Antimicrobial susceptibility trends among gram-positive and -negative clinical isolates collected between 2005 and 2012 in Mexico: results from the Tigecycline Evaluation and Surveillance Trial

Background The Tigecycline Evaluation and Surveillance Trial (T.E.S.T) is a global antimicrobial surveillance study of both gram-positive and gram-negative organisms. This report presents data on antimicrobial susceptibility among organisms collected in Mexico between 2005 and 2012 as part of T.E.S.T., and compares rates between 2005–2007 and 2008–2012. Method Each center in Mexico submitted at least 200 isolates per collection year; including 65 gram-positive isolates and 135 gram-negative isolates. Minimum inhibitory concentrations (MICs) were determined using Clinical Laboratory Standards Institute (CLSI) broth microdilution methodology and antimicrobial susceptibility was established using the 2013 CLSI-approved breakpoints. For tigecycline US Food and Drug Administration (FDA) breakpoints were applied. Isolates of E. coli and K. pneumoniae with a MIC for ceftriaxone of >1 mg/L were screened for ESBL production using the phenotypic confirmatory disk test according to CLSI guidelines. Results The rates of some key resistant phenotypes changed during this study: vancomycin resistance among Enterococcus faecium decreased from 28.6 % in 2005–2007 to 19.1 % in 2008–2012, while β-lactamase production among Haemophilus influenzae decreased from 37.6 to 18.9 %. Conversely, methicillin-resistant Staphylococcus aureus increased from 38.1 to 47.9 %, meropenem-resistant Acinetobacter spp. increased from 17.7 to 33.0 % and multidrug-resistant Acinetobacter spp. increased from 25.6 to 49.7 %. The prevalence of other resistant pathogens was stable over the study period, including extended-spectrum β-lactamase-positive Escherichia coli (39.0 %) and Klebsiella pneumoniae (25.0 %). The activity of tigecycline was maintained across the study years with MIC90s of ≤2 mg/L against Enterococcus spp., S. aureus, Streptococcus agalactiae, Streptococcus pneumoniae, Enterobacter spp., E. coli, K. pneumoniae, Klebsiella oxytoca, Serratia marcescens, H. influenzae, and Acinetobacter spp. All gram-positive organisms were susceptible to tigecycline and susceptibility among gram-negatives ranged from 95.0 % for K. pneumoniae to 99.7 % for E. coli. Conclusion Antimicrobial resistance continues to be high in Mexico. Tigecycline was active against gram-positive and gram-negative organisms, including resistant phenotypes, collected during the study.


Background
The widespread occurrence of antimicrobial resistance among bacterial pathogens is a global concern, and infections caused by resistant bacteria are now frequent events in hospitalized or community patients. Countries in Latin America are recognised to have high levels of resistance and antimicrobial susceptibility has decreased among many pathogens in Mexico in recent years [1][2][3][4][5][6][7][8].
For example, in one tertiary care hospital in Mexico susceptibility to meropenem among Acinetobacter baumannii decreased from 91.7 % in 1999 to 11.8 % in 2011 [8].
The Tigecycline Evaluation and Surveillance Trial (T.E.S.T.) is a global in vitro antimicrobial surveillance study which started in 2004 and began collecting grampositive and gram-negative isolates from Mexican centers in 2005. Susceptibility is tested against a range of antimicrobials, including tigecycline, by each medical center before shipping to a central laboratory; the central laboratory then carries out data validation and the collation of the T.E.S.T. database. In this report, we compare data for isolates collected between 2005-2007 and 2008-2012 as well as presenting data for 2005-2012 as a whole. This report serves as an update to some of the data presented in Rossi et al. [9], who presented data on antimicrobial susceptibility across Latin America between 2004 and 2007. Data for S. aureus collected across Latin America between 2004 and 2010 was previously published by Garza-González and Dowzicky [10] and data on gram-negative organisms collected across Latin America between 2004 and 2010 was previously published by Fernández-Canigia and Dowzicky [11]. These reports also contain data from Mexican centers which are included in this analysis.

Methods
In total, there were 16 centers in Mexico over the study period (1 center  . All centers did not participate in all years. The maximum number of years any one center participated for was 7 years. This was the case for two centers. One center participated for six years and four centers participated for 5 years. The remaining nine centers participated for between 2 and 4 years. All body sites were considered acceptable sources, although a maximum of 25 % of isolates could be urinary in origin. Inclusion of any isolate in the study was independent of patient medical history, previous antimicrobial use, age or gender. Only a single isolate was permitted from each patient. Ethics committee approval was not required as the study does not collect patient identifying information.

