Skip Navigation

Survey of Infections Due to Staphylococcus Species: Frequency of Occurrence and Antimicrobial Susceptibility of Isolates Collected in the United States, Canada, Latin America, Europe, and the Western Pacific Region for the SENTRY Antimicrobial Surveillance Program, 1997–1999

  1. D. J. Diekema1,
  2. M. A. Pfaller1,
  3. F. J. Schmitz2,
  4. J. Smayevsky3,
  5. J. Bell4,
  6. R. N. Jones1,a,
  7. M. Beach1,a, and
  8. the SENTRY Participants Group
  1. 1 Medical Microbiology Division, Department of Pathology, University of Iowa College of Medicine, Iowa City, Iowa
  2. 2 Institute for Medical Microbiology and Virology, University Hospital, Düsseldorf, Germany
  3. 3 Microbiology Laboratory, C.E.M.I.C., Buenos Aires, Argentina
  4. 4 Women and Children's Hospital, Adelaide, Australia
  1. Reprints or correspondence: Dr. Daniel J. Diekema, Medical Microbiology Division, C606 GH, Dept. of Pathology, University of Iowa College of Medicine, Iowa City, IA 52242 (daniel-diekema{at}uiowa.edu).
  • a Present affiliation: The JONES Microbiology Institute, North Liberty, Iowa.

Abstract

Between January 1997 and December 1999, bloodstream isolates from 15,439 patients infected with Staphylococcus aureus and 6350 patients infected with coagulase-negative Staphylococcus species (CoNS) were referred by SENTRY-participating hospitals in the United States, Canada, Latin America, Europe, and the Western Pacific region. S. aureus was found to be the most prevalent cause of bloodstream infection, skin and soft-tissue infection, and pneumonia in almost all geographic areas. A notable increase in methicillin (oxacillin) resistance among community-onset and hospital-acquired S. aureus strains was observed in the US centers. The prevalence of methicillin (oxacillin)-resistant S. aureus varied greatly by region, site of infection, and whether the infection was nosocomial or community onset. Rates of methicillin resistance were extremely high among S. aureus isolates from centers in Hong Kong and Japan. Uniformly high levels of methicillin resistance were observed among CoNS isolates. Given the increasing multidrug resistance among staphylococci and the possible emergence of vancomycin-resistant strains, global strategies are needed to control emergence and spread of multiply resistant staphylococci.

Staphylococci are a major cause of both hospital-acquired and community-onset infections. Previous reports from the SENTRY surveillance program underscore the prominence of Staphylococcus aureus as an etiologic agent of bloodstream infection [13], skin and soft-tissue infection [4], and pneumonia [5]. S. aureus was found to be the most frequently isolated pathogen causing bloodstream infection in the United States, Canada, and Latin America, accounting for 22.6% of all bloodstream infections reported by the SENTRY program during a 12-month period [2]. The same study revealed that coagulase-negative Staphylococcus species (CoNS) were the third most common cause of blood-stream infection. These findings confirm those of the National Nosocomial Infection Surveillance (NNIS) and Surveillance and Control of Pathogens of Epidemiologic Importance (SCOPE) programs regarding the emerging importance of CoNS as a cause of nosocomial bacteremia [6, 7]. CoNS are more prevalent in the NNIS and SCOPE programs, which track only nosocomial pathogens, than in the SENTRY program, which collects isolates from both nosocomial and community-onset infections.

Since methicillin resistance in S. aureus was first reported in the early 1960s [8], increasing rates of methicillin resistance among S. aureus and CoNS have been a cause for concern. The NNIS survey found that the prevalence of methicillin resistance among nosocomial S. aureus isolates increased from 2.1% in 1975 to 35% in 1991 [9]. Methicillin resistance among CoNS has also increased. NNIS data from 1980 to 1989 indicated that the incidence of CoNS resistant to methicillin increased from 20% to 60% [7]. In addition, S. aureus and CoNS, particularly those strains acquired in the hospital, have become resistant to multiple other antimicrobial agents.

