Salmonella serotypes in reptiles and humans, French Guiana

carriage was significantly higher for wild reptiles. Near two-thirds of the Salmonella


Introduction
Salmonella is a major foodborne pathogen found worldwide.Non-typhoidal Salmonella causes 93.8 million human cases of gastroenteritis and 155,000 deaths annually worldwide (Majowicz, et al., 2010).The Salmonella genus includes two species: S. enterica and S. bongori.S. enterica is subdivided into six subspecies: enterica (former subgenera I), salamae (II), arizonae (IIIa), diarizonae (IIIb), houtenae (IV) and indica (VI).Over 2,600 serotypes were identified according to the White-Kauffmann-Le Minor scheme (Grimont and Weill, 2007).The digestive tract of warm-blooded animals is the main reservoir of Salmonella subspecies enterica (Grimont and Weill, 2007), while other subspecies are mainly found in cold-blooded animals.Reptiles can harbour multiple subspecies and serotypes simultaneously (Pedersen et al., 2009).Most human salmonellosis is associated with eating contaminated raw or undercooked food of animal origin.Other transmission routes exist: human-to-human contacts, consuming contaminated vegetables, water (Wegener et al., 2003) and direct contact with infected animals, including pets.In western countries serotypes from reptiles are rarely encountered in humans (Woodward et al., 1997;O'Byrne and Mahon, 2008).Within the European Union, it is estimated that <1% of human salmonellosis are associated with reptiles (Bertrand et al., 2008).Furthermore, serotypes from reptiles seem to be responsible for more severe complications and hospitalizations (Cieslak et al., 1994).
In French Guiana, a French overseas territory (around 85,000 km 2 ) located in South America, nearly 50% of Salmonella isolates from human belong to serotypes rarely encountered in Metropolitan France (Le Hello and Weill, unpublished data).
The objectives of this study conducted around Cayenne, the main city of French Guiana, were to assess Salmonella carriage by different reptile species living in different ecosystems (zoo, forest, commensal environment or caged/contained in or around human dwellings) and to compare reptilian and human serotype distributions.

Surveillance of human Salmonella infections.
In France, human Salmonella surveillance is a voluntary laboratory-based network headed by the French National Reference Center for Salmonella (FNRC-Salm).Participating laboratories annually send 8,000-10,000 Salmonella isolates to the FNRC-Salm, which serotypes them and runs weekly outbreak-detection algorithms (David et al., 2011).In 2011, all the medical laboratories from French Guiana (8 private biomedical laboratories and 2 public hospitals) participated to the present study.

Frequency of Salmonella carriage in reptiles.
From 5 April to 6 June 2011, reptiles were captured from more-or-less human-occupied environments around Cayenne.Four living conditions were defined: zoo, for captive animals in the zoo; forest, for reptiles captured there; commensal environment, for reptiles living in or around human lodgings (e.g.Hemidactylus mabouia); reptiles caged (e.g., snake terrarium) or contained in a space from which there is no egress (e.g., turtles within a garden) in or around a private home.
Cloacal swab samples were collected and kept at room temperature (26 ± 4°C) for a maximum of 48 hr.Samples were incubated in 9 ml of buffered peptone water for 16-20 hr at 37°C.Three drops of the pre-enrichment broth were inoculated onto specific medium (Hektoen agar, BioMérieux, Marcy L'Etoile, France) and incubated at 37°C for 21-27 hr.
Reptile species were grouped as: caimans, snakes, green iguanas (Iguana iguana), other lizards and turtles.Statistical analyses were computed with Stata® v9.0 Stata Corp; chi 2 or Fisher's exact tests were used.
The frequency of Salmonella carriage in cloacal samples was 23.2% (35/151,, ranging between 14.3% and 32.4%, depending on the type of reptile (Table 1), but did not differ significantly (p = 0.5).Habitat was significantly associated with Salmonella carriage (p < 0.001), with frequencies of 5.8%-39.6%.The frequency of carriage was higher for reptiles captured in forests and caged/contained in or around private homes.

