Distribution of fsr among Enterococcus faecalis isolates from the SENTRY antimicrobial surveillance program.

Abstract

Pillai and colleagues (3) in 2002 reported a striking correlation (100.0% positive; P = 0.005) of the fsr virulence gene with Enterococcus faecalis isolated from a collection of endocarditis cases. The comparison group of stool culture isolates had a significantly lower rate of positive fsr tests (53%) by methods described earlier by Qin et al. (4, 5). The fsr locus has been noted to be widely distributed in E. faecalis (70% of isolates) and regulates virulence-associated gelatinases and serine proteases (4, 5). To expand the understanding of fsr related to contemporary prevalence, we screened a collection of E. faecalis isolates from the SENTRY Antimicrobial Surveillance Program selected to achieve a broad range of infection types, geographic samplings, years of isolation, and antibiogram features (vancomycin only). A total of 109 strains were chosen from the 1999 to 2002 SENTRY Program organism bank for the United States, Canada, Latin America, and Europe. Strains were single, unique infection episodes without duplication of strains (pulsed-field gel electrophoresis screened) indexed by patient or medical center. A total of 49 medical centers contributed samples of E. faecalis distributed geographically as follows: United States (59 strains; 24 centers), Canada (9 strains; 7 centers), Latin America (19 strains; 7 centers) and Europe (22 strains; 13 sites). These strains occurred among documented infections within the monitored years at the following body sites in patients residing in intensive care units (11 strains) or in general hospital wards (9 strains): bloodstream (51 strains sampled in all regions), lower respiratory tract (7 strains; United States only), skin and soft tissue (27 strains; all regions), and urinary tract (11 sites in all regions). The fsr locus was amplified by the methods described earlier (3, 5) using the following primer sequences: 5′-AACCAGAATCGACCAATGAAT-3′ (upstream primer) and 5′-GCCCCTCATAACTCAATACC-3′ (downstream primer). The PCR testing conditions utilized were those published by Pillai et al. (3), and the fsr-positive E. faecalis ATCC 51299 was used as a control. Table ​Table11 shows the results of the PCR fsr screen for all 109 E. faecalis strains. The gene was quite ubiquitous across all infection types, geographic areas, times, and susceptibility types. The fsr-positive E. faecalis isolates were detected in 15 of 24 (62.5%) medical centers in the United States, all Canadian hospitals (100.0%), 4 of 7 (57.1%) Latin American medical centers, and 5 of 13 (38.5%) European participant hospitals (data not shown). Rates of E. faecalis isolates with fsr were highest (63.6 to 88.9%) among urinary tract infection isolates (Table ​(Table1)1) and lowest for the bloodstream infection strains (23.5%). These findings were consistent across all regions where samples were available. Less variation in the prevalence of the fsr was noted between geographic samples of all specimen types combined. Highest fsr-positive E. faecalis rates were encountered in North America (42.4 to 44.4%). All four vancomycin-resistant E. faecalis strains were fsr positive (three isolates from the United States [New York and Massachusetts] and one isolate from Brazil). TABLE 1. Occurrences of fsr-positive strains of E. faecalis in various SENTRY Antimicrobial Surveillance Program objectives (body sites of infection) and geographic areas and nations and among vancomycin-resistant strains (109 isolates)a These results broaden the understanding of the fsr range among contemporary isolates of E. faecalis in the Americas and Europe. The prevalence of fsr can differ significantly among isolates when comparing types (sites) of infection and among geographic areas. The high prevalence in urinary tract infections described here was similar to the finding of other E. faecalis surface proteins (esp) associated with ascending urinary tract infections (6) and among hospitalized patients with antimicrobial-resistant clones (1). In contrast, the fsr-positive rate was low (23.5%) compared to that of the bacteremic endocarditis isolates reported by Pillai et al. (3), indicating a distinct difference between these two types of blood culture isolates. Some investigators have also reported that putative virulence factors among E. faecalis did not contribute to increased mortality rates (7). Other enterococcal virulence factors may be greater contributors to invasive disease and have measurable effects on patient outcomes (7). These reported findings must be further documented by resistance surveillance networks using a larger selection of enterococcal virulence genes, especially with our discovery of all of the vancomycin-resistant E. faecalis isolates (four strains) being fsr positive. Finally, these enterococcal virulence factors appear to be more prevalent among human isolates of enterococci (34.5 to 41.7%) compared to those of strains of swine or poultry origin (6 of 276 samples; 2.2%), thus minimizing food animals as a significant reservoir and source of these genetic elements (A. M. Hammerum and L. B. Jensen, Letter, J. Clin.Microbiol. 40:4396, 2002). We encourage expanded studies of these genes among human and environmental enterococcal strains to determine their epidemiologic significance.