Evaluation of Efficiency of Screening Extended-Spectrum β-Lactamase-Producing Escherichia coli and Klebsiella pneumoniae in Hospitals Where the Bacteria Are Increasingly Prevalent

Abstract

The disk screening methods for extended-spectrum β-lactamase-producing strains were evaluated. The confirmatory work is reduced significantly in settings such as those in this study, by changing the cefpodoxime breakpoint to ≤20 mm and by not testing cefoxitin-resistant isolates. Cefotaxime and ceftazidime disk screening is reliable, and the laboratory-prepared cefotaxime- and ceftazidime-clavulanic acid disks are stable at −20°C for 12 weeks.

The detection of extended-spectrum β-lactamase (ESBL)-producing strains is important for optimal therapy of infected patients. The screening and confirmation of ESBL-producing isolates using the methods of the NCCLS (12) are easily performed. A survey in Korea suggested a high prevalence of ESBL-producing isolates (7), and with the increase of cefpodoxime disk screening-positive isolates, laboratories are confronted with an increasing burden of confirming results. Particularly, the almost daily preparation of clavulanic acid solution-supplemented ceftazidime (ceftazidime-clavulanic acid) and cefotaxime (cefotaxime-clavulanic acid) disks requires time.

Isolates producing certain ESBL types with low hydrolytic activity (4) are difficult to detect (14). In Korea, only TEM-52-, SHV-2a-, and SHV-12-type ESBLs have been reported (6, 13), and the cefpodoxime disk screening-positive isolates have often been ESBL nonproducers (Y. Chong, J. K. Lim, D. Yong, J. W. Yum, K. Lee, R. Okamoto, and M. Inoue, Abstr. 40th Intersci. Conf. Antimicrob. Agents Chemother., abstr. 1608, 2000). In this study, we evaluated the efficiency of NCCLS disk screening methods in Korea and determined the stability of the laboratory-prepared cefotaxime- and ceftazidime-clavulanic acid disks in storage.

Isolates of Escherichia coli and Klebsiella pneumoniae were collected in 1999 in Korea: in phase I, from a tertiary-care hospital, and in phase II, from 28 hospitals with various bed capacities and locations. Antimicrobial susceptibility was tested by the NCCLS disk diffusion method (11) using Mueller-Hinton agar plates and antimicrobial disks (Becton Dickinson Microbiology Systems, Sparks, Md.). ESBL production was screened using cefpodoxime disks (12). To confirm ESBL production, the isolates were tested by double-disk synergy tests (8) using an amoxicillin-clavulanic acid disk plus cefotaxime, ceftazidime, and aztreonam disks and by comparing the MICs (10, 12) of cefotaxime and ceftazidime with or without 4 μg of clavulanic acid (Smith-Kline Beecham, Worthing, United Kingdom) per ml.

Some characteristics of ESBLs were determined using transconjugants obtained by agar mating (2) using recipient E. coli RG 176 (Nal^r), RG 488 (Rif^r) (kindly provided by Dong Taek Cho, College of Medicine, Kyungpook National University, Taegu, Korea), or J53 (Az^r) (kindly provided by Chik Hyun Pai, College of Medicine, Ulsan University, Ulsan, Korea). PCR was performed using heat-extracted templates as described previously (5). The isoelectric point (pI) of β-lactamase (9) was determined by electrophoresis of the cell sonicate using a pH 3 to 10 gel and a ThermoFlow unit (Novex Experimental Technology, San Diego, Calif.) and staining with 0.7 mg of nitrocefin (Unipath, Basingstoke, United Kingdom) per ml. The cefotaxime- and ceftazidime-clavulanic acid disks, prepared as recommended by the NCCLS (12), were placed in vials with desiccant and stored at 4° and −20°C, and the inhibition zone diameters were determined using ESBL-producing strains: K. pneumoniae ATCC 700603, two clinical isolates of K. pneumoniae, and one clinical isolate of E. coli.

