Antimicrobial susceptibility pattern of Gram-negative bacilli isolated from a Teaching Hospital in Jeddah , Saudi Arabia

Gram-negative bacilli (GNB) are commonly implicated in clinical diseases. However, with their increasing resistance to antimicrobial agents, treatment becomes a challenge. This study has been conducted at King Abdulaziz University Hospital in Jeddah, Saudi Arabia over a period of one month from July to August 2011. Identification and antibiotic sensitivity tests of GNB were performed using standard microbiological methods and Vitek2 system. Extended spectrum beta lactamase (ESBLs) strains were detected using double disc synergy test and Vitek 2 system. A total of 176 Gram-negative bacilli were studied. The most frequently isolated organism was E. coli (38.07%) followed by Klebsiella pneumoniae (15.91%), Pseudomonas aeruginosa (11.93%), Proteus mirabilis (9.66%) and Acinetobacter baumannii (6.82%). Other Gram-negative bacilli were less frequent. Isolates were detected most frequently from ICU patients (26.70%). Urinary tract, wound and respiratory tract infections were implicated most often. Extended spectrum beta lactamase strains accounted for 20% of all Enterobacteriacae. The vast majority of the GNB isolates were resistant to many antibiotics. Carbapenems, tigecycline and amikacin were effective against most multi-drug resistant Enterobacteriacae. Pseudomonas aeruginosa was resistant to several antibiotics; most effective agents were ceftazidime (80.95%), aztreonam (76.19%), carbapenem (90.48%), amikacin (90.48%), fluoroquinolones (80.95%) and piperacillin (61.90%). Isolates of Acinetobacter baumannii and Stenotrophomonas maltophilia were mainly multi-drug resistant to most tested antibiotics. In view of high levels of antibiotic resistance encountered, continuous surveillance of antimicrobial susceptibility patterns is warranted.


INTRODUCTION
Gram-negative bacilli (GNB) are a large heterogenous group and amongst them the Enterobacteriacae are one of the most common bacteria isolated in clinical laboratories.Enterobacteria are ubiquitous in the environment and carried in the intestinal tract of human and many animals.Generally, they are not associated with disease.However, when clinical infections occur, Escherichia coli (E.coli) is most often implicated.Clinical infections frequently involve the urinary and intestinal tract although other sites may be involved.Other members of the group such as Serratia marcescens and Enterobacter aerogenes (Ent.aerogenes) are opportunistic.Among the other GNB, Ps. aeruginosa is renowned for hospital acquired infections (Peleg and Hooper, 2010).
Antibiotics are invariably used for treatment of these infections.However, the tremendous therapeutic advantage offered by these agents is being negated by the emergence of increasingly resistant strains of bacteria on a global scale (Yezli et al., 2014).Also, in Saudi Arabia (SA), prevalence of antimicrobial resistance amongst a variety of clinical pathogens has been documented (Memish et al., 2012;Yezli et al., 2014).High levels of antibiotic resistance among GNB specially Ps. aeruginosa and A. baumannii has been reported particularly from intensive care units (ICUs) originating from a variety of infections and resulting in significant morbidity and mortality (Al-Ahmadey and Mohamed, 2013).
In recent years, the problem of drug resistance has been enhanced by the emergence of organisms producing extended spectrum beta-lactamases (ESBLs).These strains exhibit resistance to various antimicrobial agents including third generation cephalosporins, extended spectrum penicillins, and monobactams.Susceptibility to other agents such as fluoroquinilones and aminoglycosides is variable.The carbapenems are often the only active class of antibiotics against these strains (Kader and Kumar, 2004).Prevalence of these ESBLs varies globally and in SA; differing rates have been reported.Extended spectrum beta-lactamase strains have been implicated in both hospital and community settings and their increasing levels have been a cause for concern (Khanfar et al., 2009;Yezli et al., 2014).Hence, regular surveillance of clinical isolates and their susceptibility patterns is warranted.
The aim of this prospective study was to determine the antimicrobial susceptibility patterns of GNB and ESBL producing strains in clinical samples.

