The findings of the present study showed a prevalence of 49.1% for Salmonella spp. and 81.1% for oxyurids in tortoises farmed in southern Italy. Previous studies on the prevalence of Salmonella in tortoises conducted worldwide show heterogeneous results. Recovery rates of Salmonella in the present study are similar to the data reported by Savage and Baker  and Percipalle et al.  who reported a prevalence of 38.0% and 34.1%, respectively. Nevertheless, a recent survey conducted by Nowakiewicz et al.  showed levels of infection lower than those reported in the present study with a prevalence of 18.7%, whereas a study conducted by Pasmans et al.  showed levels of infection higher than those reported in our present study with a prevalence of 79.0%. The reasons for this recovery rate variation are not clear. The methods of isolation, species of tortoise, and geographical circumstances could have been involved in explaining these variable results.
The most prevalent Salmonella serotypes isolated in the present study were S. Richmond, S. Abony, and S. Hermannswerder which have been previously isolated in both captive and free-living tortoises [3, 7, 21]. It is noteworthy that some of the serotypes isolated during the present study have been associated with outbreaks of human salmonellosis worldwide. S. Kottbus have been associated with an outbreak in infants in Spain linked to contamination of commercial bottled water . Another study conducted by Ryder et al.  suggests that S. Kottbus may colonize the human mammary gland and thus be transmitted to infants by breastfeeding. S. Richmond has been isolated during an outbreak of salmonellosis in a military detachment in Spain  and during an outbreak of dysentery in children in Malaysia .
It is not possible to speculate regarding the source of Salmonella infections in the present study because they may have been initiated before or after introduction of tortoises in the farms. The higher prevalence of certain serotypes compared to others could be a result of the potential cross-infection between animals which occurred during housing in the farms.
With respect to the parasitological analysis our results showed a prevalence of 81.1% for oxyurids; higher values compared with those recently reported in a study conducted by Papini et al.  who recorded a prevalence of 29.2% for these pinworms in captive tortoises farmed in central Italy. We can explain this difference in the prevalence rates by the different diagnostic tools used in the two studies. Papini et al.  used a routine faecal flotation method to detect helminthes. In contrast, we used the highly sensitive and multivalent FLOTAC technique which was also showed as the best copromicroscopic method for assessing pinworm prevalence as reported by Rinaldi et al.  in rabbits.
Interestingly, a significant positive correlation between Salmonella spp. and oxyurid EPG was observed in our study. Tortoises with high oxyurid EPG values were also positive to Salmonella spp. It is well demonstrated that parasites may be affected, directly or indirectly, by cytokines and other immune effector molecules and parasites may themselves produce factors that affect the cells of the immune system . The consequent immunomodulation may enhance host susceptibility to other infectious pathogens  such as Salmonella spp. Furthermore, the interactions between parasites and other infectious agents have been described in various surveys in literature. Cringoli et al.  reported the co-presence of antibodies to Neospora caninum and Leishmania infantum in dogs suggesting that one protozoan may enhance the susceptibility to other one. Rinaldi et al.  showed N. caninum and Bovine herpesvirus 1 (BHV-1) coinfection in cattle, demonstrating the role of N. caninum as a primary pathogen and its presence as a risk factor for BHV-1 infection in this animal species. These studies, in line with the present study, suggest that in mixed infections the burden of one or both the infectious agents may be increased, one or both may be suppressed or one may be increased and the other suppressed. However, the correlation found between Salmonella and oxyurids in the present study does not necessarily reflect a causal relationship between the two organisms. Indeed confounding factors related to husbandry could have been involved in determining this correlation. It is very likely that this correlation might be also enhanced by the habitat (sanitation, temperature, diet) of the tortoises and further studies are needed to address this issue.