Detection of Shigella Outbreaks in Argentina Using WHONET and SatScan

1
Detection of Shigella Outbreaks in Argentina
Using WHONET and SatScan

• John Stelling, MD, MPH
• WHO Collaborating Centre for Surveillance of
  Antimicrobial Resistance
• Brigham and Womens Hospital, Boston, USA

2
MIDAS

• NIH - Models of Infectious Disease Agent Study
• Collaboration of research and informatics groups
  to develop computational models of the
  interactions between infectious agents and their
  hosts, disease spread, prediction systems, and
  response strategies.
• Harvard Medical School
• Modeling Health System Infectious Disease Data
• Early detection and monitoring of disease
  outbreaks
• Principal investigators Richard Platt and
  Martin Kulldorff

3
Project Participants

• Syndromic and disease surveillance
• Harvard Medical School and Harvard Pilgrim Health
  Care
• Kaiser Permanente Northern California
• Massachusetts Department of Health
• Laboratory-based surveillance
• Brigham And Womens Hospital, Boston
• Instituto Nacional de Enfermedades Infecciosas
  Carlos G. Malbrán
• INEI
• Antimicrobials Service and WHONET-Argentina
• Enteric Pathogens Service
• Development, evaluation, and dissemination of
  models and tools for electronic laboratory-based
  surveillance

4
(No Transcript)
5
(No Transcript)
6
WHONETSaTScan - Analysis selection
7
WHONETSaTScan Parameter selection
8
WHONETSaTScan Integrated output
9
Outbreak Detection Strategy

• Cluster detection
• By organism
• By laboratory
• By ward/clinic
• By medical service
• By antimicrobial resistance phenotype
• By combinations of the above
• Spatial component
• None
• Latitude Longitude
• Meta-locations non-parametric location
  groupings
• Ward/Clinic
• Medical Service

10
WHONET-Argentina

• Coordinated by INEI C. Malbran
• - Antimicrobials Service
• Established in 1986
• - Laboratory capacity-building
• - Surveillance of antimicrobial resistance
• At present, 70 participating labs
• - 23 provinces federal capital
• - monthly submissions of WHONET files
• - all positive routine bacteriology results
• - confirmation of important findings at INEI
• - 39 labs included in this study

11
Shigella background information

• Four species
• S. boydii 18 serotypes
• S. dysenteriae 12 serotypes (proposed)
• S. flexneri 6 serotypes (proposed)
• S. sonnei 1 serotypes
• Food- and waterborne pathogen causing diarrhea,
  dysentery, death
• Interventions
• Epidemiological and microbiological confirmation
• Alert clinicians and public health authorities
• Active identification of cases
• Therapy
• Prevention of further transmission hygiene,
  childcare practices
• Tracing of contaminated sources
• Disinfection, removal of produce

12
Shigella spp. Antimicrobial resistance

• Increased morbidity and mortality
• In Argentina
• Resistance to NAL, CIP, NIT, and FOS is rare
  (0-2)
• Susceptibility to AMP and SXT is variable but
• S. flexneri usually nonsusceptible to AMP
• S. sonnei and S. boydii usually nonsusceptible
  to SXT
• S. dysenteriae usually susceptible to both
• Recent appearance of multidrug resistant strains
• S. flexneri TOHO-1 ESBL (first report outside
  Japan)
• Shigella spp. CMY-2-type plasmid-mediated AmpC
• Relevance to outbreak detection
• Important public health issue
• Useful marker for clone identification

13
Outbreaks reported to Ministry of Health

• MOH receives reports of Shigella outbreaks
• Reporting is not mandatory
• 6 reported in 7/2006-6/2007
• INEI Enteric Pathogens Service receives isolates
  for serotyping and PFGE-typing
• National Diarrhea and Foodborne Pathogens
  Laboratory Network
• PulseNet International Network

14
Study design and laboratory selection

• Search for outbreaks in July 2006-June 2007
• Participating labs
• 30 labs full time period
• 9 labs from Jan 2006
• Excluded labs
• 14 labs incomplete data files
• 15 labs late entry into the network
• 3 labs identification of Shigella to
  genus-level only

15
Available information

• Laboratory info LatitudeLongitude
• Patient info ID, age, sex, ward/clinic
• Sample info specimen date and type
• Microbiology results
• 1st isolate per 60-day period
• Genus, species
• Antimicrobial susceptibilities
• Measurement MIC, disk diffusion zone diameter
• Interpretation Resistant, Intermediate,
  Susceptible
• Subset Serotype, PFGE pattern

16
SaTScan parameters

• Kulldorff Space-Time Permutation Scan Statistic
• Simulated prospective analysis for July 2006-June
  2007
• Historical baseline 365 days
• 9,999 Monte Carlo Simulations
• Maximum cluster length (days to first signal)
  30 days
• Spatial variable Laboratory LatitudeLongitude
• Time variable Specimen collection date
• Threshold for alert Recurrence interval gt365
  days

17
Outbreak detection levels

• Genus Shigella sp.
• Species Each of the four species
• Susceptibility pattern by species
• Susceptible
• AMP Nonsusceptible to AMP alone
• SXT Nonsusceptible to SXT alone
• AMP-SXT Nonsusceptible to both

18
Descriptive statistics

• 2,041 isolates
• Sex
• 53 Female, 47 Male
• lt1 Unknown
• Age distribution
• 53 lt6 years, 20 6-18, 8 gt18
• 19 Unknown

