Surveillance
Occurrence and antimicrobial resistance pattern comparisons among bloodstream infection isolates from the SENTRY Antimicrobial Surveillance Program (1997–2002)

https://doi.org/10.1016/j.diagmicrobio.2004.05.003Get rights and content

Abstract

The empiric treatment of patients with bloodstream infections (BSI) has become more complicated in an era of increasing antimicrobial resistance. The SENTRY Antimicrobial Surveillance Program has monitored BSI from patients in medical centers worldwide since 1997. During 1997–2002, a total of 81,213 BSI pathogens from North America, Latin America, and Europe were tested for antimicrobial susceptibility. S. aureus, E. coli, and coagulase-negative staphylococci were the three most common BSI pathogens in all three regions each year. Prevalence variability was noted in regions for some species, including higher rates of isolation of E. coli in Europe, Enterococcus spp. in North America, and Gram-negative enteric and nonenteric species in Latin America. Patient age analysis showed the most common BSI pathogen among neonates was coagulase-negative staphylococci and among elderly patients, E. coli. Resistance among BSI pathogens was much more prevalent in nosocomial infections and in patients in intensive care units (ICUs); age differences were also noted. Geographically, oxacillin-resistant S. aureus (39.1%, 2002) and vancomycin-resistant enterococci (17.7%, 2002) were highest in North America, and extended-spectrum β-lactamase–producing Klebsiella spp. (35.8–46.7%) and multidrug-resistant P. aeruginosa (18.7%, 2002) were highest in Latin America. Activity of commonly used antimicrobial agents remained relatively stable in North America, except in the case of vancomycin-resistant enterococci (20% decline between 1997 and 2002). An epidemiologic investigation of oxacillin-resistant S. aureus in North America identified 10 significant clones (ribotypes) and the common resistance patterns associated with them. Surveillance of BSI pathogens is needed to determine trends of resistance and provide useful information regarding patient risk factors and geographic differences.

Introduction

The treatment of patients with bacteremia is becoming more complicated in an era of increasing antimicrobial resistance among frequently occurring pathogens. Furthermore, the increased complexity of patients requiring hospitalization, an aging population base, and widespread use of indwelling devices has created higher risks for bacteremia. A study conducted in the United States documented a mortality rate of 17.5% among adult bacteremic patients; the study demonstrated also, however, that patients receiving appropriate antimicrobial therapy had a lower mortality rate of 13.3% (Weinstein et al., 1997). Studies have shown also that pediatric patients with nosocomial bacteremia had an overall crude mortality of 14%, and elderly patients had a mortality rate that ranged between 20% and 40% (Wisplinghoff et al., 2003, Leibovici, 1995). A mortality rate of 18% was noted among patients that had nursing home–acquired bacteremia (Mylotte et al., 2002). Patients with cancer are at high risk of acquiring bloodstream infections (BSI), and an overall mortality rate of 24.5% among patients with hematologic malignancies was observed in a recent investigation (Velasco et al., 2003). Regardless of past or current mortality related to BSI, the increasing rates of antimicrobial resistance are creating serious dilemmas for treatment of bacteremic patients, requiring the development of new therapeutic options, advanced diagnostic tests, and preventative technologies (Beekman et al., 2003, Christensen et al., 2003, Crnich and Maki, 2002). Despite such advances, accurate empiric treatment by prescribing physicians remains critical to minimize inappropriate antimicrobial therapy that may lead to poor clinical outcomes (Leibovici et al., 2001, Munson et al., 2003, Zaragoza et al., 2003).

Antimicrobial surveillance studies provide important information regarding the prevalence of pathogens responsible for bacteremia and antimicrobial resistance rate trends (Diekema et al., 1999, AMCLI Lombardia Hospital Infectious Study Group et al., 2002, Sahm et al., 1999). These types of studies can also collect patient demographic data that may be used to determine which pathogens are more prevalent among various patient populations and how antimicrobial resistance is changing among them (Collin et al., 2001, CANCER Study Group et al., 2003, Warren et al., 2001). With this information available, those who care for septic patients can have a better understanding of the possible offending pathogens and the resistance mechanisms that may be prevalent in their institution and region. If an epidemiologic element is integrated within a surveillance system network, data can be generated to better detect the global spread of key resistance mechanisms, producing more comprehensive results (SENTRY Participants Group et al., 2000, McDougal et al., 2003, Pfaller et al., 2001, Stefani and Varaldo, 2003).

