NH Department of Health and Human Services

1
Salmonella Oranienburg Infections associated with
Fruit Salad Served in Healthcare
SettingsNortheastern United States, June-July
2006

• Beth Daly, MPH
• Communicable Disease Epidemiologist
• Communicable Disease Surveillance Section

2
Salmonellosis

• US gt30,000 infections reported each year
• NH 2nd most common bacterial foodborne illness
  with 150-200 cases each year
• gt2500 different serotypes
• Poultry and cattle serve as reservoirs
• Source of outbreaks include meat, produce, ill
  food service workers, etc.
• Primarily diarrheal illness
• Illness lasts 2-5 days
• Incubation period 6-72 hours

3
Salmonella Oranienburg

• 13th most common serotype of Salmonella reported
  to CDC
• 495 cases reported in 2004 (1.4)
• 9 outbreaks reported to CDC since 1996
• Outbreaks associated with meats and produce
• Ground beef, chicken, pizza, mango

4
Salmonella Surveillance

• Mandated reporting of salmonellosis in NH
• Reports received from HCP, ICP, laboratories
• For each case reported there are an estimated 38
  cases that go unreported

5
(No Transcript)
6
Pulsed Field Gel Electrophoresis

• Isolated bacteria from each case used
• DNA from the bacteria is restricted by an enzyme
  and undergoes electophoresis to produce a pattern
  or DNA fingerprint
• Patterns compared to determine if isolated
  organisms are different, similar, or
  indistinguishable
• Two-enzyme approach used

7
PFGE Analysis
8
Use of PFGE in Surveillance

• NH PHL performs PFGE on all isolates of
• Campylobacter
• Salmonella
• Shigella
• Listeria
• E. coli O157H7
• These patterns are entered into a national and
  state database to look for matches
• This allows for recognize of local and
  multi-state outbreaks

9
Nosocomial Salmonellosis

• Uncommon in developed countries
• Healthcare facilities (HCF) dont test for
  salmonellosis in patients hospitalized gt72 hours
  with diarrhea
• C. diff the most common cause of hospital
  acquired diarrhea
• Since 1960, 56 outbreaks have been described
• Most often associated with food, feed, visitors,
  staff
• 9 occurred in the US, most recent in 2002 and 1996

10
NH Outbreak

• July 19th NH DHHS investigated a salmonellosis
  outbreak at a local hospital
• Initially 5 cases identified
• 2 patients, 2 HCP, 1 cafeteria patron
• 3 of 5 cases determined to be S. Oranienburg
• Concerns about medical devices, ill HCP, or ill
  FSW
• Challenging disease control approach

11
Immediate Infection Control

• Active surveillance and exclusion among staff in
  affected units
• Shift change sign-in
• Mandatory in-service training to staff
• Alert hospital-wide email
• Inspection of hospital kitchen
• Routine salmonellosis testing for patients with
  diarrhea

12
Multistate Outbreak Recognition

• July 19th NH DHHS and MHD investigated outbreak
  of S. Oranienburg at local hospital
• July 21st MA DPH reported S. Oranienburg cases
  at local LTCF
• NH cases and other northeastern states report
  cases with the same PFGE pattern as MA cases
• Xbal pattern uncommon and seen only 14 times
  prior to this cluster
• BlnI pattern not seen before

13
Frequency of BlnI Patterns in S. Oranienburg with
the Xbal JJXX01.0056 Pattern, 1998-2006
14
Methods- Case Finding

• Case Definition culture-confirmed cases of S.
  Oranienburg with
• Illness onset after June 1st
• Xbal pattern JJXX01.0056 and Blnl pattern
  JJA26.0017 if available
• PulseNet queried weekly
• Announcements on listserves foodborne outbreaks,
  promed mail, SHEA, APIC, EIN
• Active case finding in affected facilities

15
Methods- Hypothesis Generation

• Routine surveillance questionnaires reviewed
• Extended questionnaires administered to small
  subset of cases
• 300 exposures including 234 food items
• Facilities with cases surveyed to determine brand
  and distributor information for fruit salad
  served in the facility
• 26 unaffected facilities in NH surveyed for
  comparison

16
Methods- Case Control Study

• Case control study conducted between August 15th
  and September 6th
• Cases were eligible if
• Onset between June 15th and July 31st
• Experienced diarrhea
• Two enzyme match
• Could be interviewed
• Control selection was based on type of case and
  eligible if
• No diarrhea since June 1st
• Must be on a solid diet

