Changes in Antibiotic Resistance Level of Nosocomial Pseudomonas Aeruginosa Isolates in the Largest University Hospital of Lithuania

The aim was to estimate changes in the resistance rates of Pseudomonas aeruginosa (P. aeruginosa) strains isolated from patients treated in intensive care units of the largest university hospital. Materials and Methods. Isolates were identified with the Phoenix ID system (Becton Dickinson, USA). The minimum inhibitory concentration (MIC) of ceftazidime, ciprofloxacin, and amikacin were determined by the E-test and evaluated following the recommendations of the Clinical Laboratory Standards Institute. Results. In 2003, the proportion of P. aeruginosa strains resistant to piperacillin was greatest followed by strains resistant gentamicin and ciprofloxacin. In 2008, the resistance rates markedly changed being the highest to ciprofloxacin. An increase in the resistance rates to ciprofloxacin (+24%, P<0.001) and ceftazidime (+8.3%, P<0.05) was documented. In 2003, there were 66.7% of P. aeruginosa strains sensitive to all antibiotics tested, and this percentage decreased to 47.5% in 2008 (P<0.05). During the study, a significant increase in the median MICs for ciprofloxacin and amikacin was observed (P<0.001); however, no significant change was documented for ceftazidime. Conclusions. P. aeruginosa remains an important nosocomial pathogen with relatively high overall resistance to antimicrobial agents, and the resistance level is increasing.


Introduction
Pseudomonas aeruginosa (P.aeruginosa) is identifi ed as a microorganism of normal microfl ora in healthy individuals, but it may cause severe infections in a critically ill and immunocompromised host.Resistance rates of P. aeruginosa to antibiotics are increasing annually and vary in different settings: outpatients, inpatients, intensive care unit (ICU) patients, patients with cystic fi brosis, etc., moreover, it differs from country to country (1,2).Despite of advances in hospital care and introduction of a wide variety of antimicrobial drugs in clinical practice, it remains a dominative nosocomial pathogen in an ICU, particularly in mechanically ventilated patients (3)(4)(5).Nosocomial pneumonia takes part in more than half of all infections acquired in an ICU, and often P. aeruginosa is isolated as a causative microorganism (6,7).Treatment of P. aeruginosa infections is usually diffi cult; mortality rates are high (3,8).Worse prognosis related to higher virulence of a pathogen was reported (4), and pathogens such as P. aeruginosa were found to be associated to excess mortality, especially in case of ventilator-acquired pneumonia (9).Inap-propriate antimicrobial treatment is proven to increase mortality of ICU patients as well, and it is associated with the resistance of clinically important pathogens such as P. aeruginosa (4,10,11).They are resistant to many antibiotics such as antipseudomonal penicillins, third-generation cephalosporins, fl uoroquinolones, and aminoglycosides.Acquired P. aeruginosa drug resistance is frequently observed among nosocomial isolates, and it often involves more than one antibiotic class (12,13).Emergence of the evolution of antibiotic-resistant mutants in bacterial population under antibiotic selective pressure and development of multidrug resistance (MDR) of P. aeruginosa has become a major social issue because of increased costs of treatment and poor outcome (1,6,(14)(15)(16).MDR is identifi ed more frequently, and isolates of P. aeruginosa resistant to all antipseudomonal agents (extreme drug resistance, XDR) are being increasingly reported (17).The easy and changing acquisition of resistance in P. aeruginosa requires setting-specifi c surveillance, which is crucial for guiding physicians on the probable susceptibility in their patients especially when treating nosocomial infections, predicting future local trends, and comparing with situation in other countries.
The aim of our study was to estimate the dynamics in the resistance rates of P. aeruginosa strains isolated from patients in the ICUs of the Hospital of Lithuanian University of Health Sciences (HLUHS) (former Kaunas University of Medicine), the largest university hospital of Lithuania, by analyzing the resistance rates in 2008 and comparing them with those in 2003, which provides a reference for the estimation of increasing resistance.As the hospital had more than 2200 beds (counting 3 million population of the country) during the study, we assumed our data to be representative of the general status for Lithuania.