Setting and design
This prospective study was undertaken in a tertiary care hospital in Jeddah; King Abdulaziz University Hospital (KAUH) with 845 bed capacity, during a period of one month from 15 th of July to 15 th of August 2011.
A total of 176 Gram-negative bacilli isolates were identified from 158 patients during the study period.Clinical Gram-negative bacilli strains were obtained from various patient specimens submitted to Clinical Microbiology Laboratory at KAUH.

Data collection
Demographic data (Age, gender, nationality, specimen type and ward of hospital) of patients with Gram-negative bacterial infections (UTIs, skin infections, sepsis, pneumonia, etc.) were recorded on a standardized form.

Bacterial isolates
Strains were isolated from a variety of specimens cultured on MacConkey agar and 5% Sheep Blood agar (Saudi Prepared Media Laboratory, SPML).All isolates used in the study were maintained on nutrient agar slants at 4ºC.

Identification of Gram-negative bacilli
Preliminary identification of GNB was performed using conventional methods including: Gram-staining, culture characteristics, lactose fermentation, and oxidase test.Further identification to species level was performed using Vitek 2 (ID-GN card) automated system (BioMerieux, France) according to manufacturer's instructions.

Disc diffusion method
The susceptibility of the tested isolates was carried out by Kirby-Bauer disc diffusion method according to the Clinical and Laboratory Standard Institutes (CLSI) guidelines (CLSI, 2012).

Minimum inhibitory concentrations (MICs)
Minimum inhibitory concentrations of various antibiotics to different isolates were determined using automated microbiology system Vitek 2. Gram-negative cards (AST-N114/AST-GN26/AST-EXN8) were used according to the instructions of the manufacture.

Quality control strains
E. coli ATCC 25922, K. pneumoniae ATCC 700603 and Ps.aeruginosa ATCC 27853 were used as controls for the antimicrobial susceptibility tests.
Author(s) agree that this article remains permanently open access under the terms of the Creative Commons Attribution License 4.0 International License generation cephalosporins were tested for ESBL production by double disc synergy test (DDST).The DDST was performed as a standard disc diffusion method using Mueller-Hinton agar plates according to CLSI guidelines (CLSI, 2012).Discs containing 30 µg aztreonam and 30 µg of cefotaxime, ceftazidime, cefpodoxime, ceftriaxone, cefuroxime and cefepime each were placed 15-20 mm apart (center to center) around a disc containing amoxicillin+clavulanic acid (augmentin 20 µg+10 µg respectively); clavulanic acid is an inhibitor of beta-lactamase.The Mueller-Hinton agar plates were incubated at 37°C for 24 h.Enhancement of inhibition zone of any one of the test antibiotics towards augmentin disc was regarded as presumptive ESBL production, these isolates were subjected to Vitek 2 system (AST-EXN8) for confirmatory test.

Statistical analysis
The results were statistically analyzed using Microsoft Excel 2010.Statistical tests were presented as frequencies and percentages.

The distribution of non-ESBLs GNB and ESBLs
Enterobacteriaceae with respect to clinical specimen type is shown in Tables 1 and 2.
The test was considered positive upon the formation of a clear extension of the edges of the zones of inhibition of any of the cephalosporin antibiotics towards the disc containing clavulanic acid, often showing a characteristic  shape-zone referred to as 'keyhole' (Figures 2  and 3) due to inhibition of β-lactamase by clavulanic acid (Neu and Fu, 1978).
The antibiotic resistance patterns of non-ESBLs Enterobacteriaceae are shown in Table 5.A total of 81 isolates (72.32%) were resistant to ampicillin.
The single Stenotrophomonas maltophilia isolate exhibited MDR and was susceptible only to trimethoprime/sulfamethoxazol, fluoroquinolones, and colistin.

