Toxoplasma gondii in Retail Beef,Lamb,and Pork in Canada: Prevalence,Quantification,and Risk Factors from a Public Health Perspective

Abstract

Toxoplasmosis is a zoonosis of public health concern worldwide. Toxoplasma gondii can be transmitted to humans through the consumption of raw or undercooked meat, but little is known about the risk associated with various meat commodities consumed by Canadians. We set out to estimate the prevalence and risk factors of T. gondii-positive status of retail meat in Canada using a combination of ELISA and quantitative polymerase chain reaction, and found a prevalence of 0.3% in pork (n = 360), 0.6% (n = 329) in beef, and between 4.3% and 11.1% in lamb (n = 298), depending on province. Most of (75%) positive lamb samples (100 g each) contained more than 4000 parasites, relative to the standard curve. Lamb had a higher risk (p < 0.001) of being T. gondii positive compared to beef and pork. Using multivariable logistic regression, T. gondii positivity in lamb samples was associated (p < 0.001) with “country of origin” and “meat cut,” but no association was detected regarding store location, store category (independent, chain or butcher), or store volume. Ground lamb was identified as a risk factor compared to any other cuts. Overall, 26.4% of New Zealand imported lamb was positive to T. gondii compared to 4.0% of Canadian meat and 4.1% of Australian meat; only the difference between New Zealand and Australia was statistically significant in the multivariable model (p < 0.001). However, considering the conditions under which it is transported, frozen New Zealand (and Australian) lamb should not present a risk to Canadian public health. These results provide critical information for performing T. gondii quantitative risk assessments in food and for orienting acquired and congenital toxoplasmosis prevention strategies by targeting potential transmission sources.

Introduction

Toxoplasma gondii, a protozoan found throughout the world, is an important public health issue. While the acquired infection often remains asymptomatic, it can cause abortions in pregnant women (Remington, 1968), fatal encephalitis in immunocompromised people (Luft, 1989), and ocular lesions that can lead to blindness (Wilder, 1952; Ho-Yen et al., 1992). Congenital toxoplasmosis can lead to significant fetal defects and/or neonatal issues (Remington, 1968).

T. gondii can infect all warm-blooded animals as well as humans. The main routes of horizontal transmission to humans are the ingestion of oocysts from the environment and the ingestion of bradyzoites encysted in host tissue (meat). The World Health Organization estimated that, globally, 49% of sporadic acquired toxoplasmosis cases are foodborne (Torgerson et al., 2015). Raw or undercooked meat seems to be responsible for a considerable proportion of foodborne cases (Cook et al., 2000; Guo et al., 2015).

The parasite can be inactivated by most domestic cooking methods, with immediate destruction occurring when the meat reaches an internal temperature of 67°C (Dubey et al., 1990; Jones and Dubey, 2012). Freezing for 2 d at −18°C to −20°C has also been reported to be effective in inactivating T. gondii cysts in meat (Dubey, 1988; El Nawawi et al., 2008; Cressey et al., 2015). This is of importance in the Canadian context, where 58% of lamb meat consumed is imported, of which 96% is in a frozen state (Ministry of Agriculture, Fisheries and Food of Quebec, 2014). Various treatments of the meat such as curing, salting, and cold smoking can inactivate the parasite with varying efficiency (Robert-Gangneux and Dardé, 2012).

The gold standard for T. gondii detection in tissues is the bioassay in cat, which provides information about the viability of the parasite, but involves the use of laboratory animals. T. gondii bradyzoites in meat can also be detected by conventional DNA extraction and polymerase chain reaction (PCR); however, this technique can lead to unreliable detection and quantification, mostly because cysts are unequally spread within the tissues of the host. To overcome this issue, magnetic capture DNA extraction followed by a quantitative PCR (qPCR) has been elaborated (Opsteegh et al., 2010) and shown to be highly sensitive, allowing the detection of a single cyst in up to 100 g of meat. Although it does not provide information on T. gondii viability, it can provide accurate and reliable quantification outputs (Opsteegh et al., 2010), making it a viable tool for specific detection in several commodities. Alternatively, seropositivity (the presence of T. gondii-specific IgG antibodies) may predict the presence of T. gondii DNA in tissues of pigs, sheep, poultry, and goat, but not in cattle and horses (Opsteegh et al., 2016).

