1. Whole Genome Sequencing and its Application to Foodborne Outbreak Investigation
Bryna Warshawsky, Public Health Ontario
25th Annual FFIGS Educational Workshop
Emerging Food Trends and Their Safety… 25 Years and Counting
May 2, 2018

2. Objectives and outline
Review outbreak investigation
Describe whole genome sequencing
Describe situations where whole genome sequencing has impacted outbreak investigations
The purpose of today’s presentation is to discuss three S. Enteritidis outbreak investigations in 2014 and 2015 linked to uncooked, frozen, processed chicken products that were described in a recent publication.
We will also describe two additional outbreaks also linked to linked to uncooked, frozen, processed chicken products that occurred since the paper’s publication.
We will highlight the important role Ontario public health units played in detecting and investigating these outbreaks.
We will also highlight the importance of government and industry policies related to uncooked, frozen, processed chicken products to prevent contamination at the source level and infection at the consumer level.

3. Tools and approaches used in foodborne outbreak investigations
Epidemiological investigation
Questionnaires
Single interviewer approach
Loyalty card verification of purchases
Foodbook analyses
Food safety investigation
Traceback
Laboratory investigation
Testing of food from cases’ homes
Subtyping
Describe which occur at the local, provincial, federal levels
Three pillars of foodborne outbreak investigations.
Image Source: Microsoft Word 2013

4. Laboratory subtyping: Identifying strains that are the same or different
In outbreak investigations, subtyping allows us to differentiate the outbreak cases from all the background/baseline non-outbreak cases
Subtyping techniques with high discriminatory power allow us to identify only those cases that are related epidemiologically/biologically
Epidemiological/biological relatedness indicates that the cases have something in common, i.e. a source of illness or exposure
And that is what we want to find out; what common exposure did all these cases have that made them sick
___________________________________________________________________________________
Note:
For techniques with low discriminatory power (e.g. serotyping), epidemiological/biologically unrelated isolates are subtyped as the same
For techniques with high discriminatory power (e.g. WGS), epidemiological/biological similar isolates are subtyped as the same/related

5. Subtyping with increasing discriminatory power
Whole Genome Sequencing (WGS)
Pulsed-Field Gel Electrophoresis (PFGE)
Phage Type (PT)
This slide lists some of the available subtyping techniques in order of increasing discriminatory power, with genotyping techniques being the most discriminatory.
The order is probably debatable and what is available/useful varies by pathogen.
Starting with the least discriminatory, subtyping techniques include serotyping, phage typing…
______________________________________________________
Serotyping
Category based on immune cell response to infection (i.e. in terms if the antigen it contains or the antibodies produced against it)
Phage typing
The pattern by which a series of viruses (bacteriophages) lyse the bacterial strain
Pulse field gel electrophoresis
Digest DNA with enzymes (usually two enzymes)
Separate out pieces by size using electrophoresis on a gel
Patterns compared
MLVA
Conserved area of genome, calculate number of repeats
MLVA above PFGE because often used to differentiate organisms with similar PFGE pattern combinations
Genetic sequencing
Whole or part of the genome
Serotype
Image Source: Microsoft Word 2013

6. Pulsed-Field Gel Electrophoresis (PFGE)14 15 1 2 3 4 5 6 7 8 9 10 11 12 13 1 1
(accessed January 7, 2016)

7. Whole Genome Sequencing – Definitions
laboratory procedure that determines the order or sequence of bases in the genome/DNA of an organism
Bases - Adenine, Thymine, Cytosine and Guanine
DNA is composed of nucleotides which are composed of one of four bases
Gene
sequence of DNA that encodes a common trait or function of the organism
Allele
variant of a gene (one or more bases are changed)
Before moving onto the detailed analysis it is important to review the definitions of the following terms as used in whole genome sequencing.

