C-EnterNet 2007 Annual Report

National Integrated Enteric Pathogen Surveillance Program

10. Exposure Sources

The Exposure Sources chapter, which is new for the 2007 Annual Report, summarizes the results presented in the disease-specific chapters by the main potential exposure sources monitored by C-EnterNet in the sentinel site: agriculture, surface water, and retail food. This section is a preliminary step forward towards source attribution. The data presented in this chapter include 2005, 2006 and 2007 surveillance years. Detailed descriptions of C-EnterNet's sampling and laboratory methods are available on our web-site (http://www.phac.gc.ca/c-enternet/index.html).

10.1 Agriculture

Swine Operations

Pathogenic strains of Campylobacter spp., Salmonella enterica, Yersinia spp., Listeria spp., Giardia spp. and Cryptosporidium spp. were found on swine farms in the sentinel site (Table 10.1). Campylobacter spp. was recovered from 36% (131/359) of the samples. C. coli was the most frequent species recovered, in 90% of swine samples (118/131), but was infrequently associated with human cases (2%) (Table 10.2). Of the 28 Campylobacter isolates from swine that were subtyped, only one had an MLST pattern (ST 459) that matched a human case (see Table 3.3 in Chapter 3.). Salmonella spp. were recovered from 31% (113) of the samples; Typhimurium (including var. 5-) was the most common serotype found in both swine and humans (Table 10.2). Twenty-seven of the Typhimurium isolates (including var. 5-) had PFGE patterns that have been identified in human cases (see Table 4.2 in Chapter 4). Eighty-one percent of the isolated Yersinia strains were considered pathogenic to humans (Table 10.2). Twenty-five percent of the Cryptosporidium strains and almost half of the Giardia strains were zoonotic.

Dairy Operations

Pathogenic strains of Campylobacter spp., Salmonella enterica, Giardia spp. and Cryptosporidium spp. and VTEC were found on dairy farms in the sentinel site (Table 10.1). Campylobacter spp. were recovered from 23% (67/291) of the samples. C. jejuni was the most frequently recovered species 47% (32/67) and most frequently associated with human campylobacteriosis (Table 10.2). Eleven of the 22 Campylobacter isolates detected in dairy cattle manure were matched by MLST pattern with isolates from human cases (see Table 3.3 in Chapter 3). Twenty-four percent of the dairy cattle manure samples were positive for Salmonella. Salmonella Kentucky was the most frequent serotype detected, but was infrequently associated with human illness (Table 10.2). E. coli O157:H7 was isolated from 8% (22/291) of the dairy manure samples. Four of these twenty-two isolates matched PFGE patterns identified in human cases (see Table 5.2 in Chapter 5).

Beef Operations

Pathogenic strains of Campylobacter spp., Salmonella enterica, Listeria spp., VTEC, and Cryptosporidium spp. were found on beef farms in the sentinel site (Table 10.1). Campylobacter spp. were recovered from 13% (10/80) of the samples. Similar to results from dairy farms, C. jejuni was the most frequent species recovered (Table 10.2). Four of the seven Campylobacter isolates that were subtyped had MLST patterns that matched human Campylobacter isolates (see Table 3.3 in Chapter 3). Pathogenic E. coli O157:H7 was detected in 9% of the beef manure samples, although their PFGE patterns did not match any human cases from the sentinel site (see Table 5.2 in Chapter 5). Listeria monocytogenes was most commonly isolated in the beef manure samples (64% compared to other manure types) (Table 10.1). The two most common Listeria monocytogenes serotypes isolated in beef manure samples (1/2a and 4b) are considered important serotypes among human cases (see Table 7.2 in Chapter 7). Although pathogenic strains of Cryptosporidium were identified in 23% of samples, only non-pathogenic Giardia strains were identified (Table 10.1).

Broiler Chicken Operations

Pathogenic strains of Salmonella enterica, Listeria spp.,VTEC and Giardia spp. were found on poultry farms in the sentinel site (Table 10.1). Two common Salmonella serotypes detected in broiler chicken manure, Enteritidis and Heidelberg, were also commonly associated with human illness (Table 10.2). The PFGE patterns of these broiler chicken manure isolates matched human PFGE patterns (see Table 4.2 in Chapter 4). S. Heidelberg was only detected on chicken farms whereas the serotype Enteritidis was also detected on swine farms. Two other common Salmonella serotypes in broiler chicken manure Hadar and Kentucky, were only occasionally associated with human illness (Table 10.2).