The risk of human exposure to T. gondii by meat consumption depends also on cultural factors linked to meat consumption and preparation, on host-related factors (e.g., age and immunological status), and on parasite-related factors (e.g., viability and genotype) (Dubey, 2010; Robert-Gangneux and Dardé, 2012). This risk has been evaluated in a few modeling studies (Mie et al., 2008; Opsteegh et al., 2011a; Guo et al., 2015, 2016, 2017). Most of these studies have identified knowledge gaps that lead to assumptions and simplifications, largely regarding the presence and quantity of cysts in edible tissues (Crotta et al., 2016). In Canada, the scarcity of information on T. gondii presence in meat precludes the development of risk assessment models and ultimately, adequate risk analysis and management. The main objectives of this study were to (1) estimate the prevalence and identify risk factors for the presence of T. gondii in retail pork, lamb, and beef in Canada and (2) provide a quantitative estimate of parasites per meat portion.

Materials and Methods

Experimental design and sampling

A cross-sectional study was conducted in five provinces of Canada: Alberta, British Columbia, Quebec, Ontario, and Saskatchewan (shown in Fig. 1). Pork, beef, and lamb at the retail level were collected on a weekly basis by the Canadian Integrated Program for Antimicrobial Resistance Surveillance (Government of Canada, 2014). A random selection of census divisions within each province was performed, stratified by population size. In each selected area, a given number of sampling days was allocated and each area was weighted by population size. For each sampling day, three chain stores and one independent store or butcher were visited and one fresh raw meat sample per commodity was sampled in each store. Ground beef samples were collected and various cuts of pork and lamb were selected depending on availability, package volume, and price of purchase. The required sample size for prevalence estimation was calculated for each commodity using a confidence level of 95%, a precision of 2% (pork) or 5% (beef, lamb), and an expected prevalence of 3% (pork), 25% (beef), or 40% (lamb) according to previous publications (Lafrance-Girard, 2018).

FIG. 1.

Global Canada map showing the provinces where pork, lamb, and beef retail meat samples were collected in 2015–2017 to assess their “T. gondii status.” Lamb samples were collected in Quebec, Ontario, and British Columbia only. Color images available online at www.liebertpub.com/fpd

For each sample, information was collected on the date of sampling, location of the store, store category, store volume, country of origin, meat cut, final processing in store, and the labels “previously frozen” and “organic.” All pork (Fig. 2) and lamb (Fig. 3) samples were sent to the laboratory for ELISA testing. All seropositive samples in addition to a random selection of weak seronegative samples were then submitted to molecular detection. All beef samples were directly submitted to molecular analysis (Appendix Fig. A1), given the low reliability of serology for this animal species (Opsteegh et al., 2011b).

FIG. 2.

Description of the methodological process to determine Toxoplasma gondii status of retail pork samples. All samples were tested with ELISA on meat juice. When the optical density of the sample compared to the positive control (included in the kit) had a positivity percentage (Pos. %) equal or higher to 20%, the sample was considered seropositive. Otherwise, it was considered seronegative. Seropositive samples were to be confirmed by qPCR, but in the case of pork, none was found; so this step was not needed (dashed lines represent planned, but unneeded steps). Among seronegative samples, the median positivity percentage (−0.09%) was used to delimitate strong seronegatives (lower than median) from weak seronegatives (higher than median). Among weak seronegatives, 87 samples were selected by random stratified sampling and tested by qPCR to verify their status. If more than 250 parasites (detection limit) were detected in a sample, it was considered to have a positive status. qPCR, quantitative polymerase chain reaction. Color images available online at www.liebertpub.com/fpd

FIG. 3.

Description of the methodological process to determine Toxoplasma gondii status of retail lamb samples in this study. All samples were tested with ELISA on meat juice. When the optical density of the sample compared to the positive control (included in the kit) had a positivity percentage (Pos. %) equal or higher to 20%, the sample was considered seropositive. Otherwise, it was considered seronegative. Seropositive samples (n = 29) were tested by qPCR for confirmation. Among seronegative samples, the median positivity percentage (3.1%) was used to discern strong seronegatives (lower than median) from weak seronegatives (higher than median). Among weak seronegatives, 30 samples were selected by random stratified sampling and tested by qPCR to verify their status. If more than 250 parasites (detection limit) were detected in a sample, it was considered to have a positive status. qPCR, quantitative polymerase chain reaction. Color images available online at www.liebertpub.com/fpd

Sample processing

Samples were received at the Public Health Agency of Canada laboratory of Saint-Hyacinthe within 24 h postpurchase. They were kept at 4°C for 2 to 3 d before being brought to the Research Chair in Meat Safety laboratory in Université de Montréal. Samples were frozen (−20°C for at least 3 d) and thawed to increase meat fluid seepage, as recommended by the ELISA test manufacturer. Once ELISA was performed, tissues were refrozen and stored at −20°C for T. gondii DNA extraction in 100 g of each selected sample.