8. Whole Genome Sequencing - Data Analysis
Whole genome sequencing analyzes the genetic sequences
Takes advantage of the fact that strains from the same origin will have very similar genetic sequences
Two methods have been used in Canada
Single Nucleotide Variant (SNV)
Whole Genome Multi Locus Sequence Typing (wgMLST)
Both methods compare the genetic sequences of the bases from different perspectives
Two biocomputational methods have been used in Canada for performing the data analysis and comparisons of nucleotide sequences.
The first method is called Single Nucleotide Variant or SNV analysis. This has also been called Single Nucleotide Polymorphism or SNP analysis. Sometimes the acronyms are used interchangeably however the Canadian standard through the National Microbiology Laboratory in Winnipeg and PulseNet Canada is to use single nucleotide variant or SNV terminology.
The biocomputational tool that is used for SNV analysis is called the SNVPhyl pipeline
The second method used to analyze sequencing data is called whole genome Multi Locus Sequence Typing or wgMLST and the biocomputational tool that is used is a commercial platform called BioNumerics
Both methods compare the genetic sequences of the bases from different perspectives
* Single Nucleotide Variant PHYLogenomics

9. Whole Genome Sequencing – SNV vs. wgMLST
wgMLST analysis
Differences in genes/alleles
Used by the PulseNet Canada program through the National Microbiology Laboratory in Winnipeg for surveillance and outbreak support
Utilizes a commercial program (BioNumerics) to perform standardized data analysis between all provincial and federal laboratories
SNV analysis
Differences in bases
May be used if further laboratory evidence is required to support an investigation by providing additional genetic resolution
How are these two analytical methods currently being used……..
…….utilizes the commercial program, BioNumerics, to perform standardized data analysis between all provincial and federal laboratories.
This ensures WGS analyses performed at all PulseNet Canada laboratories produce results that are standardized and can be compared with each other
The wgMLST analysis also includes a large set of genes and therefore provides greater discrimination between isolates
Although wgMLST is routinely used the SNV analysis

10. Whole Genome Sequencing (WGS) – Implementation
Listeria monocytogenes
WGS has been performed on all isolates at the National Microbiology Laboratory since January 2017
Salmonella spp.
WGS has been performed on all isolates at the National Microbiology Laboratory since May 2017
E. coli STEC
WGS is proposed to be initiated sometime in 2018
For Listeria monocytogenes……

11. Whole Genome Sequencing – Interpretation Guidelines
The following PHAC/PulseNet Canada guidelines are being used to interpret the relatedness of isolates
Organism
SNV
differences
wgMLST / allele
Salmonella spp.
0 - 10
Listeria monocytogenes
It is also recommended to assess isolates within the 11 – 25 difference range
to compare how they relate to the 0 – 10 group
review for consistent epidemiological information
For both Salmonella and Listeria the range of 0 – 10 SNV or allele difference would indicate from a laboratory perspective that the isolates are considered to be related and require further epidemiological investigation.
It is however also recommended to assess isolates within the 11 – 25 difference range. If applicable, isolates should be compared to see how they relate to the 0 – 10 difference group and also reviewed for consistent epidemiological information.
For surveillance purposes, when WGS is performed, isolates are normally screened for any recent matches. For Salmonella the time frame for review is the last 60 days and for Listeria monocytogenes it would be the last 120 days.

12. Whole Genome Sequencing – Phylogenetic Trees
SNV or wgMLST phylogenetic tree will be very similar in presentation and include scale, branches, nodes
Scale – provides a measure of the relatedness of the isolates included in the analysis
Branches – horizontal lines show the degree of genetic change between isolates – the longer the branch the larger the amount of change
Nodes – a number is associated with a node which provides the number of SNVs or alleles difference between the two branches
The final step for the laboratory is to prepare a comparison of the isolates for surveillance or outbreak purposes.
A phylogenetic tree is a diagram that represents the evolutionary relationships among the organisms being assessed.
Both the SNV or wgMLST phylogenetic tree will be very similar in presentation and will include a scale, branches, nodes
Scale – provides a measure of the relatedness of the isolates included in the analysis – if the scale is large it means the isolates are less likely to be related and if the scale is smaller then there may be a closer relationship
Branches – are the horizontal lines which show the degree of genetic change between isolates – similar to the scale, the longer the branch, the larger the amount of change
The final item are the nodes – the nodes normally will have a number associated with it. The number indicates the number of SNVs or alleles difference between the two branches.
As a note: In this example the numbers are provided with a decimal point. The decimal point is a feature of BioNumerics and normally for reports the decimal points would be rounded up or down (0.5 goes up) and would be presented as whole numbers

13. As of Nov 14, there are 27 pages of this Salmonella tree
As of Nov 14, there are 27 pages of this Salmonella tree. This is a truncated page 1.