These results illustrate that the local food animal farms are reservoirs of important pathogens known to cause human enteric illnesses. Contact with food animals or their environment may lead to unintentional human infections and occasionally to disease.

TABLE 10.1: Pathogen prevalence in livestock farms in Sentinel Site 1 between 2005 and 2007

Pathogen	Swine n= 359	Broiler Chickens n= 36	Beef n= 80	Dairy n= 291
Campylobacter spp	131 (36%)	0	10 (13%)	67 (23%)
Cryptosporidium spp (PCR test)	31/122 (25%) c	0 c	17/72 (23%) c	20/187 (11%) c
Giardia spp (PCR test)	56/122 (46%) c	1 (3%) c	0 c	21/187 (11%) c
Listeria monocytogenes	4/122 (3%)	1 (3%)	51 (64%)	15/179 (8%)
Salmonella enterica	113 (31%)	26 (72%)	8 (10%)	34 (12%)
Verotoxigenic E. coli (VTEC)	12 (3%) a	1 (3%) a	7 (9%) b	12 (4%) a, 22 (8%) b
Yersinia spp	21 (6%)	NT	NT	NT

TABLE 10.2: Pathogen species and serotypes in positive samples from livestock farms and in human endemic cases in Sentinel Site 1 between 2005 and 2007

Pathogen species or serotypes	Swine	Broiler Chickens	Beef	Dairy	Human endemic cases
Campylobacter spp	n= 131	n= 0	n= 10	n= 67	n= 325
C. jejuni			6 (60%)	32 (47%)	315 (97%)
C. coli	118 (90%)		4 (40%)	14 (21%)	5 (2%)
C. upsaliensis					1 (0.3%)
C. lari					1 (0.3%)
Other	13 (10%)			21 (31%)	2 (1%)
Salmonella enterica	n= 113	n= 26	n= 8	n= 34	n= 199
Typhimurium (including var 5-)	47 (42%)		2 (25%)	5 (15%)	62 (31%)
Enteritidis	1 (1%)	4 (15%)			34 (17%)
Heidelberg		5 (19%)			19 (10%)
Thompson		3 (12%)		1 (3%)	7 (4%)
Infantis	6 (5%)		1 (12%)		8 (4%)
Adelaide					2 (1%)
Agona	12 (11%)			3 (9%)	5 (3%)
Hadar		6 (23%)			3 (2%)
Hartford					2 (1%)
Javiana					2 (1%)
Kentucky		5 (19%)	2 (25%)	18 (53%)	2 (1%)
Newport					5 (3%)
Derby	14 (12%)				1 (1%)
Oranienberg					4 (2%)
Other serotypes	33 (29%)	3 (12%)	3 (38%)	7 (21%)	43 (22%)
Yersinia	n= 21	NT	NT	NT	n= 41
Y. enterocolitica -pathogenic	17 (81%) a				39 (95%)
Y. enterocolitica -non-pathogenic	4 (19%)				1 (2%)
Y. intermedia -non-pathogenic					1 (2%)

10.2 Surface Water

Campylobacter spp., Cryptosporidium spp., Giardia spp., S. enterica, VTEC and Yersinia spp. were detected in untreated surface water (Table 10.3). The recovery rate was always lower for the culture-based methods compared to the molecular-based methods, which was expected (see Appendix C in C-EnterNet 2006 Annual Report). The species distribution for Campylobacter and Yersinia illustrates that a majority of isolates from the surface water samples were either non-pathogenic to humans or not amongst the most frequent ones detected in the human clinical samples (Table 10.4). However, there was some overlap between isolates recovered in the watershed and those recovered from the other potential reservoirs and from the human cases. For example, of the 17 Campylobacter jejuni isolates recovered from the Grand River watershed in 2006, only one matched the MLST pattern of an isolate detected by other surveillance efforts (found on local dairy farms, on retail chicken and pork meat, and in human clinical samples) (see Table 3.3 in Chapter 3). Among the Salmonella isolates recovered from the Grand River watershed in 2007, the most common serotype was Thompson, which has been detected on local chicken farms, retail chicken meat, and among the human clinical samples (see Table 4.1 in Chapter 4). Based on additional molecular typing, three Typhimurium isolates matched PFGE patterns of isolates recovered from human samples, while one matched a PFGE pattern recovered from dairy, beef and swine farms, as well as human clinical samples (see Table 4.2 in Chapter 4). Pathogenic strains of Yersinia were never detected in the river water, although in 2007 there were two clinical samples in the community that were positive for Y. intermedia, and Y. enterocolitica 1A, traditionally considered to be non-pathogenic. One of the most common Cryptosporidium genotypes detected in the river water was C. andersoni. This genotype was also the most frequently detected genotype on local beef farms (see Table 8.2 in Chapter 8).

