Detection of Extended-Spectrum {beta}-Lactamases by Using Boronic Acid as an AmpC {beta}-Lactamase Inhibitor in Clinical Isolates of Klebsiella spp. and Escherichia coli

We evaluated highly sensitive methods using boronic acid (BA)to detect extended-spectrum ß-lactamase (ESBL) production.A total of 182 clinical isolates of Klebsiella spp. (n = 118)and Escherichia coli (n = 64) were analyzed: 62 harbored onlyESBLs, 80 harbored both ESBLs and plasmid-mediated AmpC ß-lactamases(pAmpCs), and 40 harbored only pAmpCs. The CLSI confirmatorytest detected all isolates that produce only ESBLs but detected85% of isolates that produce both enzymes. When a $≥$ 5-mm increasein the zone diameter of either the cefotaxime (CTX) or the ceftazidime(CAZ) disk in the presence of both clavulanic acid (CA) andBA was considered to be a positive result, the test detectedall isolates that harbor ESBLs (± pAmpCs) but showedfrequent false-positive results (50%) for isolates that produceonly pAmpCs. Meanwhile, when a $≥$ 3-mm increase in the zone diameterof either the CTX/BA or the CAZ/BA disk in the presence of CAwas considered to be a positive result, the test also detectedall isolates that harbor ESBLs (± pAmpCs) and showedless frequent false-positive results (5%) in isolates that produceonly pAmpCs. The latter new interpretive guideline has enhanceddetection of ESBLs in clinical isolates of Klebsiella spp. andEscherichia coli and allowed detection of an ESBL even whenpotentially masked by a pAmpC.

INTRODUCTION

Top

Introduction

Since extended-spectrum ß-lactamases (ESBLs) wereinitially reported to be in Klebsiella ozaenae in the firsthalf of the 1980s in Germany, ESBLs have been spreading globally(2). The CLSI interpretive guidelines stated that strains ofKlebsiella spp. and Escherichia coli that produce ESBLs maybe clinically resistant to therapy with penicillins, cephalosporins,or aztreonam, despite apparent in vitro susceptibility to someof these agents (4). Therefore, detection of ESBLs in Klebsiellaspp. and E. coli isolates is crucial for optimal treatment ofpatients and to control the spread of resistance. The CLSI recommendsa phenotypic confirmatory test for ESBL production that consistsof measuring the zone diameters of cefotaxime (CTX; 30 µg)or ceftazidime (CAZ; 30 µg) disks with or without clavulanicacid (CA; 10 µg) in Klebsiella spp. and E. coli. An increasein the zone diameter of $≥$ 5 mm in the presence of CA is takento be a phenotypic confirmation of ESBL production.

There are numerous reports in which Klebsiella pneumoniae andE. coli isolates have been found to produce both plasmid-mediatedAmpC ß-lactamases (pAmpCs) and ESBLs (6, 8). In Korea,8.7% of pAmpC producers also produced ESBLs (8). AmpCs resistinhibition by CA, and hence, the presence of an ESBL can bemasked by the expression of an AmpC. Thus, the coexistence ofboth pAmpCs and ESBLs in the same strain may result in false-negativetests for the detection of ESBLs (5). There is a need, therefore,for an alternative method that can detect ESBLs in Klebsiellaspp. and E. coli isolates with a high sensitivity, even thoughthe isolates simultaneously harbor pAmpCs.

Boronic acids (BAs) were reported as reversible inhibitors ofAmpCs (1). There are a few reports of studies designed to detectpAmpC-producing isolates of Klebsiella spp. and E. coli by amethod using BA as an inhibitor that closely resembles the phenotypicconfirmatory tests for ESBLs (3, 5). And Coudron (5) recentlydescribed that a BA disk also enhances detection of isolatesthat harbor both ESBLs and pAmpCs.

The aim of this work was to modify the phenotypic confirmatorytest for ESBL production by using BA disks to improve its sensitivityfor the detection of ESBLs in Klebsiella spp. and E. coli isolateseven though the isolates simultaneously harbor pAmpCs.

MATERIALS AND METHODS

Top

Materials and Methods

Bacterial strains. A total of 182 nonrepeat ESBL- and/or pAmpC-producing clinicalisolates of Klebsiella spp. (n = 118) and E. coli (n = 64) collectedfrom 12 tertiary-care hospitals in Korea during 2004 and 2005were included in this study (Table 1) : 62 harbored ESBLs butnot pAmpCs, 80 harbored both ESBLs and pAmpCs, and 40 harboredpAmpCs but not ESBLs. The isolates were identified with API-20E systems (bioMérieux, Marcy l'Etoile, France). Searchesfor genes coding for the class A ESBLs and pAmpCs were performedby PCR amplification as described previously (7, 8). The PCRproducts were subjected to direct sequencing. Both strands ofeach PCR product were sequenced twice with an automatic sequencer(model 3730xl; Applied Biosystems, Weiterstadt, Germany).

