nfpa survey on l. monocytogenes in ready-to-eat foods yuhuan chen, jenny scott national food...
TRANSCRIPT
NFPA Survey on NFPA Survey on L. monocytogenes L. monocytogenes in Ready-to-Eat Foodsin Ready-to-Eat Foods
NFPA Survey on NFPA Survey on L. monocytogenes L. monocytogenes in Ready-to-Eat Foodsin Ready-to-Eat Foods
Yuhuan Chen, Jenny ScottYuhuan Chen, Jenny Scott
National Food Processors AssociationNational Food Processors Association
Research FoundationResearch Foundation
2
Publication: JFP 66(4) 2003Publication: JFP 66(4) 2003
1. Survey of Listeria monocytogenes in Ready-to-Eat Foods
David Gombas, Yuhuan Chen, Rocelle Clavero, Virginia Scott
3
Initiation of StudyInitiation of Study
• Risk management decision– Funding from nine industry partners through NFPA
– Seek strategy effective in reducing listeriosis• Is there an alternative to “zero tolerance”?
• Seek more than one strategy to solve the problem
• “Listeria monocytogenes: low levels equal low risk” JFP 66(4) Yuhuan Chen, William Ross, Virginia Scott, David
Gombas
4
Project ObjectiveProject Objective
• Develop data relative to the risk of listeriosis to support science-based strategy for addressing L. monocytogenes in foods
5
Study DesignStudy Design
• Product selection• Number of samples• Sampling site selection
– Sampling location within FoodNet sites– Selection of retail stores within counties
• Other aspects– Handling of samples, testing lab selection,
testing methods
6
Product SelectionProduct Selection
• Retail foods likely to contribute to consumer exposure– Products with relatively high prevalence– Products frequently consumed– Products not likely to be further treated
7
Product CategoriesProduct Categories
January 2000 – November 2001• Luncheon Meats (ham, bologna, poultry) • Deli Salads (potato, tuna, pasta, cole slaw)
October 2000 – November 2001• Bagged, Precut Leafy Vegetable Salads• Blue-veined Cheeses• Fresh Soft Cheeses• Soft Mold-ripened Cheeses • Seafood Salads• Smoked Seafood
8
Number of SamplesNumber of Samples
• Considerations– Prevalence thought to be low (< 5%) – Cost for enumeration high
• Assistance from FDA to estimate total number of samples needed
9
Sampling Site SelectionSampling Site Selection
• Areas where illness data are reliable
– CDC active surveillance for listeriosis in FoodNet Sites
– Case control study conducted• Selected two FoodNet sites widely separated
– northern California– Maryland
10
Estimate of Total Number of Positive SamplesEstimate of Total Number of Positive Samples
2
2 )1(
d
PPzn
n: total number of positive samplesP: percent positive for the population d: desired upper bound on absolute errorz: 1.96, 95% confidence level
11
Estimate of Number of SamplesEstimate of Number of Samples
• Assume P=50%– Given value of percentage unknown– Conservative
• n=125, 250, 500; d=8.8, 6.2, 4.4%
• Total number of samples 2500, 5000, 10,000 – assuming 5% prevalence
12
Total Number of Samples: Luncheon Meats Total Number of Samples: Luncheon Meats and Deli Saladsand Deli Salads
• Initial plan– 2500 samples per product per FoodNet site
• Two products: luncheon meats, deli salads
• Two sites: MD and CA
– 5000 samples per product– 10000 samples total
• Actual collection doubled– Prevalence considerably lower than expected
13
Sampling Locations in FoodNet SitesSampling Locations in FoodNet Sites
Northern CA Alameda and San Francisco counties
(counties in the FoodNet site)
• Maryland Five counties plus Baltimore City (FoodNet site)
Total 10 counties plus Baltimore City
All counties containing > 2% of population (listeriosis reporting statewide)
• Samples weighted by populations in county or city
14
Sampling RTE FoodsSampling RTE Foods In proportion to consumption
Luncheon meats and deli salads
Frequency of consumption within the geographical area based on CSFII West for CA, South for MD
Example, luncheon meats ham-bologna-turkey/chicken 50%-30%-20% in MD
43%-30%-27% in CA
15
Sampling RTE FoodsSampling RTE Foods 100 samples per week for luncheon meats or
deli salads 75% from List A stores (major super markets)
25% from List B stores (other grocers)
25 samples per week for each of six categories At least two List A and at least two List B stores
Supplementary lists of stores used as needed
16
FOOD SAMPLE
Enumeration
Divide into Two Portions
Screening
Composite
Negative,Stop
Composite Positive,
test individuals
If positive,proceed toEnumeratio
n
* 9-tube MPN: 1.0-0.1-0.01 g
MPN*(UVM-Fraser
or BLEB)
Direct plating(OXA or MOX)
Confirmation
L. monocytogenes isolate
31,707 Samples Tested in 2000-2001
MD CA MD+CAFresh Soft Cheese 4 (1450) 1 (1481) 5 (2931)Bagged Salads 8 (1465) 14 (1501) 22 (2966)BVSMR Cheese 7 (1473) 30 (1497) 37 (2970)Seafood Salads 88 (1225) 27 (1221) 115 (2446)Smoked Seafood 43 (1281) 71 (1363) 114 (2644)Luncheon Meats 54 (4599) 28 (4600) 82 (9199)Deli Salads 103 (4295) 99 (4256) 202 (8549)
No. Positives (No. Samples Tested)Products
18
Lessons from the StudyLessons from the Study
• Resource intensive– Three years/1.4 M – Industry support, government funding
• Team effort– Jenny Scott, David Gombas, et al. – Outside contractors for sample collection and analysis– Expert consultations
• Meetings with Agencies– Understand regulatory concerns– Seek feedback about study design and approaches
19
Lessons from the StudyLessons from the Study
• Industry concerns to be addressed– Confidentiality issues– Issues with unfavorable regulatory attention
• Not identify ham, bologna, or chicken/turkey
• Samples collected by third party
20
Lessons from the StudyLessons from the Study
• More information – Packaging location
• Store vs. manufacturer• Did not design sampling accordingly to sale or consumption
– Enumeration data
• An opportunity to leverage industry and government resources– Started study with two categories– Obtained funding from JIFSAN to collect data on six
additional categories
21
AcknowledgementsAcknowledgements
• Funding– Funding from industry partners through NFPA – USDA CSREES, FDA (JIFSAN)
• Technical assistance – Robert Blodgett– Jerome Schneidman– Wallace Garthright