In the phase I study, 114 of 256 (44.5%) E. coli and 55 of 109 (50.5%) K. pneumoniae isolates were cefpodoxime disk screening positive (Table 1). Among them, 62 (24.2%) E. coli and 18 (16.5%) K. pneumoniae isolates were cefoxitin resistant, and among the cefoxitin-susceptible or -intermediate isolates, only 20 (7.8%) E. coli and 33 (30.2%) K. pneumoniae isolates were ESBL positive. ESBL-positive isolates of E. coli and K. pneumoniae showed cefpodoxime zone diameters of ≤18 and ≤14 mm, respectively.

TABLE 1.

Cefpodoxime disk zone diameter distribution for ESBL-producing and -nonproducing isolates

Cefpodoxime zone diam (mm)	No. of isolates for species, cefoxitin susceptibility, and ESBL statusa
	E. coli(n = 256)				K. pneumoniae(n = 109)
	Cefoxitin S (I)		Cefoxitin R		Cefoxitin S (I)		Cefoxitin R
	ESBL positive (n = 20)	ESBL negative (n = 174)	ESBL positive (n = 6)	ESBL negative (n = 56)	ESBL positive (n = 33)	ESBL negative (n = 58)	ESBL positive (n = 8)	ESBL negative (n = 10)
<13	14 (1)	0 (1)	5	53	31 (1)	1 (1)	8	8
13–14	2	2 (1)			1
15–16	1		1	1
17–18	2	5		2
19–20		4
21–22		17 (2)				2		2
23–24		38 (1)				10 (1)
25–26		42				14 (1)
27–28		31				10
>28		30				18
No. (%) screening positive	20 (7.8)	32 (12.5)	6 (2.3)	56 (21.9)	33 (30.2)	4 (3.7)	8 (7.3)	10 (9.2)

^aBoldface indicates isolates which were cefpodoxime-screening positive. Abbreviations: S (I), susceptible (intermediate); R, resistant. 

Efficiencies of screening of ESBL-producing isolates with disks of aztreonam, cefotaxime, and ceftazidime were compared (Table 2). All of the three antimicrobial disks detected all ESBL-producing isolates, but two and four isolates of E. coli were just within the breakpoint of aztreonam (27 mm) and ceftazidime (22 mm), respectively. Among the cefoxitin-nonresistant and ESBL-negative isolates, only 23 (13.2%), 22 (12.6%), and 12 (6.9%) E. coli isolates and 17 (29.3%), 20 (34.5%), and 3 (5.2%) K. pneumoniae isolates were positive by screening with aztreonam, cefotaxime, and ceftazidime, respectively.

TABLE 2.

Aztreonam, cefotaxime, and ceftazidime disk zone diameter distribution for ESBL-producing and -nonproducing isolates

Zone diam (mm)	No. of isolates for drug, species, cefoxitin susceptibility, and ESBL production status
	Aztreonam						Cefotaxime						Ceftazidime
	E. coli(n = 256)			K. pneumoniae(n = 109)			E. coli(n = 256)			K. pneumoniae(n = 109)			E. coli(n = 256)			K. pneumoniae(n = 109)
	Cefoxitin S/I		Cefoxitin R	Cefoxitin S/I		Cefoxitin R	Cefoxitin S/I		Cefoxitin R	Cefoxitin S/I		Cefoxitin R	Cefoxitin S/I		Cefoxitin R	Cefoxitin S/I		Cefoxitin R
	ESBL production positive	ESBL production negative		ESBL production positive	ESBL production negative		ESBL production positive	ESBL production negative		ESBL production positive	ESBL production negative		ESBL production positive	ESBL production negative		ESBL production positive	ESBL production negative
<11	2		2	12	1	6	1		49	2		7	2	1	3	11		6
11–15	2		4	15	1	2	6		5	12		5	2		19	15		8
16–17			1	4		3	5	3	1	8		1	5	1	24	6		2
18–19	3	2	3	1		1		1	2	7		1	4	1	11
20–21	3	2	15			2	3	3	1	4	1	2	3	2	2	1	1
22	1	2	8		1	2	1	2	1		1		4	7			2
23–24	2	3	10	1	1		3	3	1		4			10	1		11	1
25–26	5	2	8		5		1	6			6	1		32			13	1
27	2	12	7		8	1		4	2		8	1		11	1		13
28–29		22	4		10			28			12			35	1		9
>29		129			31	1		124			26			74			9
Total	20	174	62	33	58	18	20	174	62	33	58	18	20	174	62	33	58	18