For this reason, glycopeptide agents (vancomycin and teicoplanin) have been considered the only available antibiotics uniformly active against multidrug-resistant staphylococci. Therefore, reports of staphylococci with reduced susceptibility to these agents are alarming. In 1997, Hiramatsu et al. described the first clinical S. aureus isolate with intermediate resistance to vancomycin (MIC, 8 µg/mL) [10, 11]. Subsequently, similar strains with reduced susceptibility to vancomycin were identified in Europe and the United States [12].

The emergence of CoNS strains with reduced susceptibility to glycopeptides has also been described [13, 14]. Sieradzki et al. reported an increased incidence of CoNS isolates with decreased susceptibility to teicoplanin in hospitals where glycopeptide agents were widely used [15]. Although vancomycin is still the first choice for treatment of severe multidrug-resistant staphylococcal infections, glycopeptide susceptibility of staphylococci can no longer be assumed, and vigilance is necessary to monitor the frequency of this emerging problem.

Data derived from global surveillance studies such as SENTRY can provide important information about the changing spectrum and regional variation of antimicrobial resistance patterns. This information can be used to assist in the design of empirical treatment regimens and also to plan control measures for emerging antimicrobial-resistant staphylococci.

Materials and Methods

Study design. The SENTRY Antimicrobial Surveillance Program was established to monitor the predominant pathogens and antimicrobial resistance patterns of nosocomial and community-onset infections via a broad network of sentinel hospitals distributed by geographic location and size. The monitored infections include bacteremia (objective A), outpatient respiratory infections (objective B), pneumonia in hospitalized patients (objective C), wound infections (objective D), and urinary tract infections (objective E). Participating institutions in 1997–1998 included 30 medical centers in the United States, 8 in Canada, 10 in Latin America, and 24 in Europe.

The US sites were located in 23 different states, and the Canadian sites were located in 7 provinces. The Latin American sites were located in 6 different countries (Argentina, Brazil, Chile, Colombia, Mexico, and Uruguay), and the European sites were located in 13 different countries (Albania, Austria, Belgium, France, Germany, Greece, Italy, Poland, Portugal, Spain, Switzerland, The Netherlands, and Turkey). In 1999, the number of participating sites in Europe was reduced from 24 to 10 (located in France, Greece, Israel, Spain, Switzerland, and Turkey). In 1998, the Western Pacific region was added to SENTRY, which included 17 hospitals located in 7 different nations (Taiwan, Japan, China, Australia, South Africa, Singapore, and The Philippines).

Each participating hospital contributed findings (organism identification, date of isolation, and antimicrobial susceptibility profile) on the first 20 consecutive episodes of bacteremia (yielding blood culture isolates from separate patients that were judged to be clinically significant), lower respiratory tract infection, urinary tract infection, and skin/soft-tissue infection in each calendar month. Additional demographic and epidemiological data were recorded on a data form to be included with each isolate. All isolates were saved on agar slants and sent on a weekly basis to either the University of Iowa College of Medicine (Iowa City, IA) or the Eijkmann-Winkler Institute for Microbiology, at Utrecht University (Utrecht, The Netherlands) for storage and for further characterization by reference identification methods and susceptibility testing. This report focuses on all isolates of Staphylococcus species collected in the 3 years between 1 January 1997 and 31 December 1999.

Organism identification. All isolates were identified at the participating institution by the routine methodology in use at each laboratory. On arrival at the University of Iowa, isolates were subcultured to blood agar to ensure viability and purity. The species identification was confirmed with the Vitek system (bioMérieux Vitek), API (bioMérieux) products, or conventional methods as required. Isolates were frozen at −70°C until processed.

Susceptibility testing. Antimicrobial susceptibility testing of isolates was performed by reference broth microdilution methods as described by the National Committee for Clinical Laboratory Standards (NCCLS) [16]. Microdilution trays were purchased from MicroScan and PML Microbiologicals. Antimicrobial agents were obtained from their respective manufacturers. Quality control was performed by testing with S. aureus American Type Culture Collection 29213. Interpretive criteria for each antimicrobial tested were those published by the NCCLS [17] or cited in text and tables.

Results

Between 1997 and 1999, a total of 15,439 S. aureus and 6350 CoNS isolates were referred by SENTRY participating hospitals in the United States, Canada, Latin America, Europe, and the Western Pacific region. Tables 1 and 2 summarize the total number of isolates by region and site of infection. As shown in table 3, S. aureus was the most frequent etiologic agent causing bloodstream infection, skin and soft-tissue infection, and lower respiratory tract infection in all geographic areas combined.