14. Utility of whole genome sequencing
Identifies what belongs together
Cases and food
Example: S. Enteritidis and frozen processed chicken
Identifies what is not a cluster
Examples: Listeria reports; E. coli in Quebec and Ontario
Identifies when there is a cluster but some cases or foods are not a part of it
Examples: Listeria in 2018, S. Enteritidis and frozen processed chicken
May not always be totally helpful
Examples: Listeria and prepackaged salad in 2016

15. Cluster 1 Cluster 2 Cluster 3
Mid – January 2016 Three different Listeria clusters identified by PFGE
Rare PFGE pattern
6 cases in Ontario only
Nov 2015 – Jan 2016
Cluster 1
7 cases in 5 provinces; are in Ontario
Sept 2015 – Jan 2016
Cluster 2
4 cases in Ontario
May-Dec 2015
Cluster 3
Mid-January, 2016
3 clusters of different Listeria PFGE patterns
Cluster 1 - rare PFGE pattern
6 cases; only in Ontario
Mid-November, 2015 to early January, 2016
Cluster 2
3 cases in Ontario and one case in each of New Brunswick, Quebec, Prince Edward Island, and Newfoundland and Labrador
September, 2015 to early January, 2016
Cluster 3
4 cases in Ontario from May to December, 2015

16. Salmonella Enteritidis and frozen processed chicken

17. Uncooked, frozen, processed chicken products
What are uncooked frozen, processed, chicken products?
Includes formed chicken products, such as nuggets, strips, and burgers, that are often breaded and appear fully cooked.
Image Source: Bryna Warshawsky, London, Ontario; November 19, 2017

18. Open access publication in the Journal of Food Protection1
The publication was included in the April 2017 issue of the Journal of Food protection.
Open access.

19. Routine WGS for all Salmonella isolates in Canada
Salmonella Enteritidis outbreaks in Canada linked to uncooked, frozen, processed chicken
Routine WGS for all Salmonella isolates in Canada
Outbreak 1
Outbreak 2
Outbreak 3
Outbreak 4
Outbreak 5
Will now discuss S. Enteritidis outbreaks in Canada linked to uncooked, uncooked, frozen, processed chicken.
In recent years, a total of five S. Enteritidis outbreak investigations have been linked to uncooked, frozen, processed chicken products.
The first three investigations occurred in 2014 and 2015, and are described in the paper.
Since the paper’s publication there have been two additional outbreaks, which will also be discussed.
The most recent outbreaks (outbreaks 4 and 5) were identified following the implementation of whole genome sequencing in May 2017.
Since May, all Salmonella isolates from across the country are being tested by the National Microbiology Laboratory.
This more discriminatory testing method has further illuminated the problem related to uncooked, frozen, processed chicken by detecting clusters of illness that would have otherwise been difficult to link.
Jun 2015
Nov 2014
Sept 2015
May 2017
Jun 2017
Sept 2017

20. Outbreak 1: Ontario, declared November 2014
Increase in domestically-acquired S. Enteritidis cases
Phage types 8 and 13
Uncooked, frozen, processed chicken products were identified as an important risk factor
45% of domestically-acquired S. Enteritidis cases reported uncooked, frozen, processed chicken products
19% for the Ontario population (Foodbook data)
In November 2014, a sustained increase in the number of reported domestically acquired S. Enteritidis cases (primarily among PT 8 and 13) in Ontario led to an outbreak investigation. 
Epidemiological analyses identified uncooked, frozen, processed chicken products as an important risk factor contributing to the observed increase.
About 45% of S. Enteritidis cases reported consuming uncooked, frozen, processed chicken. This was significantly higher than the expected frequency of 19% for store-bought breaded chicken for the Ontario population, based on data from Foodbook, which Bryan described earlier.

21. Outbreak 1: Case and food samples matchesPT
PFGE
WGS
Case 1
A (23 Jul 2014) 13
.0038/.0016
Sequence A
Case 2
B (7 Aug 2014) 8
.0003/.0292
Sequence B
Case 3
B (11 Sep 2014)
13a
.0006/.0007
Sequence C
Brand (production date)
Case laboratory and food testing result
As part of the investigation, PHUs collected uncooked, frozen, processed chicken samples from outbreak cases’ homes and submitted those samples to PHOL for testing.
Collecting food samples from cases’ homes, then using subtyping methods to link the food sample to the respective case is a very important step in gathering evidence to implicate a source in an investigation, and one where PHUs play a key role.
In this outbreak, as shown in the table, three of the food samples submitted matched the clinical case by PT, PFGE. Whole genome sequencing was used to further characterized the three matches, and WGS results aligned with PT and PFGE results indicated the cases were related to their respective food samples.
This was one of the first investigations where WGS was used to support PT and PFGE results.
While the food samples matched the respective cases, they differed from each other.
This finding is not unexpected as the samples included two different brand names, and were produced on three different dates.
So, to summarize, the cases matched the food, but not each other.

