Foodborne Disease Trends and Reports

Salmonella Outbreaks from Reptile Exposure

S almonella enterica is commonly isolated from reptiles and may be considered part of their normal intestinal flora. It is well known that humans may contract salmonellosis from direct or indirect contact with reptiles. Contact to these animals should always be suspected when disease is caused by S. enterica subspecies (subsp.) other than subsp. enterica (subsp. I). However, reptiles may also carry subsp. enterica strains, and infections caused by common serotypes may be associated with reptiles (Bertrand et al., 2008; Pedersen et al., 2009). Most reptile-associated infections are sporadic, but outbreaks have also been described. Two recent outbreaks caused by Salmonella serotype (ser.) Typhimurium from the United States illustrate the problem.

The first was a multistate outbreak occurring in 2008 that was associated with exposure to pet turtles (CDC, 2010a). The outbreak was detected by the Pennsylvania Department of Health and caused by a strain displaying two closely related pulsed-field gel electrophoresis patterns. A total of 135 case-patients in 25 states were subsequently identified during a 6-month period by comparison of the outbreak patterns to the national Salmonella database in PulseNet, the national molecular subtyping network for foodborne disease surveillance. The median age of the case-patients was 7 years. A case–control study showed a strong association with turtle contact; 16 out of 17 patients for which the information was available reported exposure to small turtles with a shell length <4 inches. It was not possible to trace back the source of the turtles partly because turtles are sold at flea markets and by street vendors who move frequently. This outbreak was the third salmonellosis outbreak related to small turtle exposure since 2006 (Harris et al., 2010); the first two were caused by ser. Pomona and ser. Paratyphi B var. L(+) tartrate + (formerly var. Java), both belonging to subsp. enterica. The median age in the former was 3 years and in the latter was 7 years.

The second recent ser. Typhimurium outbreak was associated with exposure to small aquatic frogs (CDC, 2010b). It occurred during most of 2009 and involved 85 case-patients in 31 states. The median age of the case-patients was 5 years. The outbreak was detected by the Utah Department of Health; again, the geographic extent of the outbreak was determined using molecular subtyping data from PulseNet. In this outbreak, pulsed-field gel electrophoresis data were supplemented with another subtyping method, multilocus variable number of tandem repeat analysis. The source of this outbreak, the aquatic African dwarf frog, was also identified in a case–control study and it was confirmed microbiologically by isolation of the outbreak strain from frogs and aquariums in patient homes, at one vendor and at the breeder, a single facility in California. Interestingly, 19 out of 36 patients knew about the risk of contracting salmonellosis from contact to reptiles, but only 11 were aware that frogs could cause the problem.

Antimicrobial Resistance Among Campylobacter Isolated from Food Animals

Macrolides such as erythromycin and azithromycin and fluoroquinolones such as ciprofloxacin are the treatments of choice for Campylobacter infections in humans. Foods of animal origin are a major source of human campylobacteriosis, and there is much debate about the extent to which antimicrobial use in food animals contributes to antimicrobial-resistant human infections. A recent cross-sectional study (Rollo et al., 2010) compared antimicrobial resistance among Campylobacter isolated from conventional swine farms (n = 60) with Campylobacter isolated from swine farms that had not used antimicrobial agents for 1 year or more (n = 35). Eight Midwestern states were included, and sample collections were performed on finisher pigs in 2002 and 2003 (all four seasons represented). Isolation from feces was done on blood agar (without Campylobacter-selective supplements) under microaerophilic conditions. Campylobacter were identified and speciated using biochemicals and polymerase chain reaction assays, and antimicrobial susceptibilities were determined by Etest (ABBiodisk, Solna, Sweden). A total of 1422 samples were processed, and Campylobacter isolation rates were similar for conventional and antimicrobial-free (AF) operations both at the farm level (95.0% and 94.3%, respectively) and at the individual animal level (35.8% and 36.4%, respectively). Not all of the 512 isolates were speciated, but of the 426 that were; most (99.6%) were Campylobacter coli. Erythromycin resistance prevalence was significantly higher for conventional farms than for AF farms both at the farm level (94.5% vs. 50.0%, respectively) and at the animal level (68.3% vs. 21.3%, respectively). A higher percentage of AF farms yielded ciprofloxacin-resistant Campylobacter compared with conventional farms (13.3% vs. 1.8%, respectively), but a slightly higher percentage of animals on conventional farms yielded ciprofloxacin-resistant Campylobacter than animals on AF farms (3.8% vs. 3.4%, respectively); neither of these results were statistically significant. Interestingly, fluoroquinolones were not approved for use in swine during the period the study was conducted. Season of Campylobacter isolation was not identified as a confounder, but herd size (using a 2000 animal cutoff ) was identified as a confounder and was included in the model used for odds ratio calculations. Erythromycin and azithromycin resistance prevalence at the animal level was significantly lower for AF farms compared with conventional farms even for farms that had only been AF for 1–2 years, and prevalence of macrolide resistance was lower the longer farms had been AF. However, tetracycline resistance prevalence was only significantly lower than conventional farm prevalence for farms that had maintained AF status for 4 years or more. Multidrug resistance (as represented by macrolides and tetracycline) was significantly higher on conventional farms, and pansusceptibility was significantly higher on AF farms. The cross-sectional study design did not allow pre-AF antimicrobial resistance prevalence to be captured for the AF farms; assuming that AF farms would have had conventional farm resistance prevalence before stopping antimicrobial use is a potential limitation of the study.

A recent survey conducted in Shandong Province, China, revealed high prevalence of clinically relevant resistance among Campylobacter isolated from chickens (Chen et al., 2010). Cecal samples were collected from chickens at five geographically representative slaughter facilities in June 2008, and Campylobacter were isolated on selective media (containing polymyxin B, vancomycin, and trimethoprim) under microaerophilic conditions. Genus and species were determined by multiplex polymerase chain reaction. Antimicrobial susceptibility testing was performed in triplicate by agar dilution. Of the 767 ceca sampled, 275 (35.9%) yielded Campylobacter. Most (75.6%) were species jejuni; 19.3% were coli; and species for 5.1% could not be determined. Of the 202 jejuni isolates available for susceptibility testing, 99.5% were resistant to ciprofloxacin, and 26.7% were resistant to erythromycin. All of the 52 coli isolates that were susceptibility tested were resistant to both ciprofloxacin and erythromycin. Resistance to other antimicrobial agents was also high for jejuni and coli,including gentamicin (27.2% and 92.3%, respectively) and tetracycline (100% for both species). One coli isolate was resistant to florfenicol, whereas 79.2% of the jejuni isolates were flofenicol resistant. It was not determined whether use of antimicrobial additives in the initial isolation selected for a more antimicrobial-resistant population of Campylobacter; nonetheless, the antimicrobial resistance prevalence estimates identified are alarmingly high.