Table 1

Total number of Staphylococcus aureus isolates collected from each SENTRY region, by site of infection, 1997–1999.

Table 2

Total number of coagulase-negative staphylococcus isolates collected from each SENTRY region, by site, 1997–1999.

Table 3

Percentage of bloodstream, lower respiratory tract, and skin/soft-tissue infections due to Staphylococcus aureus at SENTRY centers in the United States, Canada, Latin America, Europe, and the Western Pacific region, 1997–1999.

Methicillin resistance rates of S. aureus by region and site of infection are shown in figure 1. Overall, the prevalence of methicillin-resistant S. aureus (MRSA) (from all sites of infection) varied as follows: Western Pacific region, 46% (657/1427); United States, 34.2% (2455/7169); Latin America, 34.9% (682/1956); Europe, 26.3% (916/3477). Rates were substantially lower in Canada, where MRSA accounted for only 81 (5.7%) of 1410 S. aureus isolates. Isolates from patients with pneumonia demonstrated the highest overall rates of methicillin resistance (figure 1). Methicillin resistance rates were higher among S. aureus bloodstream isolates from nosocomial infections than among community-onset strains (figure 2). However, an increase in the methicillin resistance rate among S. aureus was seen among both nosocomial and community-onset strains during the 3-year period of surveillance in the United States, the region in which the greatest number of centers were involved throughout the course of the study (figure 3).

Figure 1

Methicillin resistance rates among Staphylococcus aureus isolates, by region and by site of infection, in the SENTRY program, 1997–1999

Figure 2

Methicillin resistance rates among bloodstream isolates of nosocomial vs. community-onset Staphylococcus aureus, by region, in the SENTRY program, 1997–1999.

Figure 3

Methicillin resistance rates of Staphylococcus aureus strains causing bloodstream infections in the United States, SENTRY program, 1997–1999.

Resistance of S. aureus to methicillin (oxacillin) varied greatly among countries within a region (tables 46). At European centers, methicillin resistance rates ranged from ⩽2% (Switzerland and The Netherlands) to 54.4% (Portugal), and among Western Pacific countries, percentages of MRSA strains ranged from 23.6% (Australia) to >70% (Japan and Hong Kong). It should be noted that most countries in Latin America, Europe, and the Western Pacific region were represented by only 1–3 hospitals, which may not be representative of overall “national” rates of resistance. With regard to CoNS, high rates of oxacillin resistance (>70%) were found worldwide; in contrast to S. aureus, no major differences were observed by region (figure 4).

Table 4

Variation in rates of methicillin resistance among Staphylococcus aureus isolates, by nation, at Western Hemisphere SENTRY centers, 1997–1999.

Table 5

Variation in rates of methicillin resistance among Staphylococcus aureus isolates, by nation, at European SENTRY centers, 1997–1998.

Table 6

Variation in rates of methicillin resistance among Staphylococcus aureus isolates, by region, at Western Pacific SENTRY centers, 1998–1999.

Figure 4

Methicillin resistance rates of coagulase-negative staphylococci, by region (all sites of infection), in the SENTRY program during 1997–1999

MRSA strains are also more likely to be resistant to other antimicrobial agents than are methicillin-susceptible S. aureus isolates. Regional co-resistance patterns for all MRSA isolates are shown in table 7, in which rates of resistance to representatives of 8 antimicrobial classes are delineated. Latin American MRSA isolates were resistant to a median of 6 antimicrobial classes, whereas US and Canadian strains demonstrated resistance to a median of 3 additional antimicrobial classes. Agents for which there were substantial regional differences in resistance included chloramphenicol (57.9% resistant in Latin America vs. <10% resistant in all other regions), rifampin (range, 4.9% resistant in Canada to 44.4% in Europe), tetracycline (14.8% and 15.6% resistant in the United States and Canada, respectively, vs. 82% resistant in the Western Pacific region), and gentamicin (range, 25.9% resistant in Canada to 91.2% resistant in Latin America). High levels of resistance to erythromycin, clindamycin, and ciprofloxacin were found among MRSA in all regions.

