Food supplements containing viable bacteria, so called probiotics, have been suggested to have beneficial health effects due to their influence on the normal microflora. However, there has been safety concern regarding probiotics containing Enterococcus faecium. Although part of the normal intestinal microflora in humans, enterococci can cause infections such as urinary tract infections, septicaemia, and endocarditis. Enterococci are also inclined to develop antibiotic resistance and their hardy nature promotes survival and dissemination in the hospital setting. Although the importance of E.faecium as a bloodstream isolate is increasing, little regarding its virulence is known. One virulence trait attributed to E. faecium is the enterococcal surface protein, Esp, encoded by the esp gene. Since enterococci have different roles from the human perspective, such as occurrence in food, as probiotic strains, members of the normal intestinal microflora, and as the cause of nosocomial infections, it is important to study differences and similarities between strains of different origin. The aims of the present investigation were to study some safety aspects of a probiotic product containing Efaecium, the colonization and transmission of enterococci from a nosocomial perspective, and differences between E.faecium isolates of different origin in some characteristics of probable importance for virulence. Faecal samples were collected from healthy volunteers administered an E. faecium probiotic. Half of the volunteers received simultaneous peroral vancomycin. The E. faecium isolates were subtyped and an in vitro conjugation assay was performed. The investigated strain could survive gastrointestinal transit in humans, and intake temporarily increased the number of enterococci in the faecal microflora. The vancomycin administration prevented detection of the probiotic strain. The probiotic strain could gain the vanA gene cluster under in vitro conditions. Samples from the respiratory tract and stomach were collected from 20 consecutive patients undergoing mechanical ventilation. The enterococci isolated were subsequently subtyped. Seventeen of the 20 subjects were colonized with enterococci in the respiratory tract. Genotype analyses suggested that 13 patients were involved in a transmission event, including all patients intubated more than 12 days. The E. faecium isolates were more resistant to antimicrobial agents compared to E. faecalis. The E. faecium isolates from different origins, i.e., infections, faeces, and probiotic products, were investigated for the presence of esp, their ability to conjugate, and adhere to epithelial cells in vitro. The esp gene was significantly more common among blood isolates. E. faecium strains enriched with esp adhered significantly better, were less genetically diverse, and had a higher conjugation frequency compared to esp-negative blood isolates. Antibiotic resistance was only detected among the infection-derived isolates. The adhesion among esp-negative isolates from the normal microflora was higher compared to the esp-negative bacteraemia isolates. In conclusion, the investigated probiotic E. faecium strain can survive gastrointestinal transit in humans, and intake may transiently increase the number of enterococci in the faecal microflora. The same strain can gain the vanA operon. Enterococci are frequently disseminated between mechanically ventilated patients, and isolates from different subpopulations seem to have different characteristics in terms of occurrence of esp, adhesion properties, antibiotic resistance, and conjugation frequencies.