22. Outbreak 1: Traceback of three positive food samples submitted by three different cases
Brand A food sample 1
(23 Jul 2014)
Brand B food sample 1
(7 Aug 2014)
Brand B food sample 2
(11 Sep 2014)
As part of the investigation, PHUs collected uncooked, frozen, processed chicken samples from outbreak cases’ homes and submitted those samples to PHOL for testing.
Collecting food samples from cases’ homes, then using subtyping methods to link the food sample to the respective case is a very important step in gathering evidence to implicate a source in an investigation, and one where PHUs play a key role.
In this outbreak, as shown in the table, three of the food samples submitted matched the clinical case. Whole genome sequencing was used to further characterized the three matches, and WGS indicated the cases were related to their respective food samples.
This was one of the first investigations where WGS was used to support further characterize case and food sample matches.
While the food samples matched the respective cases, they differed from each other.
This finding is not unexpected as the samples included two different brand names, and were produced on three different dates.
So, to summarize, the cases matched the food, but not each other.

23. Outbreak 1: Traceback of three positive food samples submitted by three different cases
Brand A food sample 1
(23 Jul 2014)
Brand B food sample 1
(7 Aug 2014)
Establishment A
Brand B food sample 2
(11 Sep 2014)
Traceback activities, as described by Bryna, were initiated as a result of the investigation.
Traceback activities found that the three food samples – which included two different brand names - were produced at the same establishment, establishment A. This was the first time in recent years where this establishment was implicated in an investigation and was an interesting converge.

24. Outbreak 2: Canada, declared June 2015
Related S. Enteritidis PFGE patterns not previously reported in Canada
SENXAI.0257, SENXAI.0259, and SENXAI.0266
51 outbreak cases (35 from Ontario)
Ontario case onset dates: Jan 15 to Jul 2, 2015
42% of Ontario outbreak cases reported consumption of uncooked, frozen, processed chicken products
19% for the Ontario population (Foodbook data)
In June 2015, there was a second outbreak investigation, which was national in scope.
The investigation was triggered by the identification of three closely related PFGE patterns that had not been previously reported in Canada.
Of the 51 cases reported, 35 were from Ontario.
The onset dates for Ontario cases ranged from January 15 to July 2, This is quite a wide range for case onset dates, but it is also not unexpected with this type of frozen product where the best before date is one year after the production date.
Similar to the previous investigation, a significantly higher than expected proportion (42%) of Ontario outbreak cases reported consumption of uncooked, frozen, processed chicken products compared with the expected frequency (19%) for the Ontario population.

25. Enhanced surveillance: Focused questionnaire and food sample collection
We learned from the previous investigation and early on implemented enhanced surveillance for all Salmonella cases reporting consumption of uncooked, frozen, processed chicken products.
As part of the enhanced surveillance, public health units used a focused questionnaire (which is shown on the slide) that collected specific uncooked, frozen, processed chicken product details (such as brand name, etc.) and directed public health unit to inquire about possible food samples collection from cases’ homes.
Again, the specific brand information reported and food samples collected by public health units during this process were key to implicating a specific uncooked, frozen, processed chicken product.
Ontario Agency for Health Protection and Promotion (Public Health Ontario); Frozen Processed Chicken Focused Questionnaire, 2015.

26. Outbreak 2: Case and food samples matches
D (22 Jan 2015)
.0259/
Yes
Non-Ontario case 1
C (22 Jan 2015)
.0257/
Non-Ontario case 2
C (unknown)
Case 1
Case laboratory and
food testing PFGE result
Match to outbreak
PFGE pattern
Brand (production date)
One Ontario public health unit collected two brand D samples of uncooked, frozen, processed product from their case and submitted those samples for testing.
As shown in the table, both food sample isolates were a PFGE match to the clinical case.
In addition, two brand C food samples were collected from two cases from two other provinces, and matched the one of the three outbreak PFGE patterns.
In this investigation, unlike the first, the cases and their respective food samples were related to each other by PFGE, likely due to the common Jan 22 production date across the different food samples.
Brand D and C are made by Establishment A

27. Outbreak 27
Due to the strong epidemiological and food investigation linkages of cases to a specific products, a warning was issued for a voluntary food recall by the Canadian Food Inspection Agency for brand C and D products produced on 22 January 2015, which resulted in the products being withdrawn from retail and the issuance of public warnings to avoid consuming the product.
This was the first time this type of product was recalled in Canada, and was made possible by the work of all local, provincial, and federal partners involved in the investigation.