View this table:
[in this window]
[in a new window]

TABLE 1. Bacterial strains used for the validation of test methods

ESBL detection by the CLSI confirmatory test and two different tests using BA. Disks containing BA were produced in-house by using commerciallyavailable antibiotic-containing disks (BBL, Cockeysville, MD)as described previously (5). Briefly, 20 µl of the 3-aminophenylBA stock solution (20 g/liter; Sigma, St. Louis, MO) was dispensedonto each disk containing CTX (30 µg) or CAZ (30 µg)with or without CA (10 µg). The final amount of BA onthe disks was 400 µg. The disks were allowed to dry for60 min and used immediately or stored in airtight vials withdesiccant at 4°C. The CLSI confirmatory test for ESBL productionwas carried out on Mueller-Hinton agar (Difco Laboratories,Detroit, MI) by using antibiotic-containing disks with or withoutCA. In two new different tests using BA (400 µg), (i)a $≥$ 5-mm increase in the zone diameter of the disk containingCTX (and/or CAZ) and CA tested in combination with BA (CTX/CA/BAand/or CAZ/CA/BA) versus CTX and/or CAZ alone was consideredpositive for ESBL, and (ii) a $≥$ 3-mm increase in the zone diameterof the disk containing CTX (and/or CAZ) and CA tested in combinationwith BA (CTX/CA/BA and/or CAZ/CA/BA) versus CTX and/or CAZ containingBA (CTX/BA and/or CAZ/BA) was considered positive for ESBL.CTX, CTX/CA, CAZ, and CAZ/CA disks with or without BA were spacedover the agar surface. After overnight incubation at 35°Cin ambient air, the inhibitory-zone diameters of the antibiotic-containingdisks were measured.

RESULTS AND DISCUSSION

Top

Results and Discussion

All 62 clinical isolates that produce ESBLs but not pAmpCs showed $≥$ 5-mm increases in the zone diameters of the combined disks (CTX[or CAZ]/CA or CTX [or CAZ]/CA/BA) compared with those for CTXor CAZ alone (Table 2). The average increases in the zone diametersof the CTX and CAZ disks in the presence of both CA and BA (13.1mm and 12.6 mm, respectively) were higher than those for theCLSI confirmatory test (11.3 mm and 10.9 mm, respectively).In tests with CA, the CTX and CAZ disks yielded positive testswith 60 (96.8%) and 58 (93.5%) of 62 isolates, respectively(Fig. 1A and B). In tests with both CA and BA, the CTX and CAZdisks yielded positive tests with 61 (98.4%) and 62 (100%) of62 isolates, respectively (Fig. 1C and D). Meanwhile, all 62clinical isolates showed $≥$ 3-mm increases in the zone diametersof both the CTX/BA and the CAZ/BA disks in the presence of CA,with the exception of a K. pneumoniae isolate that producesSHV-2a, which showed a 2-mm increase in the zone diameter ofa CAZ/BA disk in the presence of CA (Fig. 1E and F). The averageincreases in the zone diameters of the CTX/BA and CAZ/BA disksin the presence of CA were 11.8 mm and 11.1 mm, respectively.

View this table:
[in this window]
[in a new window]

TABLE 2. Test results for the 182 ESBL- and/or pAmpC-producing clinical isolates of Klebsiella spp. (n = 118) and E. coli (n = 64)

View larger version (52K):
[in this window]
[in a new window]

FIG. 1. Comparison of the results for the CLSI confirmatory test (A and B) and two new tests (C to F). The abbreviations (CTX/CA versus CTX, CAZ/CA versus CAZ, CTX/CA/BA versus CTX, CAZ/CA/BA versus CAZ, CTX/CA/BA versus CTX/BA, and CAZ/CA/BA versus CAZ/BA) are explained in footnote a to Table 2. Black circle, Klebsiella spp.; white circle, E. coli.

