Extended spectrum beta-lactamases ( ESBL ) in Escherichia coli and Klebsiella pneumoniae : trends in the hospital and community settings

Background: To assess the prevalence of extended spectrum beta-lactamase (ESBL) producing Escherichia coli and Klebsiella strains in nosocomial and community-acquired infections. Methodology: The study was conducted at a centralized microbiology laboratory in the Eastern Province of Saudi Arabia. Laboratory records (January 2004December 2005) were assessed. Associated resistance to a panel of antibiotics was determined. Results: A total of 6,750 Gram-negative organisms were assessed for ESBL-phenotype. ESBL was detected in 6% (409/6,750) of isolates, the majority of which were E. coli (83%). ESBL producers were significantly higher among isolates from in-patients 15.4% (143/927) versus out-patients (4.5%; 266/5,823); p < 0.05. Old age (older than 60 years) represented a significant risk for having an ESBL-producing pathogen. Urine was the major source of ESBL isolates in in-patients (46.1%) and out-patients (74.4%). The proportion of urinary E. coli isolates which were ESBL producers was significantly higher among in-patients (53/506; 10.4%) compared to out-patients (182/4,074; 4.4%); p < 0.05. Among in-patients, 60% of the ESBL associated infections were nosocomial. All were sensitive to imipenem but high levels of resistance to gentamicin, amikacin, amoxicillin-clavulanic acid and ciprofloxacin was shown. Conclusion: The findings document evidence of the spread of multiresistant ESBL-producers into the community. This has significant implications for patient management, and indicates the need for increased surveillance and molecular characterization of these isolates.


Introduction
In 1983, the first outbreak involving extended spectrum beta-lactamase (ESBL) producing organisms was reported in Germany [1].These ESBL-producing pathogens are now recognized globally as major causes of nosocomial and community-acquired infections [2].ESBLs arise because of mutations in the TEM-1, TEM-2, or SHV-1 genes, commonly found in the Enterobacteriaceae family [2,3].While these enzymes are found predominantly in Klebsiella species and Escherichia coli, they have also been described in other genera of Enterobacteriaceae including Citrobacter, Serratia, Proteus, Salmonella, and Enterobacter [2].They are capable of hydrolyzing penicillins, broad-spectrum cephalosporins and monobactams, but they do not affect the cephamycins or carbapenems and their activity is inhibited by clavulanic acid [3].
The prevalence of bacterial isolates expressing the ESBL phenotype varies across different geographical regions with low rates of 3-8% reported in Sweden, Japan and Singapore compared to much higher prevalence rates documented in studies from Portugal (34%), Italy (37%), New York (44%), Latin American countries (30-60%) and Turkey (58%) [3].In the Arabian Peninsula, reported ESBL detection rates range from 8.5-38.5% in data from the Kingdom of Saudi Arabia [4][5][6][7][8] and (31.7%) in Kuwait [9]; the highest level of 41% is from the United Arab Emirates [10].This variation is perhaps reflective of the fact that these studies focused on hospital-acquired infections or particular sites of infection (e.g.urine, blood).In this report, we present the first direct comparative data on the prevalence and characteristics of ESBL-producing Enterobacteriaceae causing hospital and communityacquired infections in Saudi Arabia.

Data collection
Retrospective analysis of laboratory records at the Saudi Aramco Dhahran Health Center, over a two-year period from January 2004 to December 2005, was conducted.The Centralized Microbiology Laboratory at this 405-bed facility also serves the 80bed Al-Hasa hospital, as well as three out-patient clinics and an emergency medical service in the surrounding communities.Patients' demographic data, clinical diagnoses, and specimen types were recorded for all E. coli and Klebsiella strains isolated during the study period.The number of isolates found to be ESBL producers and their antibiotic susceptibility profiles were noted.Only one positive culture per patient was included in the study and repeated positive cultures from the same patient were excluded from the analysis.Isolates for which it was impossible to discriminate between contamination and infection were excluded from the analysis.Hospital associated infection was defined as occurrence of infection 48 hours or more after hospital admission, without evidence that the infection was present or incubating on admission, in patients without prior history of stay in a healthcare facility [11].

