Longitudinal Assessment of Antimicrobial Susceptibility among Gram-Negative and Gram-Positive Organisms Collected from Italy as Part of the Tigecycline Evaluation and Surveillance Trial between 2004 and 2011

The Tigecycline Evaluation and Surveillance Trial (T.E.S.T.) was initiated in 2004 to longitudinally monitor the activity of the broad-spectrum glycylcycline antimicrobial tigecycline, and a suite of comparator agents, against an array of clinically important bacterial pathogens worldwide. In this report, we examine the activity of tigecycline and comparators against a collection of 13,245 clinical isolates, both Gram-positive (n = 4,078 and Gram-negative (n = 9,167), collected from 27 centres in Italy between 2004 and 2011. Susceptibility was established according to Clinical Laboratory Standards Institute guidelines. Tigecycline and linezolid exhibited very good activity against Gram-positive pathogens, with MIC90s ranging from 0.06 to 0.25 mg/L and 1–4 mg/L, respectively; vancomycin and the carbapenems also showed good activity against select Gram-positive pathogens. Tigecycline was the most active agent against Gram-negative pathogens (except P. aeruginosa), with MIC90s ranging from 0.25–2 mg/L (16 mg/L for P. aeruginosa). Amikacin and the carbapenems also possessed good activity against many Gram-negative pathogens here. ESBL-positive E. coli increased in prevalence from 2004 to 2011, while ESBL-positive Klebsiella spp., vancomycin-resistant enterococci and MRSA decreased in prevalence. Linezolid, tigecycline and vancomycin susceptibility were very stable over the course of this study, while susceptibility to ampicillin, piperacillin-tazobactam, ceftriaxone and levofloxacin varied over time according to pathogen; minocycline and cefepime susceptibility among several pathogens decreased during this study.

No breakpoints were available for tigecycline, although the lowest overall MIC 90 was observed for tigecycline (2 mg/L; Table 2). A. baumannii were most susceptible to minocycline with 81.0% susceptibility over all study years combined. Isolates of A. baumannii from Italy had low susceptibility (<40%) to most of the remaining T.E.S.T. panel agents: amikacin, cefepime, ceftazidime, ceftriaxone, levofloxacin, meropenem and piperacillin-tazobactam. The susceptibility of isolates to amikacin, cefepime, ceftazidime, ceftriaxone and levofloxacin all increased in 2005 before declining again from 2006 to 2011. From 2006 onwards, there was an overall increase in the resistance of A. baumannii isolates to amikacin, cefepime, ceftazidime, ceftriaxone, levofloxacin, and piperacillin-tazobactam ( Table 2). Very few A. baumannii isolates (<1%) were resistant to minocycline between 2004 and 2007, but this percentage rose to 7.9% and 7.6%, respectively, in 2009 and 2010 before decreasing to 2.5% in 2011. A total of 526 MDR A. baumannii isolates were collected in Italy during T.E.S.T. (Table 4), which represented 60.0% of all A. baumannii isolates. Over the surveillance period, MDR A. baumannii isolates had ≥79.5% resistance to all agents tested with the exception of minocycline (5.1%). Among A. baumannii, 79.7% were susceptible to imipenem ( Table 2) while 49.0% of MDR isolates were susceptible to imipenem (Table 4).

E. coli
Amikacin, meropenem and tigecycline were all effective against E. coli, with >95% susceptibility during all years of surveillance ( Table 2). Susceptibility of E. coli to cefepime was 93.4% in 2004, but this decreased to 77.4% in 2011, with some fluctuation in susceptibility during the years in between.  Table 5). Tigecycline, meropenem and amikacin were highly active against ESBL-positive E. coli (100%, 98.8% and 94.2% susceptible, respectively) in this study (Table 6).