^aBoldface indicates strains that screened positive. Abbreviations: S/I, susceptible or intermediate; R, resistant. 

In the phase II study (Table 3), the proportion of ESBL-positive isolates of E. coli was relatively higher (35.6%) than that in the phase I hospital, because of lower prevalence of cefoxitin-resistant isolates, but for K. pneumoniae, the rates were fairly similar in the two phases. The MICs of cefotaxime or ceftazidime were 2 μg/ml for six and two ESBL-producing isolates of E. coli and K. pneumoniae, respectively (data not shown), but all of them were positive by screening with cefpodoxime, cefotaxime, and ceftazidime disks. The bla_TEM and/or bla_SHV sequence was detected in all ESBL-producing transconjugants (Table 4), and their β-lactamases had a pI of 6.0, 7.6, or 8.2, suggesting that they were TEM-52, SHV-2a, or SHV-12 type, respectively. The mean MIC of ceftazidime was higher for isolates with a pI 8.2 β-lactamase.

TABLE 3.

Efficiency of cefpodoxime disk screening test for ESBL-producing isolates in a tertiary-care hospital and in other hospitals^a

Cefpodoxime zone diam (mm)	No. of isolates for species, cefoxitin susceptibility, and study phase
	E. coli(n = 232)				K. pneumoniae(n = 132)
	Cefoxitin S/I		Cefoxitin R		Cefoxitin S/I		Cefoxitin R
	Phase I	Phase II	Phase I	Phase II	Phase I	Phase II	Phase I	Phase II
<10	11	33	57	9	25	34	15	12
11–12	4	7	1	3	7	6	1	3
13–14	2	2			1	1
15–16	1		2
17–18	2		2				2
Total (%)	20 (17.5)	42 (35.6)	62 (54.4)	12 (10.2)	33 (62.3)	41 (51.9)	18 (34.0)	15 (19.0)

^aIsolates that were cefoxitin susceptible or intermediate were all positive for ESBL production. Results for cefoxitin-susceptible or -intermediate isolates that were negative for ESBL production are not shown. The numbers of isolates tested for each phase are as follows: phase I, 114 of E. coli and 55 of K. pneumoniae, and phase II, 118 of E. coli and 79 of K. pneumoniae. S/I, susceptible or intermediate; R, resistant. 

TABLE 4.

Comparison of MICs of expanded-spectrum cephalosporin and aztreonam for transconjugants by possession of TEM and SHV genes and pI patterns of β-lactamase

Gene(s) detected	No. of isolates			β-Lactamase with (pI)a:				Mean MIC (μg/ml) of drugb:
	E. coli	K. pneumoniae	Total (%)	5.4	6.0	7.6	8.2	CAZ	CTX	CRO	ATM
blaTEM	6	4	10 (20.0)	+	+			53	18	10	5
blaSHV	2	2	4 (8.0)			+		84	44	38	11
blaSHV	7		7 (14.0)				+	144	8	7	36
blaSHV		7	7 (14.0)				+	237	42	33	96
blaTEM + blaSHV	7	15	22 (44.0)	+			+	151	11	10	47

^aPossible β-lactamase types were TEM-52 (pI 6.0), SHV-2a (pI 7.6), and SHV-12 (pI 8.2). 

^bAbbreviations: CAZ, ceftazidime; CTX, cefotaxime; CRO, ceftriaxone; AZT, aztreonam. 

The inhibition zone diameter produced by the cefotaxime- and ceftazