In 2003, the Infectious Diseases Society of America published guidelines on the appropriate selection of antimicrobials for the management of patients with complicated intra-abdominal infections (IAIs) (
14). These guidelines addressed the choice of antimicrobials for both empirical and directed antimicrobial therapy for both community-associated and hospital-associated infections/pathogens. Several surveillance studies and reports have documented the increased levels of antimicrobial resistance to many of the recommended agents with significant awareness of the incidence of extended-spectrum-beta-lactamase (ESBL)-positive
Escherichia coli and
Klebsiella pneumoniae/oxytoca (
2,
5,
7,
8,
10). In 2002, the Study for Monitoring Antimicrobial Resistance Trends (SMART) was initiated to monitor yearly the antimicrobial susceptibility of Gram-negative bacilli associated with both community- and hospital-associated IAIs. This report documents the
in vitro susceptibility of the most common pathogens associated with IAI in the United States in 2007-2008 to ertapenem, imipenem, and 10 comparator antimicrobial agents.
(This work was presented in part as a poster [poster 814] at the 47th Annual Meeting of the Infectious Diseases Society of America, Philadelphia, PA, 29 October to 1 November 2009.)
Up to 100 consecutive aerobic or facultative anaerobic Gram-negative bacilli were collected from patients with complicated IAIs from 18 hospital sites in the United States representing all 9 CDC regions in 2007-2008. Antimicrobial susceptibility testing and ESBL detection/confirmation were performed using broth microdilution as previously described (
5). Reference strains recommended by the Clinical and Laboratory Standards Institute (CLSI) for quality control were utilized, and MICs were interpreted according to CLSI guidelines (
3). A total of 1,036 isolates were collected in 2007-2008 (Table
1). The most commonly isolated pathogens were as follows (percentage of total is shown in parentheses):
E. coli (43.1),
K. pneumoniae (17.7),
Pseudomonas aeruginosa (13.3),
Enterobacter cloacae (8.4),
Proteus mirabilis (3.0),
Klebsiella oxytoca (2.7),
Enterobacter aerogenes (2.3),
Citrobacter freundii (1.7), and
Serratia marcescens (1.5). Overall, members of the family
Enterobacteriaceae accounted for 84.7% of all pathogens and 80.4% of the group of pathogens containing only species with greater than 10 isolates represented. Bacteria that were not members of the family
Enterobacteriaceae, including
P. aeruginosa and
Acinetobacter spp., accounted for most of the remaining pathogens.
Against all members of the family
Enterobacteriaceae or members of the
Enterobacteriaceae with more than 10 isolates per species in this study, the most active agents (percent susceptible ≥ 90%) included the carbapenems (ertapenem and imipenem), amikacin, cefepime, and piperacillin-tazobactam. Ampicillin-sulbactam was the least active agent, with only 58% of all members of the
Enterobacteriaceae susceptible (Table
1).
E. coli, the most commonly isolated pathogen, was susceptible to most antimicrobials, with only ciprofloxacin, levofloxacin, and ampicillin-sulbactam displaying susceptibilities of <82%.
E. coli with an ESBL phenotype exhibited a distinctly higher level of resistance to most antimicrobials overall. The carbapenems (ertapenem and imipenem) retained their activity against ESBL-positive
E. coli, while the activity of amikacin, ciprofloxacin, levofloxacin, piperacillin-tazobactam, and ampicillin-sulbactam was considerably diminished. Table
1 shows the percent susceptibility of cephalosporins to ESBL-positive
E. coli. However, despite the reported percent susceptibilities, CLSI recommends that for clinical purposes, all cephalosporins should be reported as resistant (
3). However, in 2010, the CLSI lowered the breakpoints for several extended-spectrum cephalosporins (ceftazidime, ceftriaxone, and cefotaxime) to more accurately reflect the pharmacokinetics/pharmacodynamics of these agents and to better ensure that ESBL-positive
E. coli isolates and
K. pneumoniae/oxytoca isolates are currently reported as resistant (
4). Applying the new breakpoints to a global collection of ESBL-positive
Enterobacteriaceae resulted in 7%, 31%, and 6% of these isolates susceptible to ceftriaxone, ceftazidime, and cefotaxime, respectively (
1). Continued confirmation testing of ESBL status may still be required for epidemiologic purposes.