Discussion
Tigecycline is licensed in Italy to treat complicated intra-abdominal and skin and soft tissue infections. Good clinical results for tigecycline have been demonstrated previously: for example, Bassetti et al. [11] showed a 76.5% (13/17) success rate for tigecycline in the treatment of complicated skin and soft tissue infections and an 82.8% (72/87) success rate against peritonitis. This clinical success is reflected by high tigecycline susceptibility for most pathogens over the course of the T.E.S.T. study between 2004 and 2011: Gram-negative pathogens showed >94% tigecycline susceptibility while Gram-positives were >98% susceptible in the present report.
The two Gram-negative organisms in this study for which tigecycline breakpoints were not available were A. baumannii and P. aeruginosa. Against A. baumannii, tigecycline had the lowest overall MIC 90 (2 mg/L) of all T.E.S.T. agents in Italy. A previously published Italian surveillance study comprising nine hospitals also determined an MIC 90 of 2 mg/L for tigecycline against A.
baumannii isolates collected between 2003 and 2004 [12]. As in the current analysis, the one-year study found that A. baumannii isolates had low (≤51%) susceptibility to amikacin, cefepime, ceftazidime, meropenem and piperacillin-tazobactam. The proportion of A. baumannii resistant to piperacillin-tazobactam in the 2003-2004 study (44%) was lower than to the overall value in this T.E.S.T. paper (63.3%; 2004-2011). However, the 2003-2004 frequency of imipenem-resistant isolates (50%) was around 36% higher than T.E.S.T. (13.7%; 2004-2006). Both studies used CLSI methodologies to determine imipenem susceptibility, but there were 27 countrywide centres collecting isolates for T.E.S.T., compared with nine centres in the central-south region of Italy for the 2003-2004 study. The high incidence of imipenem resistance in the one-year study could, therefore, be due to a localised outbreak of imipenem-resistant A. baumannii infections in this region of Italy during 2003-2004. A. baumannii gained resistance to most antimicrobials on the panel over the course of the T.E.S.T. study in Italy as described in this report, and the greatest increase in resistance was to piperacillin-tazobactam (49.1% from 2004 to 2011). There were also >40% increases in resistance to amikacin and levofloxacin.
Epidemiological studies of A. baumannii in Italy have focussed mostly on carbapenem-resistant or MDR isolates. Lambiase et al. [13] examined 567 A. baumannii isolates from an ICU in Naples between 2007 and 2010, and found that all isolates were MDR, including carbapenems; these isolates were clonal in nature, all possessing the bla OXA-51-like and bla OXA-58-like genes. D'Arezzo et al. [14] reported a high (60.5%) prevalence of elevated (MIC≥128 mg/L) resistance to imipenem among 111 A. baumannii isolates, associated with bla OXA-58-like (22.8%) or bla OXA-51-like (71.1%) genes. Most of these isolates (95.6%) were related to international clonal lineage II. In a study of six hospitals in Florence, Donnarumma et al. [15] showed three main clonal groups of A. baumannii, A1, A2 and A3; A1 was genetically related to the European EU II clone. All isolates possesses the bla OXA-51-like gene, and 65% of these isolates were resistant to imipenem. As reported globally [16], clonal lineages appear to predominate among A. baumannii isolates in Italy, although there may be variation between different geographical areas.
Compared with A. baumannii, a lower overall proportion of MDR P. aeruginosa (19.1%) was recorded in Italy during T.E.S.T. In 2010, a similar percentage of invasive P. aeruginosa isolates from Italy (20.8%) was reported to have resistance to three or more antibiotic classes among aminoglycosides, carbapenems, ceftazidime, fluoroquinolones, and piperacillin-tazobactam [19]. The individual resistance values for ceftazidime and piperacillin-tazobactam against these invasive isolates were 21.2% and 17.7%, respectively, slightly lower than the 2010 T.E.S.T. results in Italy (32.9% resistance for both agents).
The highest proportion of resistant Gram-negative pathogens in the current Italian study was observed among ESBL-producing E. coli (24.8% from 2004-2011). This is a dramatic increase compared to the 10.8% occurrence of ESBL-positive E. coli isolates nationally in 1999 [20]. An Italian single-hospital surveillance report from 2004 to 2007 identified 23.5% of E. coli as ESBL producers [21]. A more recent (2009-2010) study of another Italian hospital found that the most frequently observed multidrug-resistant pathogen was ESBL-producing E. coli (18.6% of all multidrug-resistant isolates) [22]. These data are in line with the findings from the current T.E.S.T. manuscript.