Serological analysis

T. gondii-specific IgG antibodies were detected in meat fluids with host-specific commercial indirect ELISA kits (prioCHECK^® Toxoplasma AB). The protocols provided by the manufacturer were followed, with the exception that the plate reader filter was set at 640 nm instead of 620 nm because of equipment limitations. According to Buholzer et al. (2010), this ELISA test applied on the meat fluids of pork had 97% sensitivity and 100% specificity relative to indirect fluorescent antibody test (IFAT), western blot (WB), and/or ELISA. With lamb, the sensitivity and specificity were 91.3–97.9% and 94.9–97.9%, respectively, on meat juice relative to IFAT and indirect hemagglutination assay (IHA) (Glor et al., 2013).

Controls for molecular detection

Samples of 100 g of rear leg muscle from seronegative pigs were used as molecular detection controls. These tissues came from experimental animals raised in a biocontainment level 2 facility at the Université de Montréal. For positive controls and standard curve DNA extracts, a known number of parasites (a minimum of 2000 in positive controls) were spiked into weighted and prepared meat from which DNA was extracted by magnetic capture. Negative controls consisted of blank control samples from which DNA was extracted. Type 1 tachyzoites from the RH strain were used to spike control samples. These parasites were first stored dry at −80°C, and afterward suspended in sterile water and stored at −20°C until use.

DNA extraction and qPCR

A specific and highly conserved fragment of T. gondii genome—the repetitive “529-bp” element—was extracted from meat by magnetic capture and quantified by qPCR (magnetic capture polymerase chain reaction [MC-PCR]) (Opsteegh et al., 2010) with minor modifications. Some reagents (proteinase K), PCR components (probes master mix), and equipment (LightCycler 96 thermo-cycler; Roche, Laval, Canada) were different from the original protocol, and the competitive internal amplification control was not used in our study (Technical Appendix A1). The detection limit of our MC-PCR method was 250 parasites per portion of 100 g of meat, which corresponds to the lowest concentration that was systematically detected.

A positive and negative meat control were included in every magnetic capture run and tested by qPCR with the samples processed in that run. The quantification was relative to the Cp-values of the PCR standard curve, which was made of extracted DNA (by MC-PCR) from pork spiked with a known number of parasites. Five extractions, from 250 to 4000 parasites with a dilution factor of 2, were included in the standard curve. Positive samples were visualized on agarose gel electrophoresis to confirm amplicon size (162 bp). The amplicons of one positive sample of each commodity and a positive control were submitted to a Sanger-type sequencing to confirm their identity.

Statistical analysis

Prevalence of antibody-positive and T. gondii-positive samples with 95% exact confidence limits was estimated for each commodity. T. gondii-positive status was defined as the molecularly confirmed presence of more than 250 parasites (detection limit), notwithstanding serological status. T. gondii-negative status was defined as the molecularly confirmed absence (less than 250 parasites) of T. gondii or as seronegative samples that were not selected for PCR confirmation.

Multivariable logistic regression was used to model T. gondii status according to the commodity, adjusted for the location of the store. For each commodity, associations between T. gondii status and risk factors of interest (Table 3) were then evaluated using a chi-square test; only variables with p < 0.25 were selected for multivariable modeling. Pairwise correlations between selected variables were assessed using chi-square tests. In the presence of strong pairwise correlation, the variable deemed to be the most directly related to the presence of T. gondii DNA in meat from a biological point of view was included. Multivariable logistic regression models were built using backward selection, with p > 0.05 (log-likelihood ratio test) as criterion for rejection. The Hosmer–Lemeshow test was performed to assess the fit of our data in the final model. The results of the final model were expressed as adjusted odds ratio (adj. OR) with 95% confidence intervals (CIs). All statistical analyses were performed in SAS^® 9.4.

Results

Sampling

Between September 2015 and January 2017, a total of 363 pork and 329 beef samples were collected from the five provinces. Lamb samples (n = 302) were collected from December 2015 to November 2016 in Ontario, British Columbia, and Quebec. Due to time and budget constraints, a convenience sampling occurred in December 2016 and January 2017, representing 52 lamb samples that were collected exclusively in Quebec using the same store selection protocol, but including as many different cuts as possible from each store.

Seroprevalence and the presence of T. gondii DNA

Prevalence results are detailed in Table 1. For prevalence estimation, data from all sampled provinces were aggregated for pork and beef, but separated for lamb due to differences in the sampling effort per province. The highest seroprevalence and T. gondii-positive status prevalence were observed in lamb collected in Quebec. There were no seropositive pork samples.

Table 1.