DISCUSSION
In recent years, incidence of multidrug resistant pathogenic organisms has increased notably.Thus, therapeutic proposals have been modified according to the emergence of drug resistance and adapted to epidemiological markers of individual infectious processes, geographical variations of these markers and the availability of new antibacterial agents (Alvarez-Lerma et al., 2006).
In the present study, 176 Gram-negative bacilli isolates were obtained from 158 patients, and the ratio of males (41.14%) to females (58.86%) was similar to Al-Ghamdi et al. (2002).Other investigators reported a higher proportion of males (El-Amin and Faidah, 2012).
In our experience most patients were above 50 years (48.73%) which is comparable to previous reports from the same hospital (Madani, 2002) and other hospitals in Saudi Arabia (Asghar and Momenah, 2006;Khanfar et al., 2009).Results from this study corroborate the generally accepted rule that older patients are predisposed to infection.
The distribution of Saudi (32.28%) and non-Saudi (67.72%) observed here is similar to a previous report for the same hospital by Madani (2002).Studies in other Saudi hospitals however have indicated a higher proportion of Saudi patients (Asghar, 2006;Al-Anazi, 2009).Demographic data collected indicated that there was no selection in the terms of gender, nationality or age.
In this investigation, Ps. aeruginosa (52.38%) were cultured mostly from respiratory secretions other investigations have documented similar finding (Asghar    were also associated most frequently with ICUs patients.Similar results were reported previously in Saudi Arabia (Asghar and Faidah, 2009;Al-Ahmady and Mohamed, 2013).
Other work has previously shown that ICU patients are implicated most frequently in infections with other microorganisms also (Helmi et al., 2013).The incidence of multidrug resistant GNB among ICU patients may be due to prior antibiotic use, long antibiotics exposure, inadequate antibiotic therapy, or other reasons.
In our experience, ESBLs were rapidly identified by using DDST and Vitek 2 system.In addition to these methods, other studies have used E-test, combined discs or PCR methods for detecting ESBLs in GNB (Khanfar et  , 2009;Al-Sweih et al., 2011;Grover et al., 2013).Single disc diffusion method was not suitable for detecting ESBLs (Babay, 2002).However, further confirmatory tests are needed to confirm the presence of ESBL enzymes in such isolates.
In general, the results demonstrated that GNB have high rates of resistance to commonly used antibiotics.Attention should be paid to the use of ampicillin which shows high levels of resistance (> 50%).
Tazobactam in combination with piperacillin has excellent clinical efficacy in various infections and is a promising beta-lactamase inhibitor which has its own antibiotic activity (Niki, 2001).Resistance rate to piperacillin-tazobactam (16.07%) ciprofloxacin (13.39%)Beta-lactam drugs are commonly included in the empirical antibiotics treatment of Gram-negative bacteria; however ESBLs producing bacteria may not be susceptible to such treatment.ESBLs prevalence shows wide variation from country to country and within the same country over a period of time but is generally on the increase (Yezli et al., 2014).This study shows that 20% (28/140) of Enterobacteriaceae were ESBLs producing strains including 20 isolates of E. coli and 8 isolates of K. pneumoniae.This is higher than earlier reports in Saudi Arabia by Kader and Kumar (4.8%) (Al-Zarouni et al., 2008;Mokaddas et al., 2008).
ESBLs producing strains were resistant to most commonly used antibiotic such as ampicillin (100%) and ciprofloxacin (78.57%) which is higher than reported in Bangladesh by Mowla et al. (2011).Khanfar et al. (2009) reported similarly high rates of resistance to ciprofloxacin (80%) in ESBL producers isolated from the eastern province of Saudi Arabia, but other studies found much lower rates in ESBL K. pneumoniae from Al-Qassim (9.1%) and Riyadh (11%), both in Saudi Arabia (Tawfik et al., 2011;Al-Agamy et al., 2009).
In this study, the carbapenems, tigecycline and amikacin were the most active agents against MDR Enterobacteriaceae such as E. coli, K. pneumoniae, Ent.aerogenes, Ent.cloacae, S. marcescens, and highly active against the ESBLs producing isolates.Similar results were reported in other studies (Kader and Kumar, 2004;Khanfar et al., 2009;Iqbal et al., 2014).
It has also been reported that the susceptibility of Enterobacteriaceae to the carbapenems was unaffected by the production of ESBLs (Turner, 2005).This could explain the high sensitivity of Enterobacteriaceae to carbapenems in this study.
However, as reduced carbapenems susceptibility in ESBLs producing strains has been documented in Saudi Arabia (Balkhy et al., 2012) and in other countries (Jamal et al., 2011;El-Herte et al., 2012), hospitals should have proper policies and guidelines for the prudent use of antimicrobial agents and adequate infection control in order to avoid the emergence of further resistant isolates.
Our findings of high levels of resistance to non-beta lactam classes of antibiotics in ESBLs producing organisms are in agreement with other reports.Often attributed to self-transmissible R-plasmids are implicated (Shibl et al., 2012).
On the other hand, Ps. aeruginosa demonstrated high resistance rate to several antibiotics.During this survey, imipenem and amikacin show a higher susceptibility (90.48%) against Ps.aeruginosa followed by fluoroquinolones and cefepime (80.95% for each) then piperacillin-tazobactam (61.90%).This is similar to a previous study done in the same hospital (Eltahawy and Khalaf, 2001) as well as other hospitals in Saudi Arabia (Asghar, 2006;Memish et al., 2012;Al-Ahmady and Mohamed, 2013) and other countries (Al-Yaqoubi and Elhag, 2008;Benachinmardi et al., 2014).Another study, showed a lower susceptibility of Ps. aeruginosa to imipenem and amikacin compared to our observation (Jamshidi et al., 2009) Our work shows that A. baumannii was MDR to a variety of antibiotics, including ampicillins, cephalosporins, carbapenem, fluoroquinolones and gentamicin confirming earlier reports in KSA (Memish et al., 2012;Al-Ahmadey and Mohamed, 2013;Al Masoudi et al., 2013).All Acinetobacter spp.were sensitive to colistin (100%) similar finding has been documented by Asaad et al. (2013).
This study shows that the single Stenotrophomonas maltophilia, was resistant to all antibiotics tested except trimethoprime/sulfamethoxazol, fluoroquinolones and colistin.Other studies showed similar results (Nicodemo and Garcia Paez, 2007;Asaad et al., 2013).

