Background

Tigecycline is a glycylcycline licensed by the US Food and Drug Administration (FDA) for the treatment of complicated skin and skin structure infections (cSSSI), complicated intra-abdominal infections (cIAIs) and community acquired bacterial pneumonia (CAP). The Tigecycline Evaluation and Surveillance Trial (T.E.S.T.) is a global surveillance study with the aim of assessing and reporting the antimicrobial susceptibility of tigecycline and comparator agents globally, regionally, and for individual countries. T.E.S.T. was initiated in 2004 and to date 60 countries have contributed with Gram-positive and Gram-negative isolates and susceptibility data. Antimicrobial surveillance studies, such as T.E.S.T., play a key role in charting antimicrobial resistance.

The Latin American region is recognized as facing a significant challenge with high levels of antimicrobial resistance among important Gram-negative organisms including Escherichia coli and Klebsiella spp. and the non-fermenters Acinetobacter spp. and Pseudomonas aeruginosa[1–3]. In recent years, extended-spectrum β-lactamases (ESBLs) have increased in type and frequency among Enterobacteriaceae and carbapenemases have emerged[4, 5]. In the case of the non-fermentative Gram-negative bacilli multidrug-resistance is an increasing problem with limited, or no treatment option[6].

In this report we present data from the Latin American region of Gram-negative isolates collected between 2004 and 2010. The isolates collected between 2004 and 2007 were previously reported by Rossi et al.[7].

Methods

Organism collection

Gram-negative isolates were collected from 12 countries in Latin America between 2004 and 2010. Centers were distributed as follows: 12 in Argentina, 3 in Brazil, 5 in Chile, 14 in Colombia, 1 in El Salvador, 4 in Guatemala, 2 in Honduras, 1 in Jamaica, 15 in Mexico, 1 in Nicaragua, 2 in Panama, and 6 in Venezuela. The Gram-negative isolates submitted were consecutive and determined to be clinically significant using local criteria. Permissible clinical sources included blood, respiratory tract, urine (limited to no more than 25% of all isolates), skin, wound, and fluids. For each year, each participant center was required to identify and conduct susceptibility tests on Acinetobacter spp. (15), E. coli (25), Enterobacter spp. (25), Serratia spp. (10), Klebsiella spp. (25) and Haemophylus influenzae (15). A single isolate per patient was accepted. Inclusion in the study was independent of the patient’s medical history, previous antimicrobial use, sex and age. No banked or stored isolates were accepted.

Antimicrobial susceptibility testing

Each study center carried out antimicrobial susceptibility testing using broth microdilution methodology (Sensititre® plates [TREK Diagnostic Systems, West Sussex, England] or MicroScan® panels [Siemens, Sacramento, CA, USA]) as described by the Clinical and Laboratory Standards Institute (CLSI)[8]. Gram-negative isolates were tested against amikacin, amoxicillin-clavulanate, ampicillin, cefepime, ceftazidime, ceftriaxone, imipenem, levofloxacin, meropenem, minocycline, piperacillin-tazobactam, and tigecycline. In 2006, unreliability of the imipenem testing led to a switch from MicroScan® panels with imipenem to Sensititre® plates with meropenem. The presence or abscence of β-lactamase among H. influenzae was determined using the preferred method of each center.

Quality control strains used in the testing were E. coli ATCC 25922 and P. aeruginosa ATCC 27853. Confirmation of isolate identification and management of a centralized database were performed by a central laboratory (Laboratories International for Microbiology Studies, a division of International Health Management Associates, Inc. [IHMA, Schaumburg, IL, USA]).

Antimicrobial susceptibility was determined using CLSI interpretive criteria[9, 10]. For tigecycline, the FDA approved breakpoints, as provided in the package insert, were used[11].

Extended spectrum β-lactamase (ESBL) determination

Testing for ESBL production was carried out on isolates of E. coli and Klebsiella spp. according to the CLSI guidelines[9]. The methodology used Mueller-Hinton agar (Remel, Inc., Lenexa, KS, USA) and cefotaxime (30 μg), cefotaxime-clavulanic acid (30/10 μg), ceftazidime (30 μg), and ceftazidime-clavulanic acid (30/10 μg) discs (Oxoid, Inc., Ogdensburg, NY, USA). Quality control was carried out using K. pneumoniae ATCC 700603 (ESBL-positive) and E. coli ATCC 25922 (ESBL-negative).

Multidrug-resistant Acinetobacter baumannii

Multidrug resistance among isolates of A. baumannii was defined as resistance to levofloxacin, amikacin, carbapenems (imipenem and/or meropenem), ceftazidime and piperacillin-tazobactam.

Results

Antimicrobial susceptibility data on 16 232 Gram-negative isolates collected in Latin America between 2004 and 2010 are presented in Table1. Susceptibility among the E. coli isolates (both ESBL and non-ESBL producers) was >95% for carbapenems and tigecycline. Susceptibility to amikacin was also >95% against non-ESBL producing E. coli (MIC90 8 μg/mL) but decreased to 89.7% against ESBL producers (MIC90 32 μg/mL). A total of 24.3% of the E. coli collected from Latin America were identified as ESBL producers with percentages of ESBL production varying from 11.2% (58/519) in Colombia to 40.3% (31/77) in Honduras (Figure1). Data on susceptibility to imipenem and meropenem by country are presented in Table2. Among E. coli isolates, ESBL producers displayed slightly lower susceptibility to meropenem than non-ESBL producing isolates.