TABLE 10.3: Contamination of raw surface water in Sentinel Site 1 in 2005, 2006 and 2007

Pathogen	Culture based method	Molecular based method
Entamoeba spp	NT	NT
Campylobacter spp	13%	69%
Cryptosporidium spp a	90%	NT
Cyclospora spp	NT	NT
Giardia spp	97%	NT
Hepatitis A virus	NT	NT
Listeria monocytogenes	NT	NT
Salmonella enterica	14%	24%
Shigella spp	NT	NT
Verotoxigenic E. coli (VTEC)	2% b,c	26% d
Yersinia spp	29% b	45% b

The monthly average parasite counts detected in the surface water samples exhibit important seasonal variations (see Chapter 8, Sections 8.1 and 8.2), implying that the water contamination is dynamic. These dynamics and the transmission of enteric pathogens from both point and non-point sources within the watershed could be further explored.

The surface water in Sentinel Site 1 is not pristine and contamination may come from three main sources: the human population, the local farming activities (i.e. cattle grazing and manure spreading on land) and wildlife, from both upstream and local sources.

TABLE 10.4: Pathogen species and serotypes in positive samples from untreated surface water and in human endemic cases in Sentinel Site 1 in 2005, 2006 and 2007

Pathogen species or serotypes	Water	Human endemic cases
Campylobacter	n=40	n=325
C. jejuni	33%	96.9%
C. coli	9%	1.5%
C. upsaliensis	0%	0.3%
C. lari	61%	0.3%
Not typed		0.9%
Cryptosporidium	n=64	n=36
C. andersonia	66%	NT
C. baileyi	5%	NT
C. cervinea	8%	NT
C. parvum (bovine genotype) a	5%	NT
C. hominis a,b	9%	NT
C. muskrat genotype I	5%	NT
C. muskrat genotype II	3%	NT
Other	14%	NT
Salmonella enterica	n=48	n=199
Typhimurium	10%	28.7%
Enteritidis		15.7%
Heidelberg	4%	8.8%
Thompson	17%	3.2%
Infantis	2%	3.7%
Adelaide		0.9%
Agona	2%	2.3%
Hartford		0.9%
Javiana		0.9%
Kentucky	13%	0.9%
Newport	6%	2.3%
Oranienberg		1.9%
Other serotypes	46%	21.8%
Not typed		7.9%
Yersinia c	n=97	n=33
Y. aldovae - non-pathogenic	6%
Y. bercovieri - non-pathogenic	9%
Y. enterocolitica - pathogenic	0%	95%
Y. enterocolitica - non-pathogenic	13%	2%
Y. frederiksenii - non-pathogenic	23%
Y. intermedia - non-pathogenic	29%	2%
Y. kristensenli - non-pathogenic	2%
Y. mollaretti - non-pathogenic	6%
Y. rohdei - non-pathogenic	1%

10.3 Retail Food

Retail Pork

Pathogenic strains of Campylobacter spp., Salmonella enterica, VTEC, and Listeria spp., were found on raw retail pork chops, although at relatively low levels (Table 10.5). Of the Salmonella serotypes detected, 25% were Typhimurium (including var. 5-), the most common Salmonella serotype associated with human illness (Table 10.6). Thirty-four percent (14/41) of the Listeria monocytogenes isolates subtyped matched PFGE patterns that are among the top ten patterns associated with human illness in Canada (see Table 7.3 in Chapter 7). Although Yersinia spp. were detected, further subtyping determined that the strains were not pathogenic to humans (Table 10.6).