Table 7

Co-resistance patterns of MRSA, by region, among all MRSA isolates recovered during 1997–1999 (n = 4788).

The in vitro activity of selected antimicrobial agents tested against S. aureus and CoNS isolates is summarized in tables 817. Data for β-lactam agents (cefazolin, ceftriaxone, cefepime, amoxicillin-clavulanate, and imipenem) are reported only for methicillin-susceptible strains. As outlined in table 7, Latin American MRSA isolates tended to be more resistant to almost all antimicrobial agents tested than were isolates from other geographic areas. Methicillin-susceptible strains, on the other hand, were generally >90% susceptible to other tested agents, with the exception of erythromycin (range, 12% resistant in Canada to 22% resistant in the United States).

Table 8

In vitro broth microdilution susceptibility findings for Staphylococcus aureus isolates from United States SENTRY participants, 1997–1999.

Table 9

In vitro broth microdilution susceptibility findings for Staphylococcus aureus isolates from Canadian hospitals in the SENTRY program, 1997–1999.

Table 10

In vitro broth microdilution susceptibility findings for Staphylococcus aureus isolates from Latin America SENTRY participants, 1997–1999.

Table 11

In vitro broth microdilution susceptibility findings for Staphylococcus aureus isolates from European hospitals in the SENTRY program, 1997–1999.

Table 12

In vitro broth microdilution susceptibility findings for Staphylococcus aureus isolates from Western Pacific hospitals in the SENTRY program, 1998–1999.

Table 13

In vitro broth microdilution susceptibility findings for coagulase-negative Staphylococcus species (CoNS) isolates from United States hospitals in the SENTRY program, 1997–1999.

Table 14

In vitro broth microdilution susceptibility findings for coagulase-negative Staphylococcus species (CoNS) isolates from Canadian hospitals in the SENTRY program, 1997–1999.

Table 15

In vitro broth microdilution susceptibility findings for coagulase-negative Staphylococcus species (CoNS) isolates from Latin America hospitals in the SENTRY program, 1997–1999.

Table 16

In vitro broth microdilution susceptibility findings for coagulase-negative Staphylococcus species (CoNS) isolates from European hospitals in the SENTRY program, 1997–1999.

Table 17

In vitro broth microdilution susceptibility findings for coagulase-negative Staphylococcus species (CoNS) isolates from Western Pacific hospitals in the SENTRY program, 1998–1999.

Table 18 compares glycopeptide resistance rates among S. aureus and CoNS strains from different regions. Fewer than 1% of S. aureus isolates in our 3 years of global surveillance had vancomycin MICs of 4 µg/mL, and all but 1 remained susceptible to vancomycin. For 1 strain from a center in Hong Kong, the broth microdilution MIC was 8 µg/mL. In addition, 5 S. aureus strains isolated in the United States, Latin America, and Europe (0.03%–0.1%) showed reduced susceptibility to teicoplanin (MIC, >16 µg/mL).

Table 18

Rates of glycopeptide resistance among Staphylococcus species isolates in the SENTRY program, 1997–1999.

CoNS isolates were also uniformly susceptible to vancomycin, although a small fraction (0.6%–2.6%) of CoNS strains for which vancomycin MICs were 4 µg/mL were observed in all geographic areas. In addition, CoNS isolates with reduced susceptibility to teicoplanin were found in all regions. Percentages of strains for which teicoplanin MICs were >16 µg/mL ranged from 0.4% (9 of 2068 European isolates) to 3.9% (33 of 848 Latin American isolates).

The in vitro activity of the new agents quinupristin-dalfopristin and linezolid, compared with that of vancomycin, is summarized in table 19. These agents demonstrated excellent in vitro activity against both S. aureus and CoNS isolates. All strains but 1 (a methicillin-resistant CoNS isolate) were inhibited by linezolid at an MIC ⩽4 µg/mL. Linezolid yielded MIC90 values against S. aureus strains that were 1 dilution higher than those for CoNS isolates. Quinupristin-dalfopristin was equally active in vitro against both S. aureus and CoNS isolates, with 98.2%–99.8% of strains being susceptible to this agent.