Table 1 Antimicrobial activity against Gram-negative organisms collected from Latin America (2004 – 2010)
Full size table
Figure 1
figure 1

Percentage of Escherichia coli and Klebsiella pneumoniae isolates identified as ESBL producers in each Latin American countryainvolved in T.E.S.T. (2004–2010). E. coli N values: Argentina, 101/769; Brazil, 43/247; Chile, 94/271; Colombia, 58/519; Guatemala, 81/263; Honduras, 31/77; Mexico, 398/1044; Panama, 16/100; Venezuela, 32/218; Latin America, 870/3581. K. pneumoniae N values: Argentina, 270/694; Brazil, 105/214; Chile, 147/243; Colombia, 81/432; Guatemala, 96/189; Honduras, 55/75; Mexico, 191/754; Panama, 35/89; Venezuela, 36/209; Latin America, 1045/2962. a Data from El Salvador, Jamaica and Nicaragua are not included in the analysis by country because fewer than 50 isolates were collected; however, their data are included in the total for Latin America.

Full size image
Table 2 Antimicrobial susceptibility (%S) to the carbapenems among Gram-negative organisms collected from individual countries (2004 – 2010)
Full size table

The most active antimicrobial agents against non-ESBL producing K. pneumoniae were tigecycline (MIC90 1 μg/mL), carbapenems (imipenem MIC90 0.5 μg/mL and meropenem MIC90 0.25 μg/mL) and amikacin (MIC90 8 μg/mL) (Table1). All tested antimicrobial agents displayed reduced activity against ESBL-producing K. pneumoniae, with only imipenem and tigecycline recording percentage susceptibilities of >90% (96.0% and 93.7%, respectively). In particular, susceptibilities to levofloxacin against ESBL-producing isolates of E. coli and K. pneumoniae were lower when compared with non-ESBL-producing strains (11.5% vs. 60.9% and 38.2% vs 80.1%, respectively) (Table1). Among K. pneumoniae 35.3% were ESBL producers and percentages ranged from 17.2% (36/209) in Venezuela to 73.3% (55/75) in Honduras (Figure1). Both ESBL and non-ESBL-producing K. pneumoniae displayed higher resistance levels to carbapenemes than E. coli in all countries (Table2).

Amikacin, carbapenems and tigecycline were the most active agents against K. oxytoca (>94% susceptibility) and Enterobacter spp. (>89% susceptibility). Against isolates of S. marcescens the carbapenems and tigecycline were the most active agents (>91% susceptibility) (Table1). Among these three species rates of susceptibility to the carbapenems were ≥90% in all countries where data were available, with the exception of susceptibility to meropenem among isolates of Enterobacter spp. collected in Guatemala and Honduras (79.0% and 85.3%, respectively) and susceptibility to imipenem among isolates of S. marcescens from Mexico (88.5%) (Table2).

Almost all of antimicrobials in the panel were active against H. influenzae with susceptibility varying from 78.7% for ampicillin to 100% for ceftriaxone, imipenem, levofloxacin, and meropenem (Table1). Almost 20% of isolates (181/908) were β-lactamase producers.

For A. baumannii susceptibility was less than 50% for seven of the nine antimicrobial agents (Table1). The most active agents were minocycline (89.4%, MIC90 8 μg/mL) and imipenem (62.5%, MIC90 ≥32 μg/mL). Tigecycline showed good activity against A. baumannii: although no breakpoints are available for this agent, 95.8% of the isolates displayed an MIC ≤2 μg/mL. Low rates of carbapenem susceptibility were observed in most countries (Table2); the lowest rates were reported for meropenem among isolates from Argentina (15.0%) and Panama (16.7%). A total of 600 isolates (33.2%) were multidrug-resistant, among them the MIC90 for minocycline and tigecycline were 8 and 2 μg/mL, respectively.

Among P. aeruginosa collected the most active agents were piperacillin-tazobactam, with 75.3% of isolates susceptible (MIC90 ≥256 μg/mL), and amikacin with 71.8% (MIC90 ≥128 μg/mL) (Table1).

Discussion

This study reports on rates of antimicrobial susceptibility among important Gram-negative organisms collected from centers in Latin America between 2004 and 2010. It provides an update to the work of Rossi et al.[7] who reported on Gram-negative and Gram-positive organisms collected as part of T.E.S.T. between 2004 and 2007. The isolates reported on by Rossi et al.[7] are also included in the dataset studied in this report. Rates of ESBL-producing E. coli and K. pneumoniae are similar to the mentioned study and are also similar to those reported by Villegas et al.[3] for Latin American isolates collected in 2008 as part of the SMART study.