Discussion
The global frequency of Salmonella carriage by reptiles in the Cayenne area was 23.2%.Salmonella carriage frequency did not differ significantly among groups (caimans, green iguanas, other lizards, snakes and turtles), the highest was documented in snake (32.4%) and the lowest in Caiman (14.3%).However, each group is composed of various species; the frequency of carriage might differ depending on the species (Geue and Löschner, 2002).
Carriage frequency also differed according to the reptile's habitat.Reptiles captured in the forest had the highest rate (39.6%) and zoo dwellers the lowest (5.8%).Those observations seem to contradict literature data, as some authors reported higher frequencies (47%-88.9%)for zoo-housed reptiles (turtle, snake and lizard) versus 14%-58.8%for their wild counterparts (Geue and Löschner, 2002;Scheelings et al., 2011).That low frequency of carriage among sampled zoo reptiles is difficult to explain as no antibiotics had been administered during the months preceding the study.Indeed, captive conditions should favor inter-individual and interspecies transmission; their captivity and exposure to the general public (e.g.noise) could be sources of stress, inducing the onset of symptoms and Salmonella excretion (Friendship et al., 2009;Maciel et al., 2010).Salmonella frequency of carriage by reptiles caged/contained in or around private homes was high at 38.1% which tends to confirm higher frequency of contamination in wild reptiles.All of the latter had been caught in their natural environment and were captive for < 6 months.A limitation in our study is that each reptile type was not sampled in all habitats, e.g., green iguanas were sampled only at the zoo.However, considering all reptiles sampled, the forest dwellers were, on the whole, fairly wellrepresented (31.8%), in contrast to most of studies that focused on captives (Sá and Solari, 2001;Pedersen et al., 2009).
Comparing our data to those in the literature was difficult as no standard methodology exists for Salmonella isolation from reptiles but globally we could have underestimated frequency of carriage.Cloacal swabbing without enrichment step and use of simplified culture methods mainly target enterica subspecies, the major source of human and animal clinical infections, might underestimate carriage frequencies.We might also have underestimated arizonae and diarizonae subspecies, because respectively, 25% and 75% of these subspecies ferment lactose (Grimont and Weill, 2007).
Moreover, due to intermittent Salmonella shedding in feces (Bauwens et al., 2006), cloacal swabbing is less sensitive than using digestive tracts from sacrificed animals.However, regulations protecting reptiles restrict it.In addition, only one colony was tested for each animal, consequently coinfection was unlikely.Only one reptile specimen showed multiple infections while reptiles can harbour several Salmonella serotypes simultaneously (Chiodini and Sundberg, 1981).
We observed high serotype diversity among reptile samples obtained around Cayenne, with identification of 14 different serotypes among the 23 isolates sent to the FNRC-Salm.At least three different serotypes were isolated from each reptile type (except for caimans which harboured only one Salmonella serotyped).Lizards exhibited a wide Salmonella diversity, with seven different serotypes found.No patterns of host-specific serotype association were observed, as Briones et al. (2004) claimed.In addition, almost half (47.2%) of Salmonella isolates belonged to the subspecies enterica, thereby confirming that reptiles serve as reservoirs of this subspecies normally associated with the digestive tract of warm-blooded animals (Geue and Löschner, 2002).
In our study, 64.3% of Salmonella serotypes isolated from reptiles were also isolated from patients in French Guiana.In contrast only 13% of serotypes isolated were also found in inspected farm chickens (Food, Agriculture and Forestry of French Guiana, unpublished data), it might indicate low environmental contaminations by reptiles.However, no veterinary controls are performed on poultry units producing less than 250 chickens and products are sold directly to consumers.The distribution of Salmonella serotypes could differ in such small poultry farming, as facilities (livestock buildings, granary) are less efficient and interactions with reptiles possibly higher.
Thus, reptiles might be the direct or indirect source of a high percentage of human Salmonella infections in French Guiana.

Conclusion
Reptiles are Salmonella reservoirs and might represent public health risk, especially in French Guiana, where a huge diversity of reptile species is observed and some in close contact with humans (Hemidactylus mabouia having ready access into homes…) and probably livestock.Furthermore, they represent food sources, particularly green iguanas.Therefore, hygiene measures concerning the handling of reptiles, cooking their meat and cleaning kitchen utensils should be promoted.

Table 1 Reptile
Salmonella carriage according to reptile type or living conditions

TABLE 2
Distribution of S. enterica subspecies among sampled reptiles

TABLE 3
Salmonella serotypes isolated from humans in French Guiana in 2011 *One per patient.