ESBL production has been linked with third-generation cephalosporin resistance [23]. The T.E.S.T. data showed that the resistance of E. coli to ceftriaxone in Italy increased more than two-fold between 2007 and 2008, during which time the prevalence of ESBL-producing E. coli isolates almost doubled. In 2010, the ECDC annual report found that in one Italian centre, all 23 of the invasive E. coli isolates resistant to third-generation cephalosporins were ESBL-producers [19]. In a 2007-2008 single hospital study of 13 countries, including Italy, higher mortality rates and longer hospital stays were associated with third-generation cephalosporin-resistant E. coli bloodstream infections [24]. These findings suggest that E. coli resistance in Italy is increasing, possibly due to the spread of ESBL-positive strains.
High proportions of ESBL-producing K. pneumoniae were also measured during T.E.S.T. (24.1% over all years). ESBL production has been associated with reduced carbapenem susceptibility in K. pneumoniae due to a loss of bacterial membrane permeability in some ESBL-producing isolates [25]. Some important differences are observed when resistance rates in Italy are compared to global rates, as reported in Pfizer's online T.E.S.T. database [36]. In general, resistance levels to β-lactams (particularly ceftriaxone) and levofloxacin appear to be higher in Italy than globally. A. baumannii resistance to most antimicrobial agents is higher in Italy by approximately 15%, with the exceptions of amikacin and levofloxacin (resistance in Italy is > 20% higher) as well as imipenem and minocycline (resistance is similar). E. coli resistance is also high in Italy: ampicillin and cefepime resistance are around 7% higher, while ceftriaxone and levofloxacin resistance are approximately 12% higher than globally. Among Enterobacter spp., resistance was roughly 10% higher in Italy to ceftriaxone, levofloxacin and piperacillin-tazobactam. Amoxicillin-clavulanate, cefepime, ceftriaxone, levofloxacin and piperacillin-tazobactam resistance among Italian isolates of K. pneumoniae are approximately 6-8% higher than isolates globally. Resistance is roughly 6-10% higher among isolates of P. aeruginosa from Italy to all antimicrobial agents (excluding amikacin, to which resistance is the same in Italy as globally). Among S. pneumoniae, macrolide, minocycline and clindamycin resistance are about 10%, 12% and 14% higher in Italy, respectively, while penicillin resistance is approximately 7% lower than the global average. Implementation of and strict adherence to resistance control measures, such as ongoing resistance surveillance, improved hand hygiene/increased glove use and/or the use of antimicrobial stewardship programs [37,38], would almost certainly help to reduce the high levels of resistance observed in Italy.
The T.E.S.T. study, like all surveillance studies, suffers from inherent limitations. Although several centres participated in the T.E.S.T. study in Italy over 2004-2011, some contributed isolates over several years while others participated in a few years or even just one, causing fluctuations in isolate contribution both geographically and over time. Thus, regional variations in resistance in a given study year may have had a disproportionate influence on apparent national resistance levels.
Over all T.E.S.T. years, rates of β-lactamase-producing H. influenzae, ESBL-producing K. oxytoca, vancomycin-resistant E. faecalis and PRSP were ≤11.4% in Italy. Another encouraging result of this study was the identification of no vancomycin-resistant Enterococci in Italy in 2011. The above findings may indicate that these drug-resistant organisms are becoming less prevalent in Italian hospitals, and therefore, less of a threat to the welfare of patients. Other pathogens in Italy, including E. coli and K. pneumoniae, have shown increased resistance in recent years, due largely to the spread of ESBL-positive strains. Thus, these organisms must continue to be monitored for further changes in susceptibility in the future. The results of surveillance studies such as T.E.S.T. help members of the healthcare industry to monitor rates of in vitro susceptibility among important pathogens to widely used antimicrobial agents, both globally and regionally.

Conclusions
Tigecycline and linezolid exhibited very good activity against Gram-positive pathogens in Italy, with MIC 90 s ranging from 0.06 to 0.25 mg/L and 1-4 mg/L, respectively. Vancomycin and the carbapenems also showed good activity against select Gram-positive pathogens. Tigecycline was the most active agent against Gram-negative pathogens (with the exception of P. aeruginosa), with MIC 90 s between 0.25 and 2 mg/L (but 16 mg/L for P. aeruginosa), while amikacin and the carbapenems also possessed good activity against many Gram-negative pathogens. Linezolid, tigecycline and vancomycin susceptibility were stable over the course of this study, but ampicillin, piperacillin-tazobactam, ceftriaxone and levofloxacin susceptibility varied by pathogen; minocycline and cefepime susceptibility decreased among several pathogens. ESBL-positive E. coli increased while ESBL-positive Klebsiella spp., vancomycin-resistant enterococci and MRSA decreased in prevalence during this study.