Prevalence of Clostridium difficile in retail pork

Abstract number: P2040

Metcalf D.S., Reid-Smith R.J., Avery B.P., Weese J.S.

Objective:Clostridium difficile has been isolated from varying percentages of healthy animals, and some strains found in food animals are those implicated in infections in humans. To date, only a few studies assessing C. difficile contamination of food products have been performed and few have used systematic sampling methods over broad geographic regions. The objective of this research was to determine the prevalence of C. difficile in Canadian retail pork products and to characterise these isolates.

Methods: Retail pork was collected from 4 Canadian provinces between November 2007 and May 2008. Five grams of each sample were incubated anaerobically in a C. difficile selective medium for 7 days and plated onto blood agar plates. The colony identities were confirmed using biochemical tests and isolates were characterised using standard typing techniques including ribotyping and toxinotyping, in addition to being screened for the presence of tcdA, tcdB, binary toxin gene (cdtB), and tcdC sequence analysis.

Results:C. difficile was isolated from 1.8% (7/393) retail pork samples; 4/296 (1.4%) ground pork samples, and 3/97 (3.1%) pork chops. Five different ribotypes and 3 different toxinotypes were identified. Three isolates were ribotype 027 and toxinotype III, with genes encoding toxins A, B and binary toxin, an 18 bp tcdC deletion and a truncating mutation in tcdC. One strain had a different ribotype from 027 but was toxinotype III, positive for all 3 toxin genes and had the same tcdC deletion and mutation as 027. One isolate was toxinotype XXVI and possessed genes for toxins A and B and an unaltered tcdC gene, while another was a toxinotype 0 strain possessing the same toxin gene characteristics. One isolate was non toxigenic. All toxigenic strains have been found in people in Canada. There were no statistically significant associations between prevalence of C. difficile and province (p = 0.28) or sample type (p = 0.37).

Conclusions: Although the implications for food safety practices remain elusive, the frequency of toxigenic isolates and isolates indistinguishable from known human pathogenic strains suggests contaminated pork may be a source of C. difficile in humans. While the prevalence of contamination was lower here than in some other reports, further investigation of sources of contamination and clinical relevance are needed.