Detection of ESBL isolates
E. coli and Klebsiella isolates reported as cephalosporin resistant or ESBL using the automated VITEK Gram Negative Susceptibility System with cards GNS 206 and 121 (bioMérieux, Vitek Inc, Hazelwood, USA) were screened for ESBL detection using the disk diffusion method.In keeping with the Clinical and Laboratory Standards Institute (CLSI) recommended guidelines [12], ESBL screening was performed by disk diffusion using ceftazidime (10ug) cefpodoxime (30ug) and aztreonam (30ug) disks.Confirmation of ESBL phenotype was performed by the double disk diffusion method using antibiotic disks containing a combination of cephalosporin plus clavulanic acid in conjunction with a corresponding cephalosporin disk alone.The following antibiotic disks were used: ceftazidime (CAZ 30µg), ceftazidime plus clavulanic acid (CAZ/CA 30/10µg), cefotaxime (CTX 30µg) and cefotaxime plus clavulanic acid (CTX/CA 30/10µg).Antibiotics were all obtained from from Becton Dickinson, USA.The tests were interpreted according to CLSI guidelines [12] and ≥ 5 mm increase in the zone of inhibition for the CAZ/CLA and CTX/CLA containing disk versus the corresponding CAZ or CTX disk was considered positive for ESBL.
The antibiotic susceptibility pattern of the ESBLproducing isolates to a panel of antibiotics including amikacin, amoxicillin plus clavulanic acid, ciprofloxacin, gentamicin, imipenem, piperacillin and ticarcillin was recorded.E. coli ATCC 25922 was used to control for growth of isolates and potency of antibiotic disks.Data analysis (chi-squared test) was done using SigmaStat version 3.5 (Systat Software Inc, San Jose California, USA) and p < 0.05 was considered statistically significant.
Urine was the major source of ESBL-producing isolates in both in-patients (46.1%) and out-patients (74.4%) (Table 2).As the majority of the ESBL isolates in both patient populations were E. coli and of urinary source, we determined the proportion of urinary E. coli isolates which were ESBL producers in both patient populations.Among in-patients, 506 E. coli isolates came from urine specimens compared to 4,074 among out-patients.The proportion of urinary E. coli isolates which were ESBL producers was significantly higher among in-patients (53/506; 10.4%) compared to out-patients (182/4,074; 4.4%); p < 0.05.Among the 143 in-patients from whom ESBL isolates were cultured, there was available data which enabled us to categorize whether the infection was nosocomial or community acquired in 126 patients (88%).Of these, 75 (60%) were nosocomial compared to 51 (40%) which were community acquired.The majority (74%) of K. pneumoniae infections were nosocomial in origin.In contrast, distribution of E. coli isolates in nosocomial and community-acquired infections were 54% and 45% respectively.
All ESBL-producing isolates were sensitive to imipenem.Resistance to a panel of antibiotics including gentamycin, amikacin, augmentin, ciprofloxacin, ticarcillin and piperacillin was demonstrated.All the ESBL Klebsiella isolates were resistant to both piperacillin and ticarcillin and only two E. coli isolates were sensitive to these two antibiotics.The proportions of ESBL producing E. coli and Klebsiella isolates showing resistance to the other antibiotics were comparable.Comparison of | antibiotic resistance profiles revealed that with the exception of amikacin, higher levels of resistance was demonstrated in ESBL-producing isolates obtained from in-patients (Figure 1)