17
Control Selection

• Patients exposed in HCF
• 7 day period before onset calculated
• Dates of hospitalization during these 7 days
• List of patients hospitalized on same days
  created
• List randomized and selected until 3 controls
  found
• Healthcare Employees
• 7 day period before onset calculated
• Dates case worked during these 7 days
• List of coworkers on same days created
• List randomized and selected until 3 controls
  found
• Community Cases
• List of neighbors generated using reverse phone
• Closest neighbors called until 3 controls found

18
Results- Geographic Distribution

• 42 cases in 10 states
• and Canada
• MA (12)
• NH (10)
• NY (4)
• PA (3)
• VT (3)
• Canada (2)
• KY (2)
• MD (2)
• ME (2)
• CT (1)
• NJ (1)

19
Results- Case Details

• Age 8 months to 96 years (median59) 31 over
  age 70
• Sex 28 females (67)
• Healthcare Relationship
• 21 hospitalized patients or LTCF residents
• 9 healthcare employees
• 1 case who ate in hospital cafeteria
• Illness onset dates ranged from June 15th to July
  25th

20
Culture Confirmed Cases of Salmonella Oranienburg
by Date of Onset (n22)
21
Results- Hypothesis Generation

• Interviews showed a high proportion of cases
  consumed fruit salad in healthcare facilities
• 23 of 33 cases (70) consumed fruit salad
• 19 of 23 consumed Brand X fruit in a HCF
• 10 (91) of 11 facilities with cases served Brand
  X fruit salad from Ontario, Canada
• Cantaloupe and honeydew melon
• 3 (15) of 20 control facilities served Brand X
  (OR57, p-value0.0005)

22
Results- Case Control Study

• Illness significantly associated with fruit salad
  consumption
• Any Fruit Salad OR 8.9, 95 CI 2.3 - 35.5
• Fruit Salad in a HCF OR 6.0, 95 CI 1.5 - 23.5
• Many fruit salad components also significant
• Significant correlations between many individual
  components
• Multivariate analysis not feasible
• Specific component not statistically implicated

23
Results- Matched Univariate
Food Item Cases (n21) Cases (n21) Controls (n33) Controls (n33) Matched OR 95 CI
Food Item N N Matched OR 95 CI
Any Fruit Salad 14/20 70 4/30 13 8.9 2.3-35.5
Fruit Salad in a Healthcare Facility 12/20 60 4/30 13 6.0 1.5-23.5
Cantaloupe in Fruit Salad 12/20 60 2/30 7 9.9 2.2-44.5
Honeydew in Fruit Salad 11/20 55 1/30 3 16.9 2.4-119.6
Watermelon in Fruit Salad 9/19 47 2/30 7 6.9 1.4-33.7
Pineapple in Fruit Salad 8/19 42 1/29 4 40.6 3.0-548.4
Red Grapes in Fruit Salad 7/20 35 0/30 0 9.8 1.5-65.6
Uses a 0.5 continuity correction
24
Traceback and Product Investigation

• Several recent outbreaks of Salmonella associated
  with cantaloupe and honeydew
• gt 25 since 1984, 16 since 1998
• 1998 outbreak of S. Oranienburg in Ontario Canada
  associated with cantaloupe
• FDA and CFIA conducted traceback and traceforward
  investigations on August 24th
• Brand X Canadian facility inspected
• Cantaloupe and honeydew from California
• Unable to traceback to farm due to large number
  of supplying brokers

25
Challenges

• Affected population
• Other significant medical issues
• Age
• Mental status
• Distinguishing between primary and secondary
  cases
• Infection control
• Identification of the source within the fruit
  salad components

26
Conclusions

• Interventions on farm may prevent
  produce-associated outbreaks
• Nosocomial salmonellosis rare in US
• Recognition is difficult due to stool testing
  practices in hospitals
• May be difficult to distinguish between primary
  and secondary cases
• Foodborne outbreaks in healthcare settings can
  cause significant illness, costs, fear

27
Acknowledgements

• Hospital ICP
• NH Disease Control and Laboratory Staff
• Collaborating States
• MA, NY, PA, VT, KY, MD, ME, CT, NJ
• CDC
• Dr. Christine Olson, MD, MPH, EISO
• Dr. Michael Lynch, MD, MPH