Antimicrobial susceptibility testing
All participating medical centres were responsible for isolate identification and susceptibility testing. Minimum inhibitory concentrations (MICs) for all pathogens and each antimicrobial agent in the T.E.S.T. panel were determined using Clinical Laboratory Standards Institute (CLSI) broth microdilution methodology [12], and either MicroScan ® panels (Dade Microscan Inc., West Sacramento, CA, USA) or Sensititre ® plates (TREK Diagnostic Systems, East Grinstead, UK). The core T.E.S.T. antimicrobial panel included: amoxicillin-clavulanate, ampicillin, ceftriaxone, imipenem or meropenem, levofloxacin, minocycline, piperacillin-tazobactam and tigecycline. Imipenem was replaced by meropenem in 2006 due to imipenem stability issues, while MicroScan ® panels were replaced by Sensititre ® plates. Gram-positive pathogens were tested against the core antimicrobials plus linezolid, penicillin and vancomycin; gram-negative isolates were tested against the core panel as well as amikacin, cefepime and ceftazidime. The S. pneumoniae test panel was expanded in 2006 to include azithromycin, clarithromycin, erythromycin and clindamycin.
All isolates of E. coli and K. pneumoniae were tested for extended-spectrum β-lactamase (ESBL) production while all H. influenzae were tested for β-lactamase production. Isolates of E. coli and K. pneumoniae with a MIC for ceftriaxone of >1 mg/L were screened for ESBL production using the CLSI phenotypic confirmatory disk test according to CLSI guidelines [13] using cefotaxime (30 µg), cefotaxime/clavulanic acid (30/10 µg), ceftazidime (30 µg), and ceftazidime/clavulanic acid (30/10 µg) disks (Oxoid, Inc., Ogdensburg, NY, USA). Mueller-Hinton agar used in testing was manufactured by Remel, Inc. (Lenexa, KS, USA). An increase of >5 mm in the inhibition zone of the combination disk when compared to that of the cephalosporin disk alone demonstrated ESBL production. H. influenzae isolates were tested for β-lactamase production using local methodologies.
Quality MIC data were used only if the daily QC test results were within ranges considered acceptable by CLSI [13].
All isolates were sent to International Health Management Associates, Inc. (IHMA, Schaumberg, IL, USA). IHMA were responsible for the organisation of isolate collection and transport and the management of a centralized database. IHMA were also responsible for carrying out isolate identification QC checks, which were conducted on approximately 10-15 % of isolates.
Antimicrobial susceptibility was established using CLSI-approved breakpoints. The 2013 version was used for all isolates in this study [13]. Tigecycline breakpoints, as published by the US Food and Drug Administration (FDA), were used in this analysis [14]. FDA tigecycline breakpoints for E. faecalis (vancomycin-susceptible) were used for all Enterococcus isolates.

Statistical analysis
Comparison of susceptibility between the 2005-2007 and 2008-2012 time periods were analysed using the Cochran-Mantel-Haenszel test with SAS (Version 8.2). Because of the large number of hypothesis tests, significance was determined at p < 0.01. The Fisher's exact test was used to analyse changes in the percentage of resistant phenotypes between the two time periods, again using SAS (version 8.2). In this test significance was defined at p < 0.05.

Results
Demographic and source data for the isolates in this study are presented in Table 1. Table 2 presents data on the antimicrobial susceptibility of gram-positive and gram-negative isolates collected in Mexico between 2005 and 2012. Among gram-positive isolates ≥99 % were susceptible to linezolid, tigecycline and vancomycin. The one exception to this was E. faecium as only 75.0 % of isolates were susceptible to vancomycin. For S. pneumoniae susceptibility data for isolates from cerebrospinal fluid (CSF) are presented separately from data for isolates from other culture sources. Among non-CSF isolates there was a statistically significant decrease in susceptibility between the two time periods for amoxicillin-clavulanate and penicillin (p < 0.01). For isolates of E faecalis rates of susceptibility were similar between 2005-2007 and 2008-2012; however, for E. faecium rates of susceptibility were higher in 2008-2012 than in 2005-2007 for all antimicrobials with less than 100 % susceptibility and in the cases of ampicillin and penicillin these differences were statistically significant (p < 0.01).
The activity of the antimicrobial panel against the gram-negative organisms varied with susceptibility to the carbapenems and tigecycline at ≥95 % against Enterobacter spp., E. coli, K. oxytoca and K. pneumoniae when examining the 2005-2012 data. Susceptibility among Acinetobacter spp. and P. aeruginosa was lower. The MIC 90 for tigecycline against Acinetobacter spp. was 2 mg/L for the 2005-2012 and 0.5 and 2 mg/L for the 2005-2007 and 2008-2012 time periods, respectively. Decreases in susceptibility among the E. coli submitted were noted for a number of antimicrobials with the largest decreases in susceptibility seen for minocycline and piperacillin-tazobactam. Both these decreases were considered statistically significant (p < 0.0001). For K. pneumoniae statistically significant (p < 0.01) decreases in susceptibility to minocycline, piperacillintazobactam, amoxicillin-clavulanate and ceftriaxone were seen between the two time periods. Susceptibility among Acinetobacter spp. was lower in 2008-2012 than in 2005-2007 for the majority of antimicrobial agents and decreases in susceptibility to amikacin, levofloxacin, meropenem, and minocycline were considered to be statistically significant (p < 0.01).
Antimicrobial susceptibility among MRSA, methicillinsusceptible S. aureus, ESBL-positive E. coli and K. pneumoniae and MDR Acinetobacter spp. are presented in Table 3. Greater than 97 % of S. aureus were susceptible to linezolid, minocycline, tigecycline and vancomycin irrespective of methicillin status. The carbapenems and tigecycline have the highest rates of susceptibility against ESBL-positive E. coli and K. pneumoniae. The MIC 90 for tigecycline against MDR Acinetobacter spp. A total of 504 carbapenem-resistant Enterobacteriaceae, Acinetobacter spp. and P. aeruginosa were identified in this study. Susceptibility data are presented in Table 4 K. oxytoca were not included in this table as only 2 isolates were identified.

