Antimicrobial Resistance Patterns Among
|Key words:||Antimicrobial resistance, gram-negative bacilli, intensive care units, nosocomial infections|
Nosocomial infections are important problems in hospitals. In spite of the advances in infection control measures the mortality, morbidity and cost of such infections remain extremely high. In a recent European Prevalence Study it was shown that ICU-acquired infections developed in 20.6% of all patients admitted . The frequency of such infections varies significantly not only between countries, but also between different hospitals. There is an obvious necessity to know not only global trends in epidemiology of hospital infections but also the local situation.
Monitoring of antimicrobial resistance is most important in ICUs, where infection rates and consumption of antimicrobials are significantly higher in comparison with other wards. This practice also helps in optimisation of empirical antimicrobial therapy, decreasing prescription of non-rational regimens of therapy, and increasing cost-effectiveness of treatment.
The aim of this study was to determine the profile of antimicrobial resistance among aerobic gram-negative bacilli isolated from patients with ICU-acquired infections in different parts of Russia.
Between September 1995 and May 1996, 10 centres in different parts of Russia (Figure 1) which included the North-West European Region (Moscow, Saint-Petersburg, Smolensk, Nizniy Novgorod), the South European Region (Krasnodar), the Central European Region (Kazan), Ural (Ekaterinburg), and Siberia (Novosibirsk, Krasnoyarsk) were asked to submit a minimum of 100 consecutive aerobic gram-negative bacilli isolated from patients with ICU-acquired infections. The cultures were performed on the basis of clinical indications. Only non-duplicate isolates were included in this study. All microorganisms were identified to species level.
Following isolation, identification and susceptibility testing, isolates were sent to the central laboratory in Smolensk on 0.7% nutrient agar slants.
At least 20% of strains were re-identified by API 20E and API 20NE (BioMerieux, La Balme les Grottes, France). When more than 10% of discrepancies were observed, re-identification was performed for all submitted isolates.
All isolates were stored at -70oC as a heavy suspension in trypticase soy broth with added glycerol (10% v/v).
Minimal inhibitory concentrations (MICs) of 12 antimicrobials were determined by Etest® (AB BIODISK, Solna, Sweden). Interpretation of the results was performed according to National Committee for Clinical Laboratory Standards guidelines (1995). Validation of the procedures was by the determination of the MICs for Pseudomonas aeruginosa ATCC 27853 and Escherichia coli ATCC 35218. Test results were accepted only if the control strains' MICs were within performance range. As an additional quality control procedure, susceptibility testing of 20% of all strains was performed in the Research and Development Laboratory of AB BIODISK.
The following antimicrobials were tested: piperacillin, amoxicillin/clavulanate, piperacillin/tazobactam, cefuroxime, ceftriaxone, cefotaxime, ceftazidime, imipenem, gentamicin, amikacin, ciprofloxacin and trimethoprim/sulphamethoxazole.
Data management and statistical analysis was performed with SAS 6.11 (SAS Institute, USA).
During the study period, 1,005 non-duplicate consecutive strains from 863 patients were received and evaluated. The most common isolates, in order of frequency, were Pseudomonas aeruginosa (28.8%), Escherichia coli (21.4%), Klebsiella pneumoniae (16.7%), Proteus mirabilis (9.7%), Enterobacter spp. (8.2%) and Acinetobacter spp. (7.7%), representing 92.5% of all isolates. The remaining isolates included Citrobacter spp. (1.6%), Morganella morganii (1.0%), Proteus vulgaris (1.0%), Serratia marcescens (0.9%), Stenotrophomonas maltophilia (0.9%), and other gram-negative bacilli (2.1%).
The most common isolates submitted by centres are shown in the Table 1. E.coli was the most prevalent isolate in Saint-Petersburg and Krasnoyarsk, Acinetobacter spp. in Ekaterinburg, Enterobacter spp. in Moscow-2.
The body sites yielding specimens are presented in Figure 2. Among the wound isolates the most common microorganisms were P.aeruginosa (38.2%), E.coli (19.4%), P.mirabilis (14.2%), from the respiratory tract P.aeruginosa (37.4%), K.pneumoniae (23.1%), E.coli (13.2%), from the urinary tract E.coli (33.1%), P.aeruginosa (18.0%), K.pneumoniae (14.4%) and from the gastrointestinal tract E.coli (32.1%), K.pneumoniae (19.5%), E.cloacae (18.9%). K.pneumoniae (45.7%), A.baumanii (17.4%) and P.aeruginosa (15.2%) were the common isolates from blood.