Table 19

In vitro broth microdilution susceptibility testing results with quinupristin-dalfopristin, linezolid, and vancomycin against all Staphylococcus species isolates from SENTRY participating hospitals in the United States, Canada, Latin America, Europe, and the Western Pacific region, 1997–1999.

Rates of categorical agreement between reference and participant sites for oxacillin resistance (using mecA gene detection as the gold standard) among S. aureus isolates exceeded 98% (table 20). However, very major errors (false susceptibility) ranged from 0% to 26.3% per center.

Table 20

Agreement between data from reference site and SENTRY participant sites for oxacillin resistance among bloodstream isolates of Staphylococcus aureus, January–October 1999.

Discussion

The SENTRY Antimicrobial Surveillance Program was designed to track antimicrobial resistance trends and the spectrum of microbial pathogens on a global scale. The SENTRY program has unique features that distinguish it from other excellent surveillance programs, such as the NNIS and Intensive Care Antimicrobial Resistance Epidemiology (ICARE) projects [7, 9, 18], the SCOPE [6] program, and The Surveillance Network [19]. Whereas these programs may track only nosocomial infections and/or rely primarily on susceptibility testing results from participating centers, the SENTRY program monitors both nosocomial and community-onset infections on a global scale with use of validated reference identification and susceptibility testing methods at a central monitoring laboratory [13].

We found S. aureus to be the most common cause overall of bacterial infections involving the bloodstream, lower respiratory tract, and skin/soft tissue. A wide array of virulence mechanisms, an ability to persist in adverse environments, and an extraordinary potential to develop antimicrobial resistance all contribute to the success of this organism as a human pathogen [20, 21]. Our study describes the ongoing emergence of antimicrobial-resistant S. aureus worldwide.

Since the emergence of the first MRSA strains in the early 1960s [8], the spread of MRSA has been reported worldwide [2, 3, 6, 9, 18, 2226]. As our data confirm, elevated rates of methicillin resistance among S. aureus are found in many regions of the world. In the United States, which has the largest number of centers followed over the time period of the SENTRY study, a steady increase in methicillin resistance was found among both community and nosocomial S. aureus isolates. Data from the NNIS/ICARE and SCOPE programs indicate very similar rates of methicillin resistance among S. aureus isolates from US hospitals; specifically, among nosocomial bloodstream isolates, 30%–40% of S. aureus strains are now MRSA [6, 19], and the rate for the most recent time period (SENTRY 1999) is even higher (45%). There is less published information about “pan-European” rates of methicillin resistance, but we demonstrate an increase in such resistance, from a rate of 12.8% reported in the early 1990s [27] to 26.3% during SENTRY surveillance.

Our data also confirm the findings of Voss et al. [27] that methicillin resistance rates are highest in the nations of southern Europe (e.g., Italy, Greece, Portugal, and Turkey). The very highest rates we noted of methicillin resistance among S. aureus isolates were in the Asia-Pacific region (>60% in Taiwan, Singapore, Japan, and Hong Kong). Previous reports from this region have documented similarly high rates of methicillin resistance [28].

We report that many MRSA strains are also resistant to multiple other antimicrobial classes; indeed, some are only susceptible to glycopeptides. The recent emergence of S. aureus isolates with reduced susceptibility to vancomycin has therefore heightened concern about possible alternatives for treatment against these multidrug-resistant strains [1012, 29, 30]. We found only 1 strain of S. aureus and no CoNS strains that were nonsusceptible to vancomycin; however, some strains have vancomycin MICs of 4 µg/mL and teicoplanin MICs >16 µg/mL. How many of these and other MRSA strains harbor subpopulations of organisms with elevated vancomycin MICs (so-called hetero-VISA [vancomycin-intermediate S. aureus]) is not known, and this work is ongoing.

However, data from the European SENTRY centers reveal hetero-VISA strains to be rare and restricted to a few locations [31, 32]. Nonetheless, these hetero-VISA strains may represent a substrate for future development of vancomycin-resistant Staphylococcus species, with the prerequisites being deep-seated infection and prolonged exposure to vancomycin. A previous report based on the database of The Surveillance Network [19] indicates that although the percentage of S. aureus strains in the United States with vancomycin MICs of 4 µg/mL is low (<1%), it may be increasing.