28. Week outbreak was declared
Outbreak-confirmed S. Enteritidis cases by episode week, Ontario, January 1 to July 2, 2015
Week outbreak was declared
Week of production
Food recall warning
Figure shows the 35 Ontario outbreak confirmed cases by week of onset and PFGE pattern.
The arrow indicates the week in January the implicated product was produced.
Aside from the first case, the figure illustrates how the product availability on the market aligns with cases’ onset dates - three week following the January 22 production date we start to see cases.
The second arrow indicates when the outbreak was declared, and due to the challenges in identifying outbreak was a number of weeks after the earliest cases.
The third arrow indicates the week the food recall warning was issues, which occurred within a few weeks of the outbreak being declared.
Data source: Ontario Ministry of Health and Long-Term Care, integrated Public Health Information System (iPHIS) database, extracted by Public Health Ontario [2015/08/20]
Episode date is an estimate of the onset date of disease for a case. In order to determine this date, the following hierarchy is in place in iPHIS: Onset Date > Specimen Collection Date > Reported Date.

29. Outbreak 3: Ontario, declared September 2015
Increase in domestically-acquired S. Enteritidis PT13a cases
36 PT13a outbreak cases
Onset dates: Aug 5 to Oct 13, 2015
59% of outbreak cases reported consumption of uncooked, frozen, processed chicken products
13% the Ontario population from Aug to Oct (Foodbook data)
Also noted a number of S. Enteritidis PT8 cases (non- outbreak cases)
In September 2015, a sustained increase in the number of domestically acquired S. Enteritidis PT13a cases in Ontario led to a third outbreak investigation, which identified a total of 36 PT13a cases. Onset dates ranged from August 5 to October 13, 2015
Similar to the previous two investigations, a significantly higher proportion (59%) of Ontario outbreak cases reported consumption of uncooked, frozen, processed chicken products compared with the expected frequency (13%) for the Ontario population for the months of August to October, inclusive (corresponding to the range of outbreak case onset dates).
Similar to the second investigation, enhanced surveillance was implemented and public health units interviewed cases with a the same focused questionnaire used in the second outbreak investigations, and collected food samples from cases’ homes, if available.

30. Isolates from cases and food submitted by cases
Cluster 2
Isolates from cases and food submitted by cases
Cluster 1
Cluster 1
Cluster 1

31. Outbreak 3: Food samples submitted by outbreak S
Outbreak 3: Food samples submitted by outbreak S. Enteritidis cases Cases were PT13a, PFGE , WGS cluster 1 PT
PFGE
WGS
Case 1
C (9 Jul 2015) A
13a
.0006/.0007
Cluster 1
Case 2 19
Case 3
Brand (production date)
Food testing result
Establishment
Ontario public health units collected three brand C food samples, again, all of which where were produced at establishment A, from three outbreak cases, and submitted those samples for testing.
As shown in the table, all three of the samples had the same production date of 9 July and were a match to the clinical isolates by PFGE, and WGS results supported the relatedness their relatedness to the respective clinical isolates (PT matched for two of the three cases).
WGS also supported the relatedness of the cases and their respective food sample matches with each other; all cases and food isolates were all in cluster 1.
Similar to the second investigation, the relatedness of the food isolated to each other was likely due to the common Jul 9 production date. A
Case 4
C (unknown) 13
.0062/.00016
No cluster