Seroprevalence and Prevalence of Toxoplasma gondii in 363 Pork, 329 Beef, and 302 Lamb Retail Meat in Canada in 2015–2017

				T. gondii antibodies			T. gondii-positive status ^b
				n positive			n positive
Commodity	n samples	n tested ^a ELISA	n tested qPCR		%	Seroprevalence		%	Apparent prevalence
Pork	363	360	87	0	0	—	1	0.3	0.0–1.5
Beef	329	329	329	ND	ND	ND	2	0.6	0.0–2.2
Lamb	302	298	59
Quebec	173	171	39	22	12.9	8.2–18.8	19	11.1	6.8–16.8
British Columbia	72	70	14	4	5.7	1.6–14.0	3	4.3	0.9–12.0
Ontario	57	57	6	3	5.3	1.1–14.6	3	5.3	1.1–14.6

Some samples could not be tested because of an absence of meat fluids.

Positive status was defined as a seropositive sample confirmed positive with more than 250 parasites (detection limit) by qPCR or a seronegative sample randomly selected for molecular confirmation with presence of more than 250 parasites (detection limit) by qPCR.

ND, not done; qPCR, quantitative polymerase chain reaction.

For lamb, no PCR-positive samples were found in a subsample of 30 “weak seronegative” samples (Fig. 3). In seropositive lamb samples, 86% (25/29) were confirmed positive by qPCR. For pork, 87 samples were selected from the “weak seronegative” group (Fig. 2). Only one pork sample containing T. gondii DNA was detected in this subsample.

Relative quantification

Out of 28 lamb samples with a Cp-value in qPCR, 21 (75%) had, relative to the standard curve, more than 4000 parasites (positive nonquantifiable), three were under the limit of detection (considered negative), and four were more precisely quantifiable (within the standard curve) (Table 2). All pork (one) and beef (two) samples with a Cp-value were considered positive and contained between 250 and 1250 parasites.

Table 2.

Distribution of Meat Samples (Frequency) According to the Quantity of Toxoplasma gondii Detected by MC-PCR in 87 Pork, 59 Lamb and 329 Beef Retail Samples in Canada in 2015–2017

		Number of Toxoplasma gondii DNA equivalents ^a
Commodity	n samples with Cp values	<250 (considered qPCR negative)	250–4000 (qPCR positive, quantifiable)	>4000 (qPCR positive, not quantifiable)
Pork	1	0	1^b	0
Beef	2	0	2^c	0
Lamb	28	3	4^d	21

The quantification is relative to the Cp-values of the standard curve, which is made of extracted DNA (by MC-PCR) from meat spiked with a known number of parasites (250–4000, with a dilution factor of 2).

Number of parasites ± PCR interduplicate standard deviation = 313 ± 69.

Number of parasites ± PCR interduplicate standard deviation = 604 ± 275 and 723 ± 519.

Number of parasites ± PCR interduplicate standard deviation = 1603 ± 475, 2941 ± 1360, 1933 ± 243, and 1941 ± 221.

MC-PCR, magnetic capture polymerase chain reaction; qPCR, quantitative polymerase chain reaction.

Risk factors

Commodity (pork, beef, or lamb) was associated with T. gondii status of retail meat (p < 0.001, χ² test). Lamb had a higher risk of being T. gondii positive compared to beef (odds ratio 13.9, 95% CI 3.3–59.5) and pork (odds ratio 33.3, 95% CI 4.5–247.1). No difference was observed between pork and beef (p = 0.48).

Due to the very small number of positive beef and pork samples, risk factor analysis per commodity was limited to lamb. Among the variables first selected (p < 0.25) for multivariable modeling (Table 3), the “final processing in store” and “previously frozen” variables were excluded due to their strong correlation (p < 0.001) with “country of origin.” According to the final model (Table 4), “Ground” meat compared to any other specific cut (category “other” is not considered specific) and lamb originating from New Zealand rather than Australia were identified as risk factors for a T. gondii-positive status in lamb.

Table 3.

Distribution of 360 Pork, 329 Beef, and 298 Lamb Retail Samples by Commodity, Meat Characteristics, and Toxoplasma gondii Status in Canada in 2015–2017

	Pork		Beef		Lamb
Variable	n^a	% positive	n^a	% positive	n^a	% positive	χ² p
Location of the store							0.16
Alberta	49	0	63	0	—	—
British Columbia	57	0	67	0	70	4.3
Ontario	13	0	14	0	57	5.3
Quebec	206	0	155	1.3	171	11.1
Saskatchewan	35	2.8	30	0	—	—
Store category							0.3
Butcher	39	0	44	2.3	26	15.4
Chain	268	0.4	221	0	226	7.1
Independent	53	0	64	1.6	45	11.1
Meat cut							<0.001
Ground	15	6.7	328	0.6	68	23.5
Leg	—	—	—	—	47	4.3
Loin	338	0	—	—	84	3.6
Shoulder	1	0	—	—	80	3.8
Other^b	6	0	1	0	16	6.3
Country of origin							<0.001
Australia	0	—	1	0	74	4.1
Canada	137	0.7	103	0	125	4.0
New Zealand	0	—	0	—	53	26.4
Other^c	5	0	0	—	3	0
Final processing in store							0.07
Yes	282	0	247	0.8	161	5.0
No	53	1.9	71	0	135	13.0
Previously frozen							<0.001
Yes	6	16.7	9	0	56	21.4
No	8	0	161	0.6	242	5.4
Organic							NA^d
Yes	0	—	1		1
No	285	0.4	263	0.4	105	3.8
Store volume^e							0.5
Small	142	0.7	152	0.7	71	11.3
Medium	156	0	122	0	168	7.7
Large	60	0	52	0	51	12.5