K. pneumoniae, the second most commonly isolated pathogen, displayed overall susceptibility of ≥90% for the carbapenems, cefepime, and amikacin. ESBL-positive
K. pneumoniae, in comparison to ESBL-positive
E. coli, was less susceptible to ertapenem (78% versus 100%), imipenem (94% versus 100%), and ampicillin-sulbactam (3% versus 24%). ESBL production was detected in 4.7% and 17.5% of
E. coli and
K. pneumoniae isolates, respectively.
E. coli and
K. pneumoniae/K. oxytoca accounted for 63.5% of all members of the
Enterobacteriaceae isolated, indicative of their prevalence in IAI. Of the remaining
Enterobacteriaceae isolates, susceptibilities were very similar for
C. freundii,
E. aerogenes/cloacae,
P. mirabilis, and
S. marcescens. Each of these species was >90% susceptible to ertapenem, imipenem, and amikacin (Table
1). The
in vitro activity of ciprofloxacin and levofloxacin varied among the IAI pathogens. Among all members of the
Enterobacteriaceae, susceptibility of ≥90% was noted only for
E. aerogenes,
K. oxytoca, ESBL-negative
K. pneumoniae,
P. mirabilis, and
S. marcescens. Piperacillin-tazobactam activity varied with ESBL phenotype but was consistently lower than ertapenem or imipenem. Amikacin was the only antimicrobial with >90% activity against
P. aeruginosa, while 83%, 78%, 76%, and 72% of isolates were susceptible to piperacillin-tazobactam, imipenem, ceftazidime, and ciprofloxacin, respectively.
Surveillance studies investigating the susceptibility of pathogens to frequently used IAI antimicrobials have documented the decreased activity of recommended antimicrobials over the past 10 years. However, specific data for the United States describing susceptibility patterns for IAI pathogens from the United States remains limited. In this study, percent susceptibility of recommended IAI antimicrobials, such as imipenem and ciprofloxacin, varied by year (Fig.
1). Comparing
Enterobacteriaceae and non-
Enterobacteriaceae isolates each year from 2002 to 2008 in the United States from the SMART study, the susceptibility to both imipenem and ciprofloxacin vary. Only the decreased
in vitro activity of ciprofloxacin versus
Enterobacteriaceae reached statistical significance (
P < 0.001) when comparing 2002 to 2008. One focus of global surveillance studies has been to document the increasing incidence of ESBL-producing
E. coli and
K. pneumoniae/oxytoca (
6,
11-
13). The global increase in ESBL-positive and fluoroquinolone-resistant
Enterobacteriaceae has challenged standard therapeutic treatment regimens. The identification of CTX-M-type ESBLs as predominant in certain areas of the United States from 2000 to 2006 has altered antimicrobial treatment algorithms with decreased susceptibility to fluoroquinolones among CTX-M phenotypes (16%) compared to 36% among SHV or TEM phenotype (
9). The specific examination of antimicrobial susceptibility among IAI pathogens has been under way since 2002 in the SMART study with the primary goal of monitoring changes in resistance for commonly prescribed antimicrobials used for empirical and directed therapy of Gram-negative bacilli in IAIs. Between 2004 and 2007, studies of global isolates from IAIs in the SMART program demonstrated that the susceptibilities of Gram-negative isolates to the carbapenems have remained stable (
2,
6,
13), with
E. coli and
K. pneumoniae isolates, including ESBL-positive isolates, remaining between 98 and 100% susceptible to ertapenem and imipenem. Susceptibility to amikacin has remained high with the percent susceptible continuously over 90 in almost all countries examined (
6). The emergence of ESBL-positive
E. coli and
K. pneumoniae, however, has brought with it the reduced activity of fluoroquinolones (ciprofloxacin and levofloxacin) against these phenotypes (
5). In 2004, >77% of
E. coli and >84% of
K. pneumoniae of mixed phenotypes were susceptible to both ciprofloxacin and levofloxacin. By 2007, only 17% of ESBL-positive
E. coli and 70% of ESBL-negative
E. coli remained susceptible to ciprofloxacin (
5,
13). Similar declines in susceptibility were observed for
K. pneumoniae and other species of
Enterobacteriaceae (
5). In the United States, the levels of resistance to recommended antimicrobials for the treatment of IAI infections has not been examined in depth. In 2003, among 378
E. coli isolates collected from patients with IAI in 23 countries, all ESBL-negative isolates were susceptible to both ertapenem and imipenem and >80% of isolates were susceptible to ciprofloxacin. However, in ESBL-positive isolates, 96.2% remained susceptible to ertapenem and 99.1% remained susceptible to imipenem, while only 31% of isolates were susceptible to ciprofloxacin (
12). Clearly, as the incidence of ESBL-positive
E. coli and
K. pneumoniae/oxytoca phenotypes increases, surveillance studies continue to document the decreasing activity of recommended IAI antimicrobials with the exception of the carbapenems. Although
P. aeruginosa accounts for <10% of all Gram-negative pathogens isolated in IAIs in other studies (
6), increasing carbapenem, ceftazidime, and fluoroquinolone resistance in this pathogen will continue to complicate empirical antimicrobial therapy.
Acknowledgments
We thank all of the investigators and centers involved in the study.
This study was funded by Merck & Company.