Retail Chicken

Pathogenic strains of Campylobacter spp., Salmonella enterica, and Listeria spp., were consistently detected on raw retail chicken samples (Table 10.5). Of the retail meats tested, the highest prevalence of Campylobacter (33%), was observed on chicken and C. jejuni was the most commonly identified species (Table 10.6). Forty-three of the 82 Campylobacter isolates subtyped by MLST had patterns that matched human Campylobacter cases (see Table 3.3 in Chapter 3). Although the most frequent Salmonella serotype (Kentucky) was not frequently associated with human illness, the second most common serotype (Heidelberg) was identified in 10% of human endemic salmonellosis cases in the sentinel site (Table 10.6). Nineteen of the 27 S. Heidelberg isolates matched PFGE patterns with human cases (See Table 4.2 in Section 4). One third of the retail chicken samples were positive for Listeria monocytogenes. Thirty-two percent (41/128) of the subtyped isolates had PFGE patterns that are among the top ten patterns associated with human illness in Canada (see Table 7.3 in Chapter 7).

Retail Beef

Pathogenic strains of Campylobacter spp., Salmonella enterica, and VTEC were found in retail ground beef although at relatively low levels (Table 10.5). Listeria monocytogenes was found at moderate levels and 31% (30/96) of the isolates that were subtyped had PFGE patterns that are among the top ten patterns associated with human illness in Canada (see Table 7.3 in Chapter 7).

Retail pork, chicken and beef meats are potential sources of human enteric pathogens. However, more data are required to characterize and formally assess the potential risk of the foodborne transmission of enteric pathogens and to quantify the significance of each type of meat for its public health impact.

No attempts were made to compare the data from the agriculture component to the retail meat data, since a small, although unknown, proportion of meat eaten in Sentinel Site 1 originates from animals raised in the area.

TABLE 10.5: Contamination of raw retail meat in Sentinel Site 1 between 2005 and 2007

Pathogen	Pork n= 388	Chicken n= 392	Beef n= 387
Campylobacter spp	3 (1%)	129 (33%)	1 (<1%)
Listeria monocytogenes	40 (10%)	126 (32%)	90 (23%)
Salmonella enterica	12 (3%)	119 (30%)	2 (<1%)
Verotoxigenic E. coli (VTEC)	1 (<1%)	0	2 (<1%)
Yersinia spp	33 (9%)	NT	NT

TABLE 10.6: Pathogen species and serotypes detected in raw retail meat and in human endemic cases in Sentinel Site 1 in 2007

Pathogen species or serotypes	Pork	Chicken	Beef	Human endemic cases
Campylobacter	n=3	n=131	n=1	n=322
C. jejuni	2 (67%)	115 (88%)	1 (100%)	315 (97%)
C. coli	1 (33%)	16 (12%)		5 (2%)
C. upsaliensis				1 (0.3%)
C. lari				1 (0.3%)
Salmonella enterica	n=12	n=121	n=2	n=199
Typhimurium	3 (25%)	7 (6%)		62 (31%)
Enteritidis		8 (7%)	1 (50%)	34 (17%)
Heidelberg		27 (22%)		19 (10%)
Thompson	2 (17%)	1 (<1%)		7 (4%)
Infantis	1 (8%)	2 (2%)		8 (4%)
Adelaide				2 (1%)
Agona				5 (3%)
Hartford				2 (1%)
Javiana				2 (1%)
Kentucky	1 (8%)	54 (45%)		2 (1%)
Newport				5 (3%)
Oranienberg				4 (2%)
Hadar		8 (7%)		3 (2%)
Orion			1 (50%)
Other serotypes	5 (42%)	14 (12%)		44 (22%)
Yersinia	n=31	NT	NT	n=41
Y. enterocolitica -pathogenic	0			39 (95%)
Y. enterocolitica -non-pathogenic	21 (68%)			1 (2%)
Y. frederiksenii -non-pathogenic	7 (23%)
Y. intermedia -non-pathogenic	3 (10%)			1 (2%)