Discarded samples (absence of meat fluids) are excluded from this table. Samples with “Unknown” values for a variable were excluded and not represented for this specific variable.

The category “other” includes all other meat cuts found in stores such as stew cubes, neck, and so on.

United Kingdom, United States, and Spain were grouped as a “Other” category, because the number of samples in each category was very low (1 pork sample from Spain, 4 pork and 1 lamb sample from the United States, and 1 lamb sample from the United Kingdom).

The variable “Organic” was not tested, because the number of “Yes” value was too low.

“Small” store volume is considered 5 cash registers or less, “medium” goes from 6 to 10, and “large” is considered more than 10.

NA, not available.

Table 4.

Results from the Final Multivariable Logistic Regression Modeling Toxoplasma gondii-Positive Status in 298 Lamb Retail Samples in Canada in 2015–2017

	Adjusted odds ratio
Variable	Estimate	95% CI	p
Meat cut^a
Ground vs. leg	13.0	1.2–139.8	0.03
Ground vs. loin	8.6	1.9–37.8	0.005
Ground vs. shoulder	8.3	1.9–35.7	0.005
Ground vs. other	3.3	0.3–41.4	0.35
Country of origin^b
New Zealand vs. Australia	12.3	3.1–49.1	<0.001
New Zealand vs. Canada	3.2	0.9–11.7	0.08
Canada vs. Australia	3.8	0.7–21.6	0.13

Hosmer and Lemeshow goodness-of-fit χ² = 2.18, df = 7, p = 0.95

No other post hoc pairwise comparisons were significant (all p ≥ 0.35).

The category “Other” was excluded of statistical analysis due to small number of samples.

CI, confidence interval.

Sequencing

We found a similarity at the nucleotide level between our amplicons and the targeted fragment of the “529-bp” reference sequence on about 80 consecutive nucleotides out of 162 nucleotides, which we considered sufficient to confirm the identity of the detected fragment as T. gondii (data not shown).

Discussion

This project provides the first estimate of the proportion of retail meat containing T. gondii cysts in Canada. Only a few studies in North America are available, with which to compare our results.

In Canada, a seroprevalence of 0.3% among 1600 finishing pigs raised in Ontario was estimated by ELISA (Poljak et al., 2008), which is similar to the 0.3% (95% CI 0.0–1.5) we found. Using ELISA, a seroprevalence of 0.5% in retail pork and 0% in beef meat fluids in the United States was reported (Dubey et al., 2005). A seroprevalence of 17% in lamb heart extracts from the United States was reported by Dubey et al. (2008), which may be explained by differences in methodologies and matrix (heart being a predilection site), as well as the use of different serological assays and/or by discrepancies in husbandry practices (Wallander et al., 2015; Opsteegh et al., 2016).

Viable T. gondii was detected by biological assay in 0.38% and 0% of retail pork and beef from the United States, respectively (Dubey et al., 2005). The direct detection method used in our study did not assess the viability of the parasites in meat, which was the main limitation of our study. It is interesting to observe that Dubey's (2005) prevalence of viable T. gondii in meat in the United States is comparable to our “T. gondii-positive status” results. A prevalence of 13.8% of viable T. gondii was detected by mice bioassay in lamb hearts from slaughterhouses in the United States (Dubey et al., 2008). This slightly higher prevalence compared to the results reported in this study (11.1% at most), notwithstanding the different methods used, might also be explained by their use of hearts (as a matrix).

The detection limit of our adapted MC-PCR method was very similar to the one obtained in the original experiments (∼230 parasite equivalents), which seems comparable with the bioassay (Opsteegh et al., 2010). The absence of T. gondii was not confirmed by the molecular detection method in all seronegative samples, so DNA-positive samples might have been missed among the seronegative samples. Nevertheless, considering our verification in a subsample and supporting scientific opinions in the literature concerning the use of such a screening in lamb (Dubey et al., 2008; Opsteegh et al., 2010) and pork (Gamble et al., 2005; Opsteegh et al., 2016), we are confident that our approach provides reliable prevalence estimates. One single sample was revealed to contain T. gondii DNA out of the 117 randomly selected weak seronegative pork or lamb samples tested by MC-PCR. This ELISA false-negative could be explained by the imperfect sensitivity (i.e., 97%) of the serological assay and/or by a recent infection where antibodies (IgG) were not yet detectable.

