To the Editor: Enterococcus faecalis, which exists commensally in the gut in warm-blooded animals and humans, is an opportunistic pathogen that causes a variety of community-acquired and health care–associated infections, such as urinary tract and intraabdominal infections, bacteremia, and endocarditis (1). Only a few studies have assessed the relationships between clinical E. faecalis strains; strains endemic to the health care setting; and community strains residing in humans, animals, or animal-origin food (2).

Recently we showed that the emergence of high-level gentamicin-resistant (HLGR) E. faecalis among patients with infective endocarditis (IE) coincided with an increase in HLGR E. faecalis in the pig population in Denmark (3). The majority of isolates belonged to the same clonal group (sequence type [ST] 16), suggesting that pigs constitute a community reservoir of HLGR E. faecalis. We investigated human and porcine community reservoirs of other E. faecalis clonal types associated with IE in humans in Denmark.

A total of 20 consecutive gentamicin-susceptible E. faecalis isolates were obtained from IE patients in North Denmark Region during 1996–2002 (Table A1). Cases of IE were classified as definite (n = 12) or possible (n = 8) according to the modified Duke criteria (4). A case of community-acquired E. faecalis infection (n = 6) was defined in accordance with strict criteria applied for methicillin-resistant Staphylococcusaureus (5); otherwise, cases were deemed to be health care associated (n = 14) (Table A1). HLGR ST16 isolates recovered from 2 IE patients during the study period have been characterized (3) and were excluded from the present study.

Using multilocus sequence typing (6), we identified 14 STs among the 20 IE isolates (Table A1), then compared them with STs from 2 collections of E. faecalis isolates collected as part of the Danish Integrated Antimicrobial Resistance Monitoring and Research Program (www.danmap.org): 1) all 14 isolates recovered from community-dwelling humans in North Denmark Region during 2002–2006 with approval from the local ethics committee ([KF] 01-006/02), which were classified into 10 STs in this study (Table A1); and 2) 19 pig isolates from 2001 that were shown in a previous study to belong to 12 STs (7).

Among the 14 STs identified in IE isolates, 4 (ST19, ST21, ST72, and ST306) and 2 (ST40 and ST97) were also found among isolates from community-dwelling humans and pigs, respectively (Table A1). Isolates belonging to these 6 STs were further characterized by pulsed-field gel electrophoresis (PFGE) by using SmaI and grouped into PFGE pulsotypes as described (3). STs and PFGE pulsotypes (A–F) were largely concordant (ST97:A, ST72:B, ST19:C, ST40:D, ST21:E, and ST306:F), except for 2 isolates belonging to ST72 and ST40, for which PFGE banding patterns (U1 and U2, respectively) were unrelated to the major PFGE pulsotypes (A–F), and 1 ST306 isolate exhibiting the ST21-like PFGE banding pattern E (Table A1).

These findings confirm the genetic relatedness of IE isolates with those from community-dwelling humans (ST72:B, ST19:C, ST21:E, and ST306:F) and pigs (ST97:A and ST40:D). Seven (64%) of 11 IE isolates belonging to these 6 clonal types originated from IE patients with health care–associated risk factors (Table A1), which suggests that health care users are predisposed to colonization and infection with E. faecalis strains residing in human and porcine community reservoirs.

Previous reports have shown that epidemiologically distinct E. faecalis populations differ in terms of biofilm formation, virulence gene content, and antimicrobial drug susceptibility profiles (2,8). Therefore, we characterized all isolates with respect to these traits. Isolates were categorized into strong, medium, weak, and nonbiofilm formers by using the method of Mohamed et al. (8). The presence of 12 virulence-associated and pathogenicity island genes (ebpA, gelE, ef1824, hylA, ef1896, ef2347, ef2505, hylB, ace, cbh, esp, and ef0571) was investigated by using colony lysates and probes that have been described elsewhere (9). The antimicrobial drug susceptibility profiles (ampicillin, chloramphenicol, ciprofloxacin, erythromycin, gentamicin, kanamycin, linezolid, penicillin, streptomycin, teicoplanin, tetracycline, and vancomycin) were determined by the Sensititre system (Trek Diagnostic Systems, East Grinstead, UK) in accordance with Clinical and Laboratory Standards Institute guidelines (10). The isolates were generally homogenous within each clonal type in terms of biofilm formation, presence of virulence-associated and pathogenicity island genes, and resistance profiles (Table A1), further supporting that IE isolates are genetically related to those from community-dwelling humans and pigs, respectively. Notably, most IE isolates were susceptible to ampicillin (100%), penicillin (100%), vancomycin (100%), high-level gentamicin (100%), and high-level streptomycin (80%), which are the drugs of choice in therapeutic regiments for E. faecalis endocarditis.

In conclusion, our results suggest that the normal intestinal microflora of humans and pigs are community reservoirs of clinical E. faecalis and link 2 porcine-origin clonal types of gentamicin-susceptible E. faecalis, ST97:A, and ST40:D to IE in humans in Denmark. This finding strengthens existing evidence that pigs can be a source of serious infections in humans.