Administration of Antimicrobials
A total of 854 patients (99.0%) were receiving antimicrobials on the day samples were taken. Aminoglycosides were the most frequently administered (35.3% of total prescriptions), and gentamicin was the most commonly administered antibiotic (27.3% of total prescriptions), followed by third-generation cephalosporins (13.2%) and ampicillin (11.5%). Overall, 312 patients (36.2%) were given more than one antimicrobial agent.
Comparison of activity
There were no major discrepancies between results obtained in Russia and those in the Research and Development laboratory of AB Biodisk in Sweden.
Summaries of MIC50, MIC90, MIC ranges and levels of resistance for the most common isolates are presented in Tables 2a-2c.
The most active b-lactam antibiotic against P.aeruginosa was imipenem (7% of resistance), followed closely by ceftazidime (11%). Imipenem also was the most active b-lactam against E.coli, followed by third-generation cephalosporins (ceftazidime, ceftriaxone and cefotaxime). Against K.pneumoniae, imipenem showed the highest activity among the b-lactams tested, followed by third-generation cephalosporins. Imipenem possessed the highest activity against P.mirabilis, Enterobacter spp. and Acinetobacter spp. We observed quite low activity of piperacillin and piperacillin/tazobactam against both enterobacteria and non-fermenters.
Amikacin was significantly more active than gentamicin against all tested microorganisms. The resistance rates to amikacin varied from 0% in K.pneumoniae to 7% in P.aeruginosa and Acinetobacter spp. compared with the resistance rates to gentamicin of 58% in K.pneumoniae, 75% in P.aeruginosa and 91% in Acinetobacter spp.
Ciprofloxacin was selected as the most active drug among the currently available fluoroquinolones. It showed good activity against enterobacteria (1-5% resistant), but was significantly less active against non-fermenters (15% resistance in P.aeruginosa and 53% resistance in Acinetobacter spp.).
P.aeruginosa (Fig. 3)
The resistance of P.aeruginosa to piperacillin ranged from 10% in Saint-Petersburg to 79% in Smolensk. The similar tendency have been observed on activity of other antimicrobials. The resistance to ceftazidime was the lowest in Ekaterinburg and Krasnodar (5%) and the highest in Nizniy Novgorod (26%). Resistance to imipenem was comparatively low in all centres (0-11%), excluding Kazan (29%). Gentamicin possessed relatively low activity against P.aeruginosa in all centres (from 40% resistance in Krasnoyarsk to 90% in Krasnodar). Amikacin showed good activity in all centres with a resistance range of 0-11%, excluding Nizniy Novgorod (18%) and Kazan (19%). Resistance to ciprofloxacin varied between 6% and 42%.
As indicated on Figure 4, piperacillin showed quite poor activity against E.coli isolated in different centres, with the resistance ranging from 9% in Moscow-2 to 75% in Smolensk. Piperacillin/tazobactam was more active not only on comparison with piperacillin, but also with co-amoxiclav.
Generally, piperacillin, piperacillin/tazobactam and co-amoxiclav possessed the lowest activity against K.pneumoniae isolates with the ranges of resistance 58-100%, 17-89% and 25-94%, respectively (Figure 5). Another problem is a high level of resistance to third-generation cephalosporins (more than 30%) in Kazan, Novosibirsk, Smolensk and Krasnoyarsk.
The resistance of P.mirabilis in different centres is presented in Figure 6. Piperacillin showed relatively poor activity in all centres with the resistance ranging from 47% in Krasnodar to 91% in Smolensk. Piperacillin/tazobactam and co-amoxiclav had similar activity with a resistance range of 0-64%.
To our knowledge this study was the first multicentre study of ICU-acquired infections carried out in Russia. The design of this study included the simultaneous collection of clinical material in all participating centres, which excluded the influence of time factors on patterns and frequency of hospital infections. Also of the importance was the use of an internationally approved susceptibility testing method. We would also like to emphasise that selection of tested antimicrobials was based on the drugs available in Russian hospitals. We feel that these data will be of real value for physicians in different hospitals especially in the choice of empirical therapy.