As far as we know, this project is the first to identify risk factors for the presence of the parasite among intrinsic characteristics of retail meat. The main risk factor was commodity, with higher odds of positivity for lamb products compared to pork and beef. Consumption of rare lamb has been identified as a major risk factor for human toxoplasmosis in the United States (Jones et al., 2009), which supports our finding. Our results are consistent with the suspected resistance of cattle to T. gondii (Dubey et al., 2010). Moreover, most pigs produced in Canada are kept in confined housing, therefore reducing the risk from T. gondii infection, while lambs are reared within differing environments and with different husbandry practices, including confined housing, pasture-based farming, and a combination of both (Ontario Sheep Marketing Agency, 2014). Organic and free-range animal productions, which have been associated with higher T. gondii prevalence (Kijlstra et al., 2004; Jones and Dubey, 2012), represent a marginal part of the Canadian industry (Jacques et al., 2012).

The significant effect of New Zealand origin on the T. gondii-positive status of lamb meat compared to Australian origin was unexpected. Indeed, lambs in both countries are raised on pasture with a likely significant exposure to T. gondii oocysts due to the high density of feral cat populations (Langham and Charleston, 1990; Reynolds and Frame, 2005; Dempster et al., 2011; Government of Australia, 2011). We hypothesize that the higher level of rain reported in New Zealand compared to Australia could be involved. In fact, T. gondii oocyst infectivity declines at a faster rate in dry compared to humid environments (Lélu et al., 2012). However, the public health risk posed by New Zealand (and Australian) lamb to Canadian consumers is probably low since most (∼75% in 2012) of this imported meat is kept frozen at an expected temperature of −40°C during the 2 to 3 weeks of ship transportation (Kennedy, 2013; Ministry of Agriculture, Fisheries and Food of Quebec, 2014), which likely subverts the viability of T. gondii.

Ground lamb meat was associated with higher odds of positivity compared to other cuts. This association is biologically plausible as ground meat is often made of pooled skeletal muscles from several individuals.

Conclusions

Retail lamb was more likely to be T. gondii positive than pork or beef in Canada, with the majority (75%) of positive samples containing more than 4000 parasites (in 100 g of tissues). Lamb ground meat compared to other cuts and lamb coming from New Zealand in comparison to Australia had increased risks of having a positive T. gondii status. However, lamb imported from New Zealand (and Australia) is generally frozen and thus might not represent a public health concern. The presented work provides essential information for acquired and congenital toxoplasmosis risk assessment from retail meat at a national and international scale.

Footnotes

Acknowledgments

We thank Danielle Daignault, Julie Roy, and Sindy Cleary from the Canadian Integrated Program for Antimicrobial Resistance Surveillance (CIPARS) for coordinating the sampling. We thank Manon Racicot from the Canadian Food Inspection Agency (CFIA) for her contribution to the project. We thank the Research Chair in Meat Safety team for their support. We thank Louis-Philippe Leroux from INRS-Institut Armand-Frappier in Laval (Québec) for providing our control parasites. Finally, we thank Thermo Fisher Scientific (Prionics) for their contribution to laboratory supplies.

Disclosure Statement

No competing financial interests exist.

Technical Appendix A1. Detailed Method for Molecular Detection of Toxoplasma gondii in Retail Meat

References

Buholzer

, Pürro

, Schacher

, Haupt-Gerber

, Hehl

, Schares

, Deplazes

, Raeber

Evaluation of the PrioCHECK^® Toxoplasma AB Porcine ELISA for surveillance of Toxoplasma infections in pigs. In: Proceedings of the AAVLD Annual Conference, Minneapolis, MN, 2010.

Cook

, Gilbert

, Buffolano

, Zufferey

, Petersen

, Jenum

, Foulon

, Semprini

, Dunn

, Holliman

. Sources of Toxoplasma infection in pregnant women: European multicentre case-control study, Commentary: Congenital toxoplasmosis—further thought for food. Br Med J, 2000; 321:142–147.

Cressey

, Lake

Risk Profile Update: Toxoplasma gondii in Red Meat and Meat Products . Christchurch, New-Zealand: Institute of Environmental Science & Research Limited. Prepared for the Ministry for Primary Industries. Report No. FW15009, 2015, pp. 109.

Crotta

, Limon

, Blake

, Guitian

. Knowledge gaps in host-parasite interaction preclude accurate assessment of meat-borne exposure to Toxoplasma gondii . Int J Food Microbiol, 2016; 261:95–101.