S. flexneri 80 Susc 10 AMP 48 SXT 6
AMP-SXT 35 S. sonnei 17 Susc 25
AMP 6 SXT 58 AMP-SXT 11 S. boydii 2
Susc 19 AMP 0 SXT 78 AMP-SXT 4 S.
dysenteriae 0.3 Susc 67 AMP 0 SXT
17 AMP-SXT 17
19
Event 5 One laboratory in La PampaJan 2006
Jan 2007
20
SaTScan signals

• 31 statistical clusters
• Genus 5
• Species 10
• Resistance profile 16
• Grouped into 19 epidemiological events
• Resistance profiles were the most specific and
  useful
• Resistance profile missed three events
• In two cases, susceptibility testing was
  incomplete
• In one, the event was considered a false alert (a
  genus alert involving two different species,
  three different serotypes)
• Outbreak detection solely by resistance profile
  may be sufficient if susceptibility tests are
  routinely performed

21
MIDAS Shigella outbreaks in Argentina
Reported to MOH
Suggested by SaTScan
22
Outbreak A

• Outbreak details
• S. flexneri serotype 2 in San Luis October 2006
• 10 patients
• First isolate sent 11/14/2006
• 2 isolates sent - PFGE ARJZXX01.0001
• No concordance
• This laboratory joined the network in March 2006
• Results were excluded from the analyses because
  of insufficient historical data

23
Outbreak B

• Outbreak details
• S. flexneri serotype 2 in Buenos Aires December
  2006
• 4 patients
• First isolate sent 12/20/2006
• 3 isolates sent, 2 out of 3 were ARJZXX01.0002
• No concordance
• Possibly missed because of the small number of
  isolates in a large city

24
Outbreak C

• Outbreak details
• S. boydii serotype 2 in Neuquén January 2007
• 55 patients students on a field trip to Mar del
  Plata
• First isolate sent 1/17/2007
• 7 isolates sent, 6 out of 7 were ARJZGX01.0001
• Possible concordance
• Event 5
• S. boydiiSXT serotype 2 in Río Negro, 3/13/2007
  - 3/27/2007
• Date of first signal -3/17/2007
• Neighboring city in next province with a
  two-month gap
• Local epidemiologists felt that there could be a
  link given the infrequent occurrence of this
  species in their region

25
Outbreak D

• Outbreak details
• S. sonneiSXT La Pampa, Jan-Feb 2007
• Unknown number of patients
• First isolate to INEI 3/8/2007
• 12 of 19 isolates had PFGE Pattern ARJ16X01.0014
• Possibly related to produce from Mendoza
• Good concordance
• Event 5 S. sonneiSXT La Pampa,
  12/22/2006-1/31/2007
• Date of first signal 1/10/2007
• Event 10 S. sonneiSXT La Pampa, 2/21/2007
• Date of first signal 2/21/2007
• Possible concordance
• Event 7 S. sonneiSXT Mendoza,
  1/15/2007-2/20/2007
• Date of first signal 2/7/2007

26
Outbreak E

• Outbreak details
• S. sonneiSXT in San Luis March 2007
• 4 patients
• First isolate sent 4/19/2007
• 2 isolates sent, both ARJ16X01.0009. Genetically
  similar to clone from Outbreak D
• Considered by authorities to be a regional
  extension of Outbreak D
• Possible concordance
• Event 5, 7, and 10
• S. sonneiSXT, event details provided with
  Outbreak D
• Neighboring provinces with a similar time period
• This laboratory was excluded from the analyses
  because of late entry into the network

27
Outbreak F

• Outbreak details
• S. flexneriSusceptible serotype 6 Río Negro,
  April 2007
• Unknown number of patients
• First isolate to INEI 5/22/2007
• 6 isolates sent, 3 of 6 isolates have PFGE
  Pattern ARJZXX01.0004. Other patterns are
  genetically similar
• Good concordance
• Event 13
• S. flexneriSusceptible serotype 6 in Río Negro,
  3/28/2007 - 5/9/2007
• Date of first signal 4/17/2007

28
Additional SaTScan events

• 14 events were not reported to the MOH as
  outbreaks
• 2 of these were felt to be spurious findings of
  no public health importance
• Event 9 Genus-level alert with 5 isolates
• S. flexneri serotype 6, S. flexneri serotype 2,
  and S. sonnei
• Event 19 Resistance-profile alert with 2
  isolates
• Two different laboratories in Buenos Aires
• One isolate had high-level AMP resistance, while
  the other had low-level AMP resistance
• 12 were suggestive of single clone outbreaks that
  would have merited further epidemiological
  investigation and molecular subtyping

29
Prospective real-time surveillance

• Project 1
• Collaboration with three provinces with weekly
  data submissions
• Shigella and other community pathogens
• Hospital-acquired pathogens
• Project 2 Proposed
• Collaboration with SIVILA, a national web-based
  surveillance program for mandatory and optional
  case reporting of many infectious diseases and
  laboratory findings
• Project 3 Proposed
• Collaborations with other networks
• Several thematic infectious disease networks
• 25 jurisdictional infectious diseases networks

30
SIVILA
31
Conclusions

• Electronic laboratory-based surveillance
• Practical and needed addition to routine outbreak
  detection strategies
• Rich and specific level of detail when
  investigating signals
• For greatest value, should be close to real-time
• Antimicrobial resistance
• Important public health threat
• Specific marker which enhances the sensitivity
  and specificity of outbreak detection algorithms
• WHONETSaTScan
• practical and free tools which permit the use of
  advanced statistical methods for outbreak
  detections by routine clinical and public health
  laboratories