Conclusion
Emerging antimicrobial resistance is currently the main concern of the medical community, because such resistant bacteria are becoming more difficult to treat.Our data showed a high rate of resistance among GNB but carbapenems are still effective against these organisms.However, MDR is common with Ps. aeruginosa and A. baumannii.Strict infection control measures, formulation of antibiotic policy and continual monitoring and surveillance is required globally.New antimicrobial agents and vaccines must be developed.Further studies are recommended to evaluate both antimicrobial susceptibility of GNB isolated and their mechanism of resistance in different hospitals of Saudi Arabia and other countries.

Figure 1 .
Figure 1.Distribution of most frequently isolated Gram-negative bacilli.

Figure 2 .
Figure 2. ESBLs producing E. coli, double disc synergy test.Showing enhanced zone of inhibition around more than one of the β-lactam-containing discs towards the clavulanic acidcontaining disc.

Table 1 .
Distribution of non-ESBLs Gram-negative bacilli with respect to clinical specimen.

Table 2 .
Distribution of ESBLs Enterobacteriaceae with respect to clinical specimen.

Table 3 .
Distribution of non-ESBLs Gram-negative bacilli with regard to wards.

Table 4 .
Distribution of ESBLs Enterobacteriaceae with regard towards.
*% according to total number of specified isolates; **% according to hospital wards, ***other wards such as private and isolation; ND, not detected.ESBLs: Extended spectrum-beta-lactamase.

Table 5 .
Susceptibility of 112 isolates of non-ESBLs Enterobacteriaceae to antimicrobial agents.

Table 6 .
Susceptibility of 28 isolates of ESBLs Enterobacteriaceae to antimicrobial agents.

Table 7 .
Susceptibility of 21 isolates of Pseudomonas aeruginosa to antimicrobial agents.

Table 8 .
Susceptibility of 12 isolates of Acinetobacter baumannii to antimicrobial agents.
in 2004, and Khanfar et al. (6%) in 2009 but lower than those reported by Babay (36%) in 2002.Other countries have also shown variation, within the Arabian Gulf region, high ESBLs prevalence of 31.7% in Kuwait and 41% in the United Arab Emirates have been reported