Dempster

, Wilkins

, Green

, de Lisle

. Serological survey of Toxoplasma gondii and Campylobacter fetus fetus in sheep from New Zealand. N Z Vet J, 2011; 59:155–159.

Dubey

JP.

Long-term persistence of Toxoplasma gondii in tissues of pigs inoculated with T. gondii oocysts and effect of freezing on viability of tissue cysts in pork. Am J Vet Res, 1988; 49:910–913.

Dubey

JP.

Toxoplasmosis of Animals and Humans, Second Edition. Boca Raton, FL: CRC Press. 2010.

Dubey

, Hill

, Jones

, Hightower

, Kirkland

, Roberts

, Marcet

, Lehmann

, Vianna

MCB

, Miska

, Sreekumar

, Kwok

OCH

, Shen

, Gamble

. Prevalence of viable Toxoplasma gondii in beef, chicken, and pork from retail meat stores in the United States: Risk assessment to consumers. J Parasitol, 2005; 91:1082–1093.

Dubey

, Kotula

, Sharar

, Andrews

, Lindsay

. Effect of high temperature on infectivity of Toxoplasma gondii tissue cysts in pork. J Parasitol, 1990; 76:201–204.

10.

Dubey

, Sundar

, Hill

, Velmurugan

, Bandini

, Kwok

, Majumdar

, Su

. High prevalence and abundant atypical genotypes of Toxoplasma gondii isolated from lambs destined for human consumption in the USA. Int J Parasitol, 2008; 38:999–1006.

11.

El Nawawi

, Tawfik

, Shaapan

. Methods for inactivation of Toxoplasma gondii cysts in meat and tissues of experimentally infected sheep. Foodborne Pathog Dis, 2008; 5:687–690.

12.

Gamble

, Dubey

, Lambillotte

. Comparison of a commercial ELISA with the modified agglutination test for detection of Toxoplasma infection in the domestic pig. Vet Parasitol, 2005; 128:177–181.

13.

Glor

, Edelhofer

, Grimm

, Deplazes

, Basso

. Evaluation of a commercial ELISA kit for detection of antibodies against Toxoplasma gondii in serum, plasma and meat juice from experimentally and naturally infected sheep. Parasit Vectors, 2013; 6:85.

14.

Government of Australia. The feral cats (felis catus). Commonwealth of Australia. 2011. Available at: https://www.environment.gov.au/system/files/resources/34ae02f7-9571-4223-beb0–13547688b07b/files/cat.pdf Accessed July 1, 2017 .

15.

Government of Canada. Canadian Integrated Program for Antimicrobial Resistance Surveillance (CIPARS) Annual Report. 2014. Available at: www.phac-aspc.gc.ca/cipars-picra/pubs-eng.php Accessed September 1, 2016 .

16.

Guo

, Buchanan

, Dubey

, Hill

, Lambertini

, Ying

, Gamble

, Jones

, Pradhan

. Qualitative assessment for Toxoplasma gondii exposure risk associated with meat products in the United States. J Food Prot, 2015; 78:2207–2219.

17.

Guo

, Lambertini

, Buchanan

, Dubey

, Hill

, Gamble

, Jones

, Pradhan

. Quantifying the risk of human Toxoplasma gondii infection due to consumption of fresh pork in the United States. Food Control, 2017; 73:1210–1222.

18.

Guo

, Mishra

, Buchanan

, Dubey

, Hill

, Gamble

, Pradhan

. Quantifying the risk of human Toxoplasma gondii infection due to consumption of domestically produced lamb in the United States. J Food Prot, 2016; 79:1181–1187.

19.

Ho-Yen

DO.

Clinical features. In: Ho-Yen

, Joss

AWL

, eds. Human Toxoplasmosis. Oxford: Oxford University Press, 1992:56–78.

20.

Jacques

, Louvel

, Lessard

, Sanscartier

. Market analysis of organic products based on the development of the organic sector in Quebec [in French]. La filière biologique du Québec. 2012. Available at: https://www.agrireseau.net/agriculturebiologique/documents/%C3%89coRessources%20Fili%C3%A8re%20bio%202012%20Analyse%20des%20march%C3%A9s_2012-03-06_VF.pdf

21.

Jones

, Dargelas

, Roberts

, Press

, Remington

, Montoya

. Risk factors for Toxoplasma gondii infection in the United States. Clin Infect Dis, 2009; 49:878–884.

22.

Jones

, Dubey

Foodborne toxoplasmosis. Clin Infect Dis, 2012.

23.

Kennedy

2012. Lamb Imports to Canada. 2013. Available at: http://quartet.aps.uoguelph.ca/ONSheepProduction/?p = 973 Accessed May 1, 2018 .

24.