Resistant phenotypes
Rates of resistant phenotypes for the three time periods are presented in Table 5. Rates of MRSA, meropenemresistant Acinetobacter spp. and MDR Acinetobacter spp. increased between 2005-2007 and 2008-2012 with the rates of meropenem-resistant and MDR Acinetobacter spp. increasing significantly (p < 0.05). In comparison, β-lactamase production among H. influenzae decreased significantly (p < 0.05) between the two time periods.

Discussion
Rates of antimicrobial resistance among both grampositive and gram-negative organisms were high in this report from Mexico. Antimicrobial resistance is a recognized problem in Latin America with high levels of resistance among both gram-positive and gram-negative organisms [1][2][3]. There is a known relationship between antimicrobial use and resistance [15], and it can be surmised that one of the factors contributing to the high rates of resistance in Mexico is overuse of antimicrobials. Data on previous antimicrobial use is not collected by T.E.S.T.; however, such data would be of interest if the T.E.S.T. program were to develop in the future.      [16]. In addition, Klebsiella spp. susceptibility to piperacillintazobactam, ceftriaxone and cefepime was approximately 7-9 % higher in the SENTRY study than in T.E.S.T. for K. pneumoniae, although susceptibility was similar among other antimicrobial agents [16]. Carbapenem resistance among K. pneumoniae and E. coli was low in T.E.S.T. as has been previously reported for Mexico [17].
There was a variation in the susceptibility of S. pneumoniae to penicillin which was dependent on the breakpoints applied. For the 2005-2012 time period among non-CSF isolates susceptibility was 33.7 % using the oral or parenteral (meningitis) breakpoints but 93.1 % when using the parenteral (non-meningitis) breakpoints.
Penicillin oral breakpoints are S ≤ 0.06 mg/L; I 0.12-1 mg/L; R ≥ 2 mg/L; whereas those for parenteral administered penicillin are: non-meningitis (S ≤ 2 mg/L; I 4 mg/L; R ≥ 8 mg/L) and meningitis (S ≤ 0.06 mg/L; R ≥ 0.12 mg/L). This highlights the importance of using the correct breakpoints when interpreting the susceptibility of an organism.
When comparing rates of resistance in Mexico with other countries the rate of MRSA (45 %) was similar to that reported for other Latin American countries such as Guatemala (49 %) and Panama (47 %) [3]. With limited treatment options, concern continues about the prevalence of MRSA globally and although rates have been reported to be decreasing in some regions, most notably North America and Europe [18,19], the prevalence in other areas, particularly developing countries, is of increasing concern [20]. In their study of isolates collected between 2010 and 2014 Conceição et al. reported rates of 61.6 % in Angola, 25.5 % in São Tomé and Príncipe, 5.6 % in Cape Verde and 0.0 % in East Timor [21]. At 22 % the rate of vancomycin-resistant E. faecium was similar to that reported for enterococci in Brazil by Jones et al. (27 %) for 2011 [3]. It is also comparable to the rate of 18.5 % reported for Saudi Arabia in 2009-2010 [22]. The rates of ESBL producing E. coli and K. pneumoniae, although high, were relatively low when compared to some other countries. Sharma et al. reported that 67 % of Klebsiella spp. and 57 % of E. coli were ESBL producers  in Jaipur, India in 2011-2012 [23]. Results for amikacin were similar in this study to the results reported for E. coli collected in Egypt but for K. pneumoniae susceptibility was higher in T.E.S.T. [24]. Antimicrobial susceptibility among P. aeruginosa was similar between the two time periods in this study and susceptibility rates were similar to those reported by Jones et al. for Latin America in 2011 [3]. In contrast, rates of susceptibility to amikacin and meropenem were lower than those reported by Gad et al. for P. aeruginosa isolates collected from three Egyptian hospitals [25]. In the case of Acinetobacter spp., resistance to meropenem increased from 17.  [26,27]. In addition a rate of 26 % was reported for a single center in India in 2013 although this was a decrease from the 33 % previously reported [28]. Other countries, such as Libya, are also reporting the emergence of carbapenem resistant A. baumannii [29]. These results demonstrate the variability in antimicrobial resistance between countries and with increasing globalization the importance of a global strategy to control the spread of resistant organisms.