We compared our results with studies in the Western Europe, analysing data from studies from Germany , Belgium , the Netherlands  and Sweden  which had a similar design. All of these studies included several centres (16 in Belgium, 10 in Germany and Sweden and 8 in the Netherlands) and were focused on ICUs. Moreover, the centres were asked to submit susceptibility testing data on at least 100 consecutive gram-negative isolates. We found the relatively high level of resistance among all P.aeruginosa isolates in Russian ICUs to piperacillin (50% vs. 8.5%, 12%, 15%, 18% in Sweden, Germany, the Netherlands and Belgium, respectively) and gentamicin (75% vs. 6.8%, 40%, 41%, 48%, respectively). Resistance rates to gentamicin in E.coli in this study was also significantly higher (13% vs. 2% in Germany and 4% in both Belgium and the Netherlands). The differences in the resistance rates for the other antimicrobials were not so marked. Similar tendencies of resistance were observed in K.pneumoniae: the resistance to gentamicin was 58% in comparison with 2% in Germany and 10% in the Netherlands. Such high levels of resistance to gentamicin might be explained by its high consumption both in out-patients clinics and hospitals, even for therapy of community-acquired pneumonia.
In our analysis ceftazidime was selected as the representative third-generation cephalosporin for detection of resistance in strains that are hyperproducers of type I chromosomal b-lactamases or that carry plasmid-mediated extended spectrum b-lactamases . The comparison of our data with recently published NPSP  and NNIS ICU  data shows that the rate of resistance among our K.pneumoniae strains was significantly higher (26.2% vs. 5.8% and 2.6%, respectively). The resistance rate among Enterobacter species was also higher (56.1% vs. 32.0% and 39.6%, respectively), but the resistance among our P.aeruginosa isolates was slightly lower (10.7% vs. 14.2% and 13.2%, respectively).
We would like to draw attention to the high level of resistance of nosocomial strains of Acinetobacter spp. One of the most troubling issues is the increase of resistance to ciprofloxacin (53%) which has been widely used for the treatment of gram-negative infections in Russian hospitals during the last couple years.
Another interesting fact that the resistance to many antimicrobials in Enterobacter spp. coexists with quite good activity of trimethoprim/sulphamethoxazole, which potentially can be used for the treatment of these infections.
Comparatively high level of resistance to combinations of b-lactam and b-lactamase inhibitors among all species in different hospitals considerably restrict their usage in the treatment of nosocomial infections.
In conclusion we would like to mention the important point. During this study we created a team which will work on the control of hospital infections and, will, we hope, develop guidelines and protocols for the treatment of infections in hospitals in Russia.
We would like to thank Merck Sharp & Dohme (Russia) for financial support.
Also we are grateful to Dr. Yu. Danilevsky personally for his continuous enthusiasm during the study.
We would like to offer our gratitude to personnel of Research & Development Laboratory of AB BIODISK (Sweden) and to Dr. A. Bolmström and Dr. Å. Karlsson personally for their help in this study.
Table 1. The most common bacterial isolates received from the centres
Table 2a. MICs50 (mg/l), MICs90 (mg/l), MIC ranges (mg/l) and resistance (%) to tested antimicrobials for the most common bacterial species.
* R resistance, according to NCCLS criteria
Table 2b. MICs50 (mg/l), MICs90 (mg/l), MIC ranges (mg/l) and resistance (%) to tested antimicrobials for the most common bacterial species.
* R resistance, according to NCCLS criteria
Table 2c. MICs50 (mg/l), MICs90 (mg/l), MIC ranges (mg/l) and resistance (%) to tested antimicrobials for the most common bacterial species.
* R resistance, according to NCCLS criteria
Figure 1. Geographic distribution of the centers included in this study.
Figure 2. Distribution of body site samples.
Figure 3. Resistance (%) of Pseudomonas aeruginoza in the study centers.
Figure 4. Resistance (%) of Escherichia coli in the study centers.
Figure 5. Resistance (%) of K.pneumoniae in the study centers.
Figure 6. Resistance (%) of Proteus mirabilis in the study centers.