Kijlstra

, Eissen

, Cornelissen

, Munniksma

, Eijck

, Kortbeek

. Toxoplasma gondii infection in animal-friendly pig production systems. Invest Ophthalmol Vis Sci, 2004; 45:3165–3169.

25.

Lafrance-Girard

Toxoplasma gondii in retail meat in Canada: Prevalence, quantification and risk factors from a public health perspective [in French] (Master), Université de Montréal, Montreal, 2018. Available at: http://hdl.handle.net/1866/19890 Accessed May 1, 2018 .

26.

Langham

NPE

, Charleston

WAG

. An investigation of the potential for spread of Sarcocystis spp. and other parasites by feral cats. N Z J Agric Res, 1990; 33:429–435.

27.

Lélu

, Villena

, Dardé

M-L

, Aubert

, Geers

, Dupuis

, Marnef

, Poulle

, Gotteland

, Dumètre

Quantitative estimation of the viability of Toxoplasma gondii oocysts in soil. J Appl Environ Microbiol, 2012; 78:5127–5132.

28.

Luft

BJ.

Toxoplasma gondii. In: Walzer

, Genta

, eds. Parasitic Infections in the Compromised Host. New York: Marcel Dekker, 1989:179–279.

29.

Mie

, Pointon

, Hamilton

, Kiermeier

. A qualitative assessment of Toxoplasma gondii risk in ready-to-eat smallgoods processing. J Food Prot, 2008; 71:1442–1452.

30.

Ministry of Agriculture, Fisheries and Food of Quebec. Monography of ovine industry in Quebec [in French]. Government of Quebec. 2014. Available at: https://www.mapaq.gouv.qc.ca/fr/Publications/MonographieovineWEB.pdf Accessed October 1, 2017 .

31.

Ontario Sheep Marketing Agency. Introduction to Sheep Production in Ontario, Chapter 1: Introduction. 2014. Available at: www.ontariosheep.org/uploads/userfiles/files/a1%20introduction2014%20(Autosaved).pdf Accessed February 1, 2017 .

32.

Opsteegh

, Langelaar

, Sprong

, Den Hartog

, De Craeye

, Bokken

, Ajzenberg

, Kijlstra

, Van Der Giessen

. Direct detection and genotyping of Toxoplasma gondii in meat samples using magnetic capture and PCR. Int J Food Microbiol, 2010; 139:193–201.

33.

Opsteegh

, Prickaerts

, Frankena

, Evers

. A quantitative microbial risk assessment for meatborne Toxoplasma gondii infection in The Netherlands. Int J Food Microbiol, 2011a;150:103–114.

34.

Opsteegh

, Schares

, Blaga

, Van der Giessen

Experimental studies on Toxoplasma gondii in the main livestock species (GP/EFSA/BIOHAZ/2013/01) Final report. Parma, Italy: EFSA Supporting Publications, 2016: p. 13.

35.

Opsteegh

, Teunis

, Zuchner

, Koets

, Langelaar

, Van Der Giessen

Low predictive value of seroprevalence of Toxoplasma gondii in cattle for detection of parasite DNA. Int J Parasitol, 2011b;41:343–354.

36.

Poljak

, Dewey

, Friendship

, Martin

, Christensen

, Ojkic

, Wu

, Chow

. Pig and herd level prevalence of Toxoplasma gondii in Ontario finisher pigs in 2001, 2003, and 2004. Can J Vet Res, 2008; 72:303–310.

37.

Remington

JS.

Toxoplasmosis and congenital infection. Birth Defects, 1968; 4:49–56.

38.

Reynolds

, Frame

. Grasslands: Developments, Opportunities, Perspectives. Boca Raton, FL: CRC Press, Food & Agriculture Organization of the United Nations, 2005.

39.

Robert-Gangneux

, Dardé

. Epidemiology of and diagnostic strategies for toxoplasmosis. Clin Microbiol Rev, 2012; 25:264–296.

40.

Torgerson

, Devleesschauwer

, Praet

, Speybroeck

, Willingham

, Kasuga

, Rokni

, Zhou

, Fèvre

, Sripa

, Gargouri

, Fürst

, Budke

, Carabin

, Kirk

, Angulo

, Havelaar

, de Silva

. World Health Organization estimates of the global and regional disease burden of 11 foodborne parasitic diseases, 2010: A data synthesis. PLoS Med, 2015; 12:e1001920.

41.

Wallander

, Frossling

, Vagsholm

, Burrells

, Lunden

. (2015). “Meat juice” is not a homogeneous serological matrix. Foodborne Pathog Dis, 2015; 12:280–288.

42.

Wilder

HC.

Toxoplasma chorioretinitis in adults. AMA Arch Ophthalmol, 1952; 48:12736.