Rossi et al. [9] examined the in vitro activity of tigecycline and comparator agents against gram-positive and gram-negative isolates from Latin America, including Mexico, between 2004 and 2007 as a part of the T.E.S.T. study. These data from Mexico are included in the current report but are updated with additional isolates. The most dramatic changes in susceptibility between 2005-2007 and 2008-2012 occurred among S. pneumoniae, E. faecium, K. pneumoniae and K. oxytoca: ≥10 % changes were observed for seven antimicrobial agents against non-CSF S. pneumoniae [amoxicillin-clavulanate, azithromycin, clarithromycin, erythromycin, meropenem, minocycline and penicillin (using oral or parenteral meningitis breakpoints], five antimicrobials against E. faecium (ampicillin, levofloxacin, minocycline, penicillin and vancomycin), five agents against K. pneumoniae (amoxicillin-clavulanate, cefepime, ceftriaxone, minocycline and piperacillin-tazobactam) and four antimicrobials against K. oxytoca (amoxicillin-clavulanate, cefepime, levofloxacin and piperacillin-tazobactam). All changes for E. faecium were increases in susceptibility, while for S. pneumoniae, K. pneumoniae and K. oxytoca decreases in susceptibility were seen. These changes in antimicrobial susceptibility may be due to a number of factors. Firstly, between the two time periods there was an increase in the number of isolates coming from inpatients in this study. As isolates from inpatients and outpatients are known to have different susceptibility profiles this could impact the susceptibility profile of the isolates as a whole. Also, increases in susceptibility can be due to improved antimicrobial stewardship whereas decreases in susceptibility may occur due to failures in stewardship and center specific outbreaks. Over the counter dispensing of antimicrobials is common in Latin America and in 2010 Mexico sought to enforce existing laws to reduce their consumption. This policy has been shown to have decreased consumption [30], although a trend for decreasing consumption had already been detected [31]. The relationship between antimicrobial consumption and resistance is well known.
Linezolid    As discussed above, rates of ESBL production are high in Mexico. In the current study, all ESBL-positive E. coli isolates and 95.3 % of ESBL-positive K. pneumoniae isolates were susceptible to tigecycline (data not shown). E. cloacae and S. marcescens are not examined for ESBLs as part of the T.E.S.T. study, but low levels of tigecycline non-susceptibility were observed in this study for both Enterobacter spp. (3.4 %) and S. marcescens (4.2 %). Silva-Sanchez et al. [34] have also reported good in vitro activity for tigecycline against ESBL-positive Enterobacteriaceae in Mexico (as well as MRSA), with >94 % of 1055 isolates reported as tigecycline susceptible. Tigecycline thus appears to be a potential treatment option in Mexico, where the prevalence of pathogens resistant to commonly-used antimicrobials is high.
When <10 isolates MIC 90 data are not presented and the number of isolates susceptible or resistant are presented in parenthesis  Limitations of this study include the center repetition between years, with nine of the 16 centers participating for between two and four of the 8 years of study. The types of centers involved in surveillance studies can also influence results as large university hospitals and smaller community based hospitals can have differing levels of resistance. Both university and community based hospitals submitted isolates to the T.E.S.T. program in Mexico.
Surveillance studies such as SENTRY and T.E.S.T. are critical tools for monitoring the development and spread of resistance among important clinical pathogens, assisting healthcare professionals in making appropriate judgments for the best use of antimicrobials on regional or national levels and supporting antibiotic stewardship efforts [35,36]. Tigecycline demonstrates good in vitro activity against most of the pathogens examined in this study, and should continue to be a useful option in the treatment of infectious diseases in Mexico.