This study shows important variations in the rate of ESBL production by country, reaching values around 40% in E. coli and >50% for K. pneumoniae, which are similar to those observed in the Asia/Pacific region by Farrell et al.[12] for both organisms and by Hawser et al. 2009[13] for E. coli. However, it should be noted that these rates may be affected by the type of infection and population analyzed in each particular center or even by ward[2]. Considering that these are common nosocomial pathogens causing severe morbidity and mortality in critically ill patients and that the available choices of antibiotic treatments for these microorganisms are seriously reduced, there is increasing clinical concern for successful patient management where ESBL isolates are prevalent. Antimicrobial susceptibility rates were lower among ESBL-producing isolates when compared with non-ESBL producers with the exception of tigecycline, imipenem and meropenem where little or no changes in susceptibility (<6.0%) were observed between both groups. ESBL-producing K. pneumoniae are frequently associated with multidrug resistance[14]. In particular, susceptibility to commonly-used antimicrobials including piperacillin-tazobactam and fluoroquinolones was reduced among ESBL-producing isolates. The worrying increase in resistance to these antibiotics among ESBL-producing organisms has been associated with the simultaneous presence of other resistance determinants[15–17]. The most common risk factor for resistance to fluoroquinolones in ESBL-producing strains is a previous history of high-level consumption of both extended-spectrum cephalosporin and quinolone antibiotics. These antibiotics are widely used in the region: Wirth et al. reported an increased use of fluoroquinolones in Latin America over a period of 10 years (1997–2007), where in some countries consumption doubled or even tripled[18].

It has been previously reported that tigecycline and carbapenems, along with amikacin, are highly active against the Enterobacteriaceae collected from countries in Latin American[19, 20]. In the current study, susceptibility to tigecycline ranged between 99.8% against ESBL-producing E. coli to 93.7% against ESBL-producing K. pneumoniae. Imipenem susceptibility ranged between 100% against K. oxytoca to 91.7% against S. marcescens and meropenem susceptibility ranged between 98.6% against non-ESBL-producing E. coli to 89.0% against ESBL-producing K. pneumoniae. The range of tigecycline MICs was greater than reported by Rossi et al.[7] against E. coli, K. pneumoniae, and Enterobacter spp.; however, this was due to single isolates at the top of the testing range (MIC ≥32 mg/L).

It is worth noting that resistance to meropenem has been observed across Latin America among members of the Enterobacteriaceae. The situation may not appear as poor for imipenem, with higher rates of susceptibility reported. However, it should be noted that imipenem susceptibility testing stopped in 2006 and switched to meropenem, meaning that the results for meropenem give a more current picture of carbapenem susceptibility in Latin America. In the late 1990s and early part of the 21st century, carbapenem resistance in Enterobacteriaceae was infrequent and resistance mechanisms were related to the presence of ESBL or overproduction of AMP-C β-lactamases associated with reduced outer membrane permeability[21, 22]. Enterobacteriaceae producing carbapenemases were first reported in the USA[23] and have now been reported in various parts of the world, including several countries in Latin America where class A carbapenemase KPC-2 enzymes are prevalent[5, 24–26]. The results of this study, along with reports of decreasing susceptibility to imipenem among Klebsiella spp. in Latin America[27] demonstrate the importance of antimicrobial resistance surveillance and further analysis of the carbapenem-resistant Enterobacteriaceae identified in this dataset is warranted.

H. influenzae are frequently susceptible to available antimicrobials. In this study susceptibility was >98% to the agents tested, with the exception of ampicillin (78.7% susceptible) largely due to the production of β-lactamase. This is in agreement with the global T.E.S.T. findings published by Garrison et al.[28].

A. baumannii is a problematic organism frequently associated with multidrug resistance and 33.2% of the isolates in this study were defined as such. The antimicrobial with the highest rate of susceptibility against the whole A. baumannii population was minocycline. Tigecycline was also active, with 95.8% of isolates displaying an MIC ≤2mg/L. These results are similar to those reported by Rossi et al.[7] for Latin America isolates collected between 2004 and 2007 and Garrison et al.[24] who reported on a global collection from the T.E.S.T. study collected between 2004 and 2007. Susceptibility to the carbapenems was 62.5% for imipenem and 33.9% for meropenem which are lower than the global rates reported by Garrison et al. (82.3% and 59.0%, respectively) and lower than the Latin American rates reported by Gales et al.[29] for Acinetobacter spp. collected between 2001 and 2004 (86.4% and 83.6%, respectively). Susceptibility also varied by country, Tognim et al.[30] reported as part of the SENTRY study that carbapenem resistance among Acinetobacter spp. varied between countries within Latin America with Argentina a particular ‘hot spot’ of resistance. Our results suggest this is a continuing situation with the lowest rates of susceptibility to meropenem reported among isolates from Argentina.

Conclusions

Surveillance of antimicrobial susceptibility plays a key role in guiding appropriate antimicrobial therapy. In this study the carbapenems and tigecycline continue to be active against the Enterobacteriaceae and A. baumannii; however, there is cause for concern with carbapenem non-susceptible isolates reported in all countries included in this study. The in vitro activity (MIC90) of tigecycline was similar to that reported for isolates collected during Phase 3 clinical trials[31].