is recognized in multiple states
at once, the code is “ML.” The organism code is
the same code as used to identify the organism
in the pattern name and identifies which or-
ganism is involved in the outbreak/cluster. The
digit that follows the organism code denotes the
number of outbreak/cluster caused by that or-
ganism that month and year in that state. The
code 0510MNEHX-1 denotes the first cluster of
STEC O157 infections detected in Minnesota in
October of 2005.
WEB-BASED PULSENET TOOLS
PulseNet Listserv
The PulseNet Listserv is a closed, web-based
rapid alert and communication system for the
participating laboratories and epidemiologists.
Communications on the listserv are confiden-
tial and information posted by one participant
may not be disseminated to any third party
outside the network without permission. Par-
ticipants post information about clusters or
outbreaks of general interest to the other par-
ticipants, for example, if it is suspected that the
cluster/outbreak extends outside their state.
An outbreak posting contains summarized in-
formation about the characteristics of the out-
break strain, geographic distribution and de-
mography of infected patients, collection dates,
who is posting the message and who to contact
in case the other participants have information
relating to the cluster. A posting always in-
cludes an image of the PFGE profile of the sus-
pected outbreak pattern using at least the pri-
mary PulseNet restriction enzyme. If available,
the outbreak pattern by secondary PulseNet en-
zyme is also posted. The decision on what and
when to post is made by the participants. If the
posting laboratory has access to the national
database, it may also provide preliminary data
comparing the outbreak profile with patterns in
the national database. The database team at
CDC will confirm this information or supply it
in the cases where it has not been posted to the
listserv. Other PulseNet participants respond
to postings on the listserv indicating whether
or not they have any recent matches to the
clustered pattern. Laboratories that identify
matches to an outbreak pattern notify their lo-
cal epidemiologists to initiate an outbreak in-
vestigation. Notices of general interest such as
quarterly reports, information about upcoming
meetings and minutes of PulseNet meetings or
grant opportunities are also posted on the list-
serv. Queries and discussions about molecular
surveillance issues also take place. During 2004,
there were 178 original postings relating to dis-
ease clusters to the listserv (Salmonella 95, E. coli
55, Shigella 13, Listeria nine, Campylobacter six),
with 4131 responses (Salmonella 2565, E. coli
1256, Shigella 189, Listeria 99, Campylobacter 22)
and 116 postings on non-outbreak issues.
PulseNet web portal
Another web-based tool has been developed
for the participants of PulseNet and collabo-
PULSENET USA: A FIVE-YEAR UPDATE
13