The SENTRY Antimicrobial Surveillance Program has monitored (1997 to present) BSI pathogens, among other objectives, on a worldwide scale. Results from this study have been previously published relating to pathogen prevalence and antimicrobial resistance rates that can be compared to other local, national, and regional studies monitoring BSI pathogens. Most studies have documented Staphylococcus aureus and Escherichia coli as the most common causes of BSI. Coagulase-negative staphylococci (CoNS), although isolated frequently from blood, most often contaminants and not indicative of a true septic episode (Weinstein et al., 1997). During the first six months of the initial year of the SENTRY Program BSI surveillance, marked differences in antimicrobial resistances were noted for some pathogens between the United States and Canada (Pfaller et al., 1998). In Latin America, a unique pattern of antimicrobial resistance was also noted when compared to North American BSI isolates, including elevated rates of drug-resistant Gram-negative bacteria and the lack of vancomycin-resistant enterococci (VRE) (Sader et al., 1999). The first six months of the 1998 North American surveillance showed a similar rank order of pathogen prevalence (S. aureus ranked first) compared to the previous year with a consistent frequency of occurrence for the top five bacterial pathogens associated with BSI (Diekema et al., 2000). However, that study documented an increase in oxacillin resistance in both S. aureus and coagulase-negative staphylococci. Another surveillance program in the United States conducted between 1995 and 1997 showed an increase in vancomycin-resistant Enterococcus faecium, an increase of staphylococcal strains with a vancomycin minimum inhibitory concentration (MIC) of 4 ug/mL, and an indication that Klebsiella pneumoniae isolated from blood were more likely to have an extended-spectrum β-lactamase (ESBL) phenotype when compared to strains isolated from other infection sources (Sahm et al., 1999).

Data from more focused surveillance of bacteremic patients are also available. The number of BSI isolates originating from bone marrow transplant patients during seven years of surveillance showed that a previously higher Gram-positive organism infection rate was reversing during the study interval and that streptococci became more prevalent than staphylococci during the majority of the study years (Collin et al., 2001). Mortality in these patients was attributable mainly to Gram-negative pathogens. A study of neutropenic patients showed that the susceptibility profiles of pathogens isolated, including strains from BSI, were similar to those of other hospital strains (Mutnick et al., 2003). A notable difference between pathogens from patients in different age groups, including frequency of isolation and antimicrobial susceptibility profiles, was documented during the first three years of the SENTRY Program (Diekema et al., 2002).

This report provides a comprehensive evaluation of BSI isolates collected by the SENTRY Program during the first six years (1997–2002). The data presented include comparisons of the frequency of occurrence among three regions (North America, Latin America, and Europe) and the resistance rates among the most prevalent pathogens isolated from these regions. More detailed antimicrobial susceptibility/resistance statistics are presented for North American isolates. Pathogen prevalence is stratified by patient age, source of infection (hospital versus community), hospital location (ICU versus non-ICU patients), and by the primary diagnosis of the patient. Finally, a molecular epidemiologic study of oxacillin-resistant S. aureus isolated from BSI occurring in North American centers participating in the SENTRY Program (1998–2001) is presented.

Section snippets

Participating institutions

The geographic regions in this study included North America, Latin America, and Europe during 1997 through 2002. Medical centers were selected for their commitment to long-term participation and for representation of their respective geographic regions (North America, Latin America, Europe; 1997–2002). Although site participation varied for some medical centers during the six-year study period, overall geographic representation was maintained and included a total of 107 medical center

Results

The rank order of the top 10 bacterial pathogens that were submitted and the frequency with which they occurred in BSI in North and Latin America and Europe during the entire study period for each year monitored are presented in Table 1. Overall, the 10 listed pathogens accounted for 89.4–92.1% of all isolates. S. aureus, E. coli, and CoNS were the three most frequently isolated pathogens in all three regions. However, E. coli was isolated more frequently in Europe and ranked first each year;

Discussion

This six-year overview of a very large collection of BSI isolates from three continents (>81,000 strains) tested by reference methodologies provides both useful and interesting information regarding pathogen prevalence and rates of resistance by geography and patient demography.

The frequency of occurrence for bacterial species causing BSI in the three areas of the world was quite variable in certain instances. Overall, it was shown that Gram-positive species were causative agents in >57% of

Acknowledgements

The secretarial assistance of K Meyer is appreciated. We also recognize the efforts of Kelley Fedler, L Deshpande, and Paul Rhomberg for ribotyping and PFGE analysis and general technical guidance. The SENTRY Program is supported by an educational/research grant from Bristol-Myers Squibb.

References (30)

  • D.J. Diekema et al.

    Genetic relatedness of multidrug-resistant, methicillin (oxacillin) resistant Staphylococcus aureus bloodstream isolates from SENTRY antimicrobial resistance surveillance centers worldwide, 1998

    Microbiol Drug Resist

    (2000)
  • A.W. Karchmer

    Nosocomial bloodstream infectionsOrganisms, risk factors, and implications

    Clin Infect Dis

    (2000)
  • L. Leibovici

    Bacteraemia in the very old

    Features and treatment Drugs Aging

    (1995)
  • L. Leibovici et al.

    Departmental consumption of antibiotic drugs and subsequent resistanceA quantitative link

    J Antimicrob Chemother

    (2001)
  • F. Luzzaro et al.

    Prevalence and drug susceptibility of pathogens causing bloodstream infections in northern ItalyA two-year study in 16 hospitals

    Eur J Clin Microbiol Infect Dis

    (2002)
  • Cited by (0)

    View full text