Discussion
This is the first study from the Arabian Gulf region describing ESBL detection in Enterobacteriaceae isolated from all specimen types, from different patient populations with noscomial and community-acquired infections.Within the Arabian Gulf region, high ESBL prevalence of 31.7% in Kuwait and 41% in the United Arab Emirates [9.10] have been reported among inpatients.For Saudi Arabia, reported ESBL rates vary from 8.5-38.5% [4][5][6][7][8].Thus in comparison to regional data, the finding of 6% ESBL producers in this study is on the lower end of the spectrum.This finding is also similar to data reported from surveys in some countries in Europe and Asia [2,13,14].
Although a predominance of either K. pneumoniae or E. coli ESBL isolates has been identified in different geographical regions, E. coli has emerged as the major source of ESBLs in the community.In this study, the proportion of E. coli (83%) is higher compared to that found in other studies but similar predominance of E. coli has been reported in data from the United Arab Emirates [9,10].Nosocomial infections caused by ESBL producing pathogens are associated with risk factors such as elderly age, prolonged hospitalization, previous antibiotic use, and presence of invasive devices [2,7].The demographics in this study indicates that old age ( > 60 years) remains a risk factor in our setting.As patients in critical care units are likely to have higher use of invasive devices such as urinary and vascular catheters, it is not surprising that up to 20% of ESBL isolates were from the intensive care unit.Urine was the most common specimen for the isolation of ESBL pathogens among in-patients and out-patients.Up to 74% of the total ESBL isolates obtained from out-patients were urinary isolates in comparison to the wider diversity of specimen types among in-patients.
However, the detection of ESBL producers among urinary E. coli isolates was significantly higher among in-patients.These findings have significant implications for empirical management of  patients with urinary tract infection using third generation cephalosporins or fluoroquinolones, especially as we also found high levels of ciprofloxacin resistance in the ESBL isolates identified.
To our knowledge, this is the first report presenting data differentiating between nocosomial and community acquired ESBL infections (based on internationally defined criteria) in Saudi Arabia.About 60% of ESBL isolates were associated with noscomial infections with a significantly higher proportion of K. pneumoniae isolated in nosocomial compared to community acquired infections.However, the proportion of E. coli in both settings was similar.These findings of occurrence of E. coliproducing ESBL isolates in the community is a cause for concern and is in keeping with the global trend of emergence of community-acquired infections caused by ESBL-producing strains, in particular those which habour the CTX-M gene [15].A recent report from Saudi Arabia showed 17.7% fecal carriage of ESBL isolates with the majority (over 80%) of these being E. coli [16].This observation indicates that ESBLproducing E. coli are circulating within the community and are associated with infections emerging from these settings.Carbapenems are widely regarded as the drugs of choice for treatment of infection caused by ESBL-producing organisms and remain useful in our setting as all isolates were sensitive to imipenem.As reduced imipenem susceptibility has been described in E. coli with CTX-M-type ESBL [17], we recommend molecular characterization of ESBL isolates circulating in our setting.In addition, continued antibiotic surveillance is needed to preserve the continued usefulness of these drugs.Our finding of high levels of resistance to non-beta lactam classes of antibiotics is in keeping with other reports and arises because these plasmid-mediated enzymes which are transferable between bacterial species also incorporate genetic material coding for resistance to other antibiotics [18].Treatment failures have been reported with the use of betalactam/beta-lactamase inhibitor combinations for infections caused by ESBL-producing organisms [19].The level of resistance reported here for amoxicillin-clavulanic is much higher than that reported in other studies [20,21] and also much higher than the 25% resistance to piperacillintazobactam among ESBL isolates described in United Arab Emirates [10], suggesting that this betalactam/ beta-lactamase inhibitor combination may not be useful in our setting.
In conclusion, the findings in this study document the emerging threat of ESBL pathogens in our setting with the occurrence of these strains as aetiological agents of infection in the hospital and community.While the findings shed light on Klebsiella species and E. coli, which are the predominant ESBL producers, we recommend further work on evaluating the ESBL types in these isolates as well as the prevalence of other ESBL-producing Gram negative bacteria which are emerging as pathogens of concern in the clinical setting.We advocate increased surveillance as well comprehensive multicenter/ multinational studies to address this emerging problem of ESBL-associated infections.

Figure 1 :
Figure 1: Comparison of pattern of antibiotic resistance in ESBL producing isolates obtained from in-patients and outpatients.

Table 1 :
Total number of ESBL and non-ESBL isolates isolated during the study period from in-and outpatients

Table 2 :
Distribution of ESBL isolates obtained from different body sites *Includes one K. oxytoca isolate ; ** Includes 2 K. oxytoca isolates