CoNS traditionally have been considered low-virulence pathogens. However, since the 1980s, CoNS have been increasingly recognized as a prevalent cause of hospital-acquired infections. NNIS data reveal that the incidence of CoNS among nosocomial bloodstream infections increased from 9% to 27% between 1980 and 1989. The NNIS and SCOPE programs rank CoNS as the leading cause of nosocomial bloodstream infection, whereas they are the third most common cause of bloodstream infection (community and nosocomial combined) and the second most common cause of nosocomial bloodstream infection in the SENTRY program [2, 6, 7]. The extremely high rates of oxacillin resistance (77.3%–76.9%) that we found among CoNS isolates are consistent with those reported from the SCOPE (80.4%) and ICARE (75%) programs [6, 18]. Furthermore, most of these strains are also resistant to several other classes of antibiotics, which no doubt has contributed to the widespread use of glycopeptides over the past 2 decades.

Because of the increasing multidrug resistance among staphylococci and the possible emergence of vancomycin-resistant strains, new antimicrobials are needed as alternative agents against these multiply resistant strains. The streptogramin combination quinupristin-dalfopristin and the oxazolidinone linezolid are among the most promising drugs under investigation, showing potential activity against a variety of multiply resistant gram-positive organisms [3335].

In the SENTRY study, vancomycin remains highly active in vitro against S. aureus and CoNS isolates. However, quinupristin-dalfopristin and linezolid also demonstrated excellent activity in vitro against almost all strains tested. All strains but 1 oxacillin-resistant CoNS isolate remained susceptible to linezolid, whereas only 0.5% and 0.1% of MRSA and methicillin-resistant CoNS isolates, respectively, were resistant to quinupristin-dalfopristin. Molecular typing of the European quinupristin-dalfopristinresistant MRSA has revealed evidence of clonal relatedness among many of the strains [36]. Measures to control the transmission and spread of MRSA will undoubtedly preserve the activity of this and other agents. Linezolid and quinupristin-dalfopristin will be available as promising alternatives to vancomycin for multidrug-resistant staphylococci, pending additional data regarding clinical efficacy against serious, deep-seated staphylococcal infections.

Although high levels of methicillin resistance in CoNS isolates were noted in all regions, important differences in MRSA rates were found among different geographic areas. The reason for these regional differences in methicillin resistance rates is difficult to establish, although variations in antimicrobial usage and in infection control practice are likely explanations. High rates of multidrug resistance may be due either to the selection of antimicrobial-resistant strains under antimicrobial pressure or to the widespread horizontal transmission of multidrug-resistant strains. In the case of S. aureus, many investigators have described the ability of MRSA strains to spread widely and rapidly to become regionally prevalent strains [26, 37]. Molecular typing results in a 6-month study of genetic relatedness of multidrug-resistant MRSA isolates from the SENTRY study revealed that a small number of highly related strains have established themselves both within regions and across continents [38].

Because these “epidemic clones” are particularly successful at disseminating within and between institutions, any new resistance pattern (e.g., resistance to glycopeptides, streptogramins, or oxazolidinones) could spread rapidly. Therefore, global strategies to prevent the spread of multidrug-resistant staphylococci are urgently needed. Measures should be aimed at controlling overuse and misuse of antibiotics, but perhaps more important is the application of uniform infection-control practices to prevent transmission. For example, in The Netherlands an aggressive strategy of vigilant surveillance and infection control measures has been very effective in controlling the spread of MRSA [39]. Specific recommendations for controlling the emergence of glycopeptide-resistant staphylococci also have recently been published [28, 40].

In conclusion, increasing rates of methicillin resistance among staphylococci, the recent emergence of glycopeptide-intermediate strains, and the introduction of new antimicrobial agents (e.g., linezolid and quinupristin-dalfopristin) increase the importance of local, national, and international surveillance programs. Data generated from these programs can assist in the design of appropriate measures for controlling the emergence and spread of antimicrobial resistance in an attempt to limit the scope of this disturbing worldwide problem.

Acknowledgments

The authors thank Kay Meyer for expert assistance in preparation of the manuscript. The authors also thank all SENTRY site participants who contributed isolates to this study.

References

| Table of Contents