The rate of detection of ESBLs by the CLSI confirmatory testfor clinical isolates that produce both ESBLs and pAmpCs waslower than that for clinical isolates that produce ESBLs butnot pAmpCs (Table 2). Among 80 clinical isolates that produceboth ESBLs and pAmpCs, 68 (85%) showed $≥$ 5-mm increases in thezone diameters of either the CTX or the CAZ disks in the presenceof CA. In this test, the CTX and CAZ disks yielded positivetests with 51 (63.8%) and 45 (56.3%) of 80 isolates, respectively(Fig. 1A and B). All 80 clinical isolates, however, showed $≥$ 5-mmincreases in the zone diameters of either the CTX or the CAZdisks in the presence of both CA and BA (Table 2). In this test,CTX and CAZ disks yielded positive tests with 80 (100%) and75 (93.8%) of 80 isolates, respectively (Fig. 1C and D). Theaverage increases in the zone diameters of the CTX and CAZ disksin the presence of both CA and BA (11.9 mm and 9.9 mm, respectively)were higher than those for the CLSI confirmatory test (7.1 mmand 5.1 mm, respectively). All 80 clinical isolates also showed $≥$ 3-mm increases in the zone diameters of either the CTX/BA orthe CAZ/BA disks in the presence of CA (Table 2). In this test,76 (95%) and 68 (85%) of 80 isolates showed $≥$ 3-mm increases inthe zone diameters of the CTX/BA and CAZ/BA disks in the presenceof CA, respectively (Fig. 1E and F). The average increases inthe zone diameters of the CTX/BA and CAZ/BA disks in the presenceof CA were 8.7 mm and 7.9 mm, respectively.

Among the 40 clinical isolates that produce pAmpCs but not ESBLs,6 (15%) and 20 (50%) isolates showed $≥$ 5-mm increases in the zonediameters of either the CTX or the CAZ disks in the presenceof CA and both CA and BA, respectively (Table 2). Meanwhile,only 2 (5%) of the 40 isolates showed $≥$ 3-mm increases in thezone diameters of either the CTX/BA or the CAZ/BA disks in thepresence of CA (Table 2).

In summary, the CLSI confirmatory test detected ESBLs in allclinical isolates that produce ESBLs but not pAmpCs, but thedetection rate (85%) was lowered in clinical isolates that produceboth ESBLs and pAmpCs. The test showed false-positive resultsin 6 (15%) of 40 clinical isolates that produce pAmpCs but notESBLs. When a $≥$ 5-mm increase in the zone diameter of either theCTX or the CAZ disk in the presence of both CA and BA was consideredto be a positive result for ESBL production, the new test detectedall clinical isolates that harbor ESBLs (± pAmpCs) butshowed frequent false-positive results (20/40, 50%) for clinicalisolates that produce pAmpCs but not ESBLs. An inhibitory effectof BA on pAmpCs may have an influence on the high frequencyof false-positive results. Meanwhile, when a $≥$ 3-mm increase inthe zone diameter of either the CTX/BA or the CAZ/BA disk inthe presence of CA was considered to be a positive result forESBL production, the test also detected all clinical isolatesthat harbor ESBLs (± pAmpCs) and the test showed lessfrequent false-positive results (2/40, 5%) for clinical isolatesthat produce pAmpCs but not ESBLs. In the case of the CTX/BAdisk in the presence of CA, there were no false-positive results(Fig. 1E). The test missed none of the ESBL-producing isolates.Coudron (5) recently reported that a BA-based method using cefotetanenhanced detection of pAmpCs, but Coudron's method missed oneof the pAmpC-producing Proteus mirabilis isolates. AlthoughCoudron reported that the method can also detect isolates harboringpAmpCs and ESBLs, only three isolates were tested for the presenceof both enzymes, and they did not include E. coli isolates.

In conclusion, we describe a newly modified method with a highsensitivity and a high specificity for the detection of ESBLsin clinical isolates of Klebsiella spp. and E. coli. A $≥$ 3-mmincrease in the zone diameter of either the CTX/BA or the CAZ/BAdisk in the presence of CA is a new interpretive guideline thathas enhanced detection of ESBLs in clinical isolates of Klebsiellaspp. and E. coli and allowed detection of an ESBL even whenpotentially masked by a pAmpC. The exact detection of ESBLsin isolates that produce both ESBLs and pAmpCs is importantfor both treatment and epidemiology.

ACKNOWLEDGMENTS

This work was supported by a research grant from the Kosin UniversityCollege of Medicine and the National Institute of Health (NIH)of the KCDC in the Republic of Korea.

FOOTNOTES

* Corresponding author. Mailing address: Department of Laboratory Medicine, Kosin University College of Medicine, 34 Amnam-Dong, Suh-Gu, Busan, Republic of Korea 602-030. Phone: 82-51-990-6373. Fax: 82-51-990-3034. E-mail: kscpjsh{at}ns.kosinmed.or.kr

Published ahead of print on 14 February 2007.

REFERENCES

Top