Materials and Methods
The present study investigated the sensitivity of all P. aeruginosa strains isolated from the respiratory tract of all patients (n=191) treated in ICUs (one isolate per patient) of the HLUHS during years 2003 (n=90) and 2008 (n=101).Isolates were defi ned as nosocomial if a patient spent more than 48 hours in the hospital.Pseudomonas strains were selected on Pseudomonas agar with cetrimide (Liofi lchem, Italy) according to the manufacturer's instructions for the identifi cation of P. aeruginosa.Cetrimide inhibits a wide variety of bacterial species including Pseudomonas species other than P. aeruginosa.It develops a blue-green pigment due to pyocyanin and fl uorescein production.Isolates suspected to be P. aeruginosa or not clearly showing blue-green pigment were further identifi ed with the Phoenix ID system (Becton Dickinson, USA) to confi rm the strains of P. aeruginosa.Antimicrobial susceptibility of all the isolates was tested against piperacillin (100 μg), piperacillin/tazobactam (100 μg/10 μg), ceftazidime (30 μg), cefepime (30 μg), ciprofl oxacin (5 μg), gentamicin (10 μg), and amikacin (30 μg) (Becton Dickinson, USA) by the Kirby-Bauer disc diffusion method on Mueller-Hinton agar following the recommendations of the Clinical Laboratory Standards Institute (17).A number of paper discs, each impregnated with a calibrated concentration of an antibiotic, were placed onto an agar plate inoculated with bacteria.A visible zone of growth inhibition of susceptible bacteria formed around some discs according to their antibiotic concentration.The prevalence of MDR was investigated among isolates of P. aeruginosa tested with piperacillin, ceftazidime, ciprofl oxacin, and gentamicin.Isolates resistant to three or all four of mentioned antimicrobials like representatives of different classes of antibiotics with antipseudomonal activity were considered MDR.
The minimum inhibitory concentrations (MICs) of ceftazidime, ciprofl oxacin, and amikacin were determined by the E-test (AB Biodisk, Solna, Sweden) according to manufacturer's instructions and evaluated following the recommendations of the Clinical Laboratory Standards Institute (18).An E-test strip, which contains a gradient of an antibiotic, was placed on an inoculated agar plate, and the pattern of bacterial growth was examined after 24 hours.In using a gradient of an antibiotic, the E-test has a greater precision than the disc diffusion method, allowing better ascertainment of the actual MIC (18).The E-test was used to estimate so-called MIC creep, i.e., an increase in MICs of P. aeruginosa strains for tested antibiotics in 2003 and 2008 over time, as it refl ects a decreased sensitivity to antibiotics and is associated with decreased clinical effi cacy of antibiotics and higher mortality rates.
Statistical Analysis.Comparison of means between groups was performed by using the Student t test or Mann-Whitney U test (nonparametric values).Proportions were compared using the chi-square or Fisher exact test.Differences were considered signifi cant at P<0.05.The statistical package SPSS 13.0 for Windows release was used for data analysis.
Table 2 summarizes the contribution of piperacillin-, ceftazidime-, ciprofl oxacin-, and gentamicinresistant (as representatives of different antibiotic classes with antipseudomonal activity) P. aeruginosa strains to the development of MDR phenotypes.In 2003, the percentage of P. aeruginosa strains sensitive to all tested antibiotics was higher than resistant to one, two, or three-four antibiotics (68.9% vs. 8.9%, 14.4%, and 7.8%; P<0.05), and the ratio of sensitive to resistant strains was found to be decreased in 2008 (52.5% vs. 19.8%,14.9%, and 12.9%; P<0.05).The percentage of MDR P. aeruginosa strains increased insignifi cantly from 7.8% (7/90)  there was a signifi cant increase in the percentages of isolates with MIC of ≤ median 2003 MIC for ciprofl oxacin and amikacin, but the change in the percentage of strains with MIC of ≤ median 2003 MIC for ceftazidime was insignifi cant.
Variation in the MIC of ceftazidime, ciprofl oxacin, and amikacin for P. aeruginosa strains during the 5-year study (MIC creep) is shown in Figs.1-3.

