Population structure and antibiotic resistance of the genus enterococcus in humans, animals and the environment

Abstract

Enterococci belong to the normal intestinal flora of humans and animals. An increased prevalence of antibiotic resistant enterococci causing nosocomial infections has drawn attention to the epidemiology and emergence of antibiotic resistance in this genus.

In the present thesis we have studied the enterococcal flora in samples from humans, animals, and the environment, in order to be able to follow the movement of bacteria between different ecological niches, as well as to determine the prevalence of antibiotic resistant enterococci in these samples. In a European study, 17,157 enterococcal isolates from 2,868 samples from humans, animals and the environment in Sweden, Denmark, the United Kingdom, and Spain were studied (study I-III). The diversities of enterococci in environmental samples were generally high. Samples from hospital sewage, urban sewage, and manure contained enterococcal populations that reflected those in faecal samples of hospitalised patients, healthy humans and animals. Thus, such samples could be used as pooled faecal samples and replace cumbersome samplings from many individuals.

Vancomycin resistant enterococci (VRE, resistant to 20 mg/L vancomycin), were identified in 8.2% of all samples and most frequently and at similar levels in untreated urban sewage in Sweden, Spain and the UK (in an average of 71% of the samples). In contrast, pig faeces and manure were more often VRE-positive in Spain than in Sweden (30% vs. 1%), most probably reflecting the former use of the vancomycin analogue avoparcin as a feed additive. Most VRE were E. faecium carrying vanA both among humans and animals. Typing of VRE showed a high degree of polyclonality and no evidence were found for transmission of VRE strains between humans and animals.

The high prevalence of VRE in Swedish sewage samples (19-60% in 118 samples) was unexpected. Typing of 35 isolates revealed a high diversity (Di 0.97). Four of five VRE from hospital sewage were E. faecium with vanB, which is the most common type in infections and among hospitalised patients in Sweden. However, the origin of VRE from urban sewage remains unclear. A majority of VRE from urban sewage were E. faecium with vanA (17 of 29), but a larger proportion than found in the other countries was E. faecalis with vanA (11of 29). Either these VRE represent a higher carriage rate among healthy individuals in the community than earlier reported or perhaps they harbour in the sewage system.

An ampicillin and ciprofloxacin resistant E. faecium (ARE) strain, named FMSE1, was in a previous study found to dominate among faecal ARE isolates from patients in several Swedish hospitals. In study IV, the prevalence of the same PhP-type as the FMSE1 PhP-type, was searched for among typing data from 9676 isolates from Sweden and Denmark. FMSE1 was most common in samples of hospital sewage (50%), surface water (35%), treated sewage (28%), and untreated sewage (17%), but were rare in samples from healthy children (0.8%) and animals (2%). PFGE typing of FMSE1-like isolates from hospital sewage indicated that they were closely related to the nosocomial FMSE1 strain. According to study I-III the enterococcal flora in sewage and hospital sewage resembled that of the flora in individual faecal samples. This fact led to an idea for a new concept for monitoring antibiotic resistance in the community and in hospitals, based on samplings of sewage water.

In study V and VI the feasibility of this concept was evaluated. Up to 24 enterococcal isolates from each sample of hospital sewage (N=9), sewage treatment plants (N=14), and sewage from an anthroposophic village, were screened for resistance, using breakpoint concentrations of antibiotics in microplates. The resistance rates found for ampicillin, ciprofloxacin and erythromycin were markedly higher in hospital sewage (30, 35 and 30%) than in community sewage (4, 6 and 15%l), whereas tetracycline resistance was found at the same level in all sewage types (28%). Differences in resistance rates for enterococci isolated from different types of sewage samples were obvious and easy to monitor using this method.