32. Outbreak 3:
Food samples submitted by non-outbreak S. Enteritidis cases Cases were PT8, PFGE , WGS cluster 2 PT
PFGE
WGS
Case 1
C (9 Jul 2015) A
8, 19
.0003/.0003
Cluster 1
Case 2 8
.0006/.0007
.0007/.0005
Cluster 2
Case 3
B (unknown)
Case 4
B (6 Jul 2015)
Brand (production date)
Establishment
food testing result
.0007/.0005
.0006/.0007
.0225/.0003
While this third outbreak investigation focus on PT13a cases, as part of the enhanced surveillance implemented, public health units collected additional food samples from other S. Enteritidis cases that reported consumption of uncooked, frozen, processed chicken products.
As shown in the table, four additional samples were collected.
Two of the samples were the same brand C sample with the Jul 9 production date that was collected for the outbreak cases.
The remaining two samples were brand B product, only one of what had a production date available.
Again, all of the samples were produced at establishment A.
All four of the samples matched respective case by at least two of the three subtyping methods used, so by PT, PFGE, and/or WGS.

33. Outbreak 3: Same product, two subtypes; Same cluster, two products
Food samples from four non-outbreak PT8 cases, cluster 2
Food samples from three PT13a outbreak cases, cluster 1
Brand C 9 Jul 2015
Cluster 1
Brand C 9 Jul 2015
Cluster 1
Brand C 9 Jul 2015
Cluster 1 and Cluster 2
Brand C Jul 2015
Cluster 1
Brand C 9 Jul 2015
Cluster 1
Brand B unknown
Cluster 2
Since the same brand C product with the same Jul 9 production date was linked to both PT13a outbreak cases as well as the non-outbreak PT8 cases, we learned that the same product can be contaminated multiple Salmonella subtypes.
Brand B 6 Jul 2015
Cluster 2

34. Cluster 2
6 non-outbreak cases (PT8)
Brand C 9 July 2015
Brand B 6 July 2015
4 outbreak cases (PT13a) not part of cluster 1
One part of cluster 2
Cluster 1
18 outbreak cases (13a)
Brand C 9 July 2015

35. Outbreak 3: Lessons learned
Despite the strong epidemiological and food investigation linkages warning for a food recall warning was not issued, although we would have liked to have seen one.
Despite recall in second outbreak, we were still unclear on process required for recall for this type of product
As we were trying to work though the process, no new cases had been reported, and outbreak appeared to be over.
However, we did learn from this experience.
We better understood the process required to implicate a product for recall warning, which paved the way for recalls in later outbreaks.
Reinforced the pattern of traceback to establishment A.
And in order to help us put these pieces together, we summarize the evidence to date implicating uncooked, frozen, processed chicken, which in our investigations all traced back to establishment A, and submitted the paper for publication.

36. Outbreak 4: Canada, declared June 2017
13 S. Enteritidis cases related by WGS
Five from Ontario
Ontario case onset dates: May 3 to Jun 2, 2017
Significantly higher than expected proportion of outbreak cases reported consumption of uncooked, frozen, processed chicken products
Investigation implicated brand E uncooked, frozen, processed chicken product
Since the paper’s publication, additional outbreaks have been linked to uncooked, frozen, processed chicken, both of which lead to the issuance of food recall warnings.
As mentioned, these more recent outbreaks were identified following the implementation of routine whole genome sequencing for all Salmonalla isolates as of May 2017.
A fourth national outbreak occurred in June 2017.
13 cases were identified by WGS, including five from Ontario.
A significantly higher than expected proportion of outbreak cases reported consumption of uncooked, frozen, processed chicken.
Investigation implicated brand E uncooked, frozen, processed chicken product produced at establishment A.

37. Outbreak 4 7
Due to the strong epidemiological and food investigation linkages of cases to a specific product, a warning was issued for a voluntary food recall for the implicated product.

38. Outbreak 5: Canada, declared September 2017
Ongoing investigation
22 S. Enteritidis cases in two WGS clusters (12 from Ontario)
Ontario case onset dates: Jul 2 to Oct 13, 2017
The two clusters are not related by WGS, but are linked by a common product
Investigation implicated Brand B uncooked, frozen, processed chicken product
A fifth national outbreak investigation started in September 2017, and this investigation is ongoing.
The investigation includes 22 S. Enteritidis cases from two separate WGS clusters, of which 12 are from Ontario.
The two WGS clusters included in the investigation are considered unrelated, but are are linked by a common product. 
The same uncooked, frozen, processed chicken product (brand B produced on the same date by establishment A) were submitted for testing by a case from another province from cluster 1 and an Ontario case from cluster 2.
Whole genome sequencing results of products provided by the cases matched the respective clinical cases, but not each other.