Discussion
Increasing resistance rates of P. aeruginosa strains, particularly hospital strains, to different antipseudomonal agents have been reported worldwide, and it presents a serious therapeutic problem in the management of diseases due to these organisms (19,20).Susceptibility of P. aeruginosa isolates from the HLUHS to various antibiotics was examined in our study.There is a clear trend toward an increase in resistance in all classes of antibiotics as the resistance level in general, which is signifi cantly expressed in some classes of antipseudomonal drugs.
In our study, the resistance to piperacillin decreased insignifi cantly from 23.3% in 2003 to 17.8% (-5.5%) in 2008.The same decline of 4.4% in resistance to piperacillin/tazobactam was documented.In our opinion, decreased resistance to piperacillin and piperacillin/tazobactam can be associated with the restricted usage of this β-lactam antibiotic in our hospital recently.The similar resistance to piperacillin and piperacillin/tazobactam was found in Belgium (24.0% and 17.5%) and Greece (20.2% and 18.6%) (1,19).Much higher resistance to piperacillin was observed in Russia (79.0%) (21).The lower resistance to piperacillin was in Italy (12.0%) and Spain (10.0%) (22,23).The lower resistance to piperacillin/tazobactam was found in the United Kingdom and Ireland and in the United States (5.2% and 8.4%, respectively) (10,24).The United States reports low resistance rates to piperacillin/ tazobactam, and the study by Mutnick et al. has revealed further decreasing resistance to piperacillin/ tazobactam (-2.5%) (24).
In our study, the resistance to ceftazidime increased signifi cantly from 5.6% in 2003 to 13.9% in 2008 (+8.3%) (P=0.05).The documented resistance rate to ceftazidime (13.9%) is close to that observed in the study by Bonfi glio et al. from Italy (13.4%) and Bouza et al. from Spain (15.0%) (22,23).Data of the studies from Belgium and Greece showed the resistance to ceftazidime to be twofold higher than in our study: 28.5% and 25.5%, respectively (1,19).In some other countries such as the United Kingdom and Ireland, France, and the United States, the resistance to ceftazidime was lower than in Lithuania (3.8%, 6.2%, and 9.7%, respectively), and a 1.2% decrease was documented in the United States from 1999 to 2002 (10,24,25).An increase in the ceftazidime MIC for P. aeruginosa strains over time is a phenomenon, known as ceftazidime creep.The prevalence of P. aeruginosa strains with MIC values of 1-2 μg/mL changed during 6 years, and the level of resistance to ceftazidime increased as strains with raised MIC became dominant (2-6 μg/mL).Higher MIC is associated with higher mortality rates, and it was reported that even small increases in MIC below the susceptibility breakpoint could affect the clinical effi cacy of antibiotics (26).
In our study, the resistance rate to ciprofl oxacin increased signifi cantly from 15.6% in 2003 to 39.6% in 2008 (+24%, P<0.001).An MIC creep during the 6-year study was observed as the predominance of P. aeruginosa strains with the MIC values of 0.125-0.25 μg/mL in 2003 changed to the predominance of P. aeruginosa strains with the MIC values of 1-3 μg/mL and 15-165 μg/mL in 2008.High resistance to ciprofl oxacin has been reported worldwide: in Greece, 18.6%; in United States, 22.7% (increase, +10.8%); in Spain, 23.0%; in Belgium, 24.0%; and in Italy, 31.9% (1,19,(22)(23)(24).Only in the United Kingdom and Ireland, the resistance to ciprofl oxacin was estimated to be constant (7.4%) (10).Fluoroquinolones, such as ciprofl oxacin, are broad-spectrum antimicrobials widely used to treat different infections (27).The broad use of them has led to increased resistance rates of P. aeruginosa strains to fl uoroquinolones as well as increased multidrug resistance.This causes a serious problem in clinical practice (28).
The resistance to gentamicin in our study did not change: 20.0% in 2003 and 19.8% in 2008.In Spain and the Russian Federation, the resistance rates to gentamicin were higher (31.0% and 75.0%, respectively) (21,23), but in the United States and the United Kingdom and Ireland -lower (8.4% and 6.3%, respectively) (10,24).
The resistance rate of P. aeruginosa strains to amikacin was 5.9% in our study.In 2003, the MIC values of P. aeruginosa strains for amikacin were 2-6 μg/mL, and during the study period, it increased as strains with raised MIC (4-12 μg/mL) became dominant.This MIC creep confi rms the level of resistance to amikacin to be increased.In Belgium, Italy, and Spain, the resistance rates to amikacin were higher reaching 9% to 10% (1,22,23), while the highest resistance rate to amikacin was reported in Greece (27.5%) (19).
Belgium scientists reported that despite the resistance of P. aeruginosa to penicillins, cephalosporins, fl uoroquinolones, and aminoglycosides varies among Belgian hospitals, the level of resistance is increasing in general (1).The surveys carried out in Italy and Spain documented the resistance rates of P. aeruginosa similar to our study, and an increase in resistance to vast majority of antibiotic classes is being observed as well (22,23).
The percentage of P. aeruginosa strains sensitive to all tested antibiotics signifi cantly decreased from 66.7% (60 of the 90) in 2003 to 47.5% (48 of the 101) in 2008 (P<0.05).
In our study, MDR rate increased from 7.1% in 2003 to 12.8% in 2008 (+5.7%).The same trend was found in the US study (29).Several studies have showed that previous antibiotic use increases the risk of a P. aeruginosa infection resistant to many commonly used antimicrobials especially fl uoroquinolones (15,30).
A signifi cant increase in MICs for ciprofl oxacin and amikacin was documented as the percentage of isolates with MIC lower than median MIC markedly decreased in 2008 as compared with 2003, and it was not a case for ceftazidime.
Antimicrobial drug use is one of the well-established risk factors for the development of antimicrobial resistance (24).Substantial regional variations in the resistance patterns worldwide have been observed, and it is probably related to the different antibiotic treatment regimens used traditionally in different countries (31,32).
Surveillance of the resistance level of P. aeruginosa strains, so called "diffi cult to treat pathogen," gives clinicians a better understanding of epidemiologic status in a particular setting and allows making an adequate empirical choice of antibiotics in the treatment of life-threatening pseudomonal infections especially in the ICU.

Table 1 .
in 2003 to 12.9% (13/101) in 2008 (P=0.25).Table 3 presents changes in the prevalence of nosocomial P. aeruginosa isolates, sensitive to different antibiotics, with MIC of ≤ median MIC of year 2003 comparing years 2003 and 2008.In 2008, Changes in the Percentages of Pseudomonas aeruginosa Nosocomial Isolates Resistant to Tested Antibiotics in the Intensive Care Units of HLUHS in 2003 and 2008

Table 2 .
Changes in the Contribution of Piperacillin-, Ceftazidime-, Ciprofl oxacin-and Gentamicin-Resistant P. aeruginosa Strains to the Development of Multidrug Resistant Phenotypes

Table 3 .
Changes in the Percentages of Pseudomonas aeruginosa Nosocomial Isolates Sensitive to Different Antipseudomonal Antibiotics With the MIC Values of ≤ Median MIC in 2003