39. Outbreak 5: Summary of WGS results
Cluster 2
Cluster 1
10 outbreak-confirmed cases
Brand B
food samples
(12 May 2017)
2 outbreak-confirmed cases
Two unrelated WGS clusters, with outbreak cases in both clusters, with the same product (same brand produced on the same date by establishment A) included in both clusters
As demonstrated in the third outbreak investigation, findings further indicated the same product can be contaminated multiple Salmonella subtypes.

40. Outbreak 5 7
Due to the strong epidemiological and food investigation linkages of cases to a specific product, a warning was issued for a voluntary food recall for the implicated product.
As well as an associated product that shared common ingredients.

41. Summary: Evidence from Canadian outbreaks
Est. A
Outbreak 1
Outbreak 5
Outbreak 4
Outbreak 3
Outbreak 2
To summarize the information presented;
Ontario cases were implicated in five outbreak investigations in Canada that linked human illness to uncooked, frozen, processed chicken products.
Evidence from the five Canadian investigations implicated products from the same establishment, establishment A, which suggests a common root cause for all five outbreaks.
In three of these investigations, strong epidemiological and food investigations linkages implicated specific lots of products and lead to voluntary recall warnings for the implicated product.
Implicating a source in these outbreaks has in a large part been facilitated by two huge improvements in our ability to investigate case clusters, and includes the use of;
PHAC’s Foodbook data to compare foods eaten by outbreak cases to the general population. This depends on public health units collecting comprehensive food histories during case interviews through the implementation of standardized questionnaires.
And the second is WGS, which has allowed us to detect clusters that would have otherwise been difficult to link, as well as more definitively link cases to food samples, collected by public health units.

42. WGS and cluster detection
Routine WGS for all Salmonella isolates in Canada
Outbreak 1
Outbreak 2
Outbreak 3
Cluster 2
Cluster 1
Outbreak 4
Outbreak 5
Cluster 3
Cluster 4
Cluster 5
Cluster 6…
. The implementation of routine WGS by the NML, IN additional to the OB that have been identified there have been numerous other clusters that have suggested a linke to FPCP, but to date link had not been strong wnought to declare an outbraek.
Jun 2015
Nov 2014
Sept 2015
May 2017
Jun 2017
Sept 2017

43. Challenges and limitations of investigations
Under-reporting of Salmonella cases
For every one reported case, there are 13 to 37 other cases6
Previous subtyping methods likely missed clusters/outbreaks of S. Enteritidis
Long turn around times for WGS results lead to delays in identifying clusters/outbreaks
Delays in follow-up result in:
Greater difficulty for cases to remember the foods they ate
Food not available for testing or missing product identifiers
Reported cases represent only a small fraction of the number of people with Salmonella infection; therefore, cases, and potentially outbreaks due to uncooked, frozen, processed chicken, may go unrecognized. 13 to 37 cases go unreported
Information collected during outbreak investigations is limited by the cases’ ability to recall relevant exposures because cases are often interviewed up to 2 weeks or more after symptom onset. Furthermore, due to delays in reporting, food samples were not available for many cases. Thus, brand and lot information can not be necessarily verified, samples can not necessarily be tested.
Reporting delays have increased with the implementations of WGS. There is a long turn-around time for obtaining WGS results, often several weeks after submission.
So identifying uncooked, frozen, processed chicken in the first interview is important.
In addition, laboratory findings in the investigations identified multiple PT and PFGE patterns and/or WGS results in food sample isolates tested, demonstrating that processed chicken products can be contaminated with multiple Salmonella subtypes, which presents a challenge in identifying case and food sample matches.
The market share of establishment A is not known. It is possible that establishment A produces the majority of uncooked, frozen, processed chicken products available in Ontario and, thus, would be more frequently identified in outbreak investigations.

44. Why does uncooked, frozen, processed chicken pose a risk?
Consumer handling
Manufacturing process

45. Marketed as a quick, convenient and inexpensive food
Consumer handling
Marketed as a quick, convenient and inexpensive food
Appears cooked when it is not
Don’t realize that hands are contaminated after handling it
Cross contaminates surfaces
Not cooked as per package directions (time/temperature)
Microwaved when it is not intended for microwave

46. Manufacturing processes
Uncooked, frozen, processed chicken products can be made from:
Whole chicken pieces
Multiple chicken sources mixed together
If multiple sources of chicken mixed together:
Increased risk of contamination, potentially with multiple strains
Mixed throughout the product
Tests of random samples of uncooked, frozen, processed chicken nuggets from grocery stores in three Canadian provinces conducted by FoodNet Canada in 2014 and 2015 found that the proportion of chicken nugget samples positive for Salmonella and Salmonella Enteritidis was as high as or higher than the proportion of raw chicken breasts that were positive.

47. Directive from the Canadian Food Inspection Agency (CFIA)
On March 13, 2018, the Canadian Food Inspection Agency issued a directive regarding raw, frozen, breaded chicken products to the poultry industry:
requires that over a 12 month period industry implement changes in the manufacturing process to reduce Salmonella in raw, frozen, breaded chicken products to non-detectable levels.
Canadian Food Inspection Agency
working-with-the-poultry-industry-to-reduce-the-risk-of-salmonella-illness-from-frozen-raw- breaded-chicken-products.html Access April 29, 2018

48. Whole genome sequencing and outbreak investigation
Advantages
Detected clusters that previously were difficult to link
More definitively linking case and food sample matches
Can rule out non-clusters and cases that don’t belong to cluster
Can identify serotype and antibiotic resistance
Disadvantages
Slow to get results
Difficulty with displaying and communicating data
Could pose challenges when collaborating with other jurisdictions
New technology
Determining what is and isn’t connected
Implicating a source in these outbreaks has in a large part been facilitated by two huge improvements in our ability to investigate case clusters, and includes the use of;
PHAC’s Foodbook data to compare foods eaten by outbreak cases to the general population. This depends on public health units collecting comprehensive food histories during case initial interviews, through the implementation of standardized questionnaires.
And the second is WGS, which has allowed us to detect clusters that would have otherwise been difficult to link, as well as more definitively link cases to food samples, collected by public health units.
If public health unit investigator believe that uncooked, frozen, processed chicken is a likely source of illness for a case, collecting that food sample at the initial interview could lead to lead to implicating a source in a outbreak that is later declared.

49. Acknowledgments Leigh Hobbs and publication co-authors
Anne Maki, Sandra Zittermann, Allana Murphy, Anna Majury, and Dean Middleton
Dean Middleton and Anne Maki for WGS slides
Ontario public health units
Ministry of Health and Long-Term Care
Public Health Agency of Canada & FoodNet Canada
Canadian Food Inspection Agency
Health Canada
EZVBD team and other PHO/PHOL staff
Thank you to Ontario public health units for reporting case information in iPHIS, and collecting food samples.

50. References
Hobbs JL, Warshawsky B, Maki A, Zittermann S, Murphy A, Majury A, et al. Nuggets of wisdom: Salmonella Enteritidis outbreaks and the case for new rules on uncooked frozen processed chicken. J Food Prot. 2017;80(4): Available from:
Middleton D, Savage R, Tighe MK, Vrbova L, Walton R, Whitfield Y,et al. Risk factors for sporadic domestically acquired Salmonella serovar Enteritidis infections: a case-control study in Ontario, Canada, Epidemiol Infect. 2014;142(7):
Currie A, MacDougall L, Aramini J, Gaulin C, Ahmed R, Isaacs S. Frozen chicken nuggets and strips and eggs are leading risk factors for Salmonella Heidelberg infections in Canada. Epidemiol Infect. 2005;133(5):
MacDougall L, Fyfe M, McIntyre L, Paccagnella A, Cordner K, Kerr A, Aramini J. Frozen chicken nuggets and strips--a newly identified risk factor for Salmonella Heidelberg infection in British Columbia, Canada. J Food Prot. 2004;67(6):

51. References
Public Health Agency of Canada. Foodbook report. Ottawa, ON: Her Majesty the Queen in Right of Canada; Available from: canada/migration/healthy-canadians/publications/eating- nutrition/foodbook-2015/alt/pub-eng.pdf
Thomas MK, Majowicz SE, Pollari F, Sockett PN. Burden of acute gastrointestinal illness in Canada, : interim summary of NSAGI activities. Can Commun Dis Rep. 2008;34(5):8-15. Available from: publications/canada-communicable-disease-report-ccdr/monthly- issue/ /34-05/burden-gastrointestinal-illness-canada.html
Canadian Food Inspection Agency. 2016. The Canadian food safety system: food recalls. 5 January Available at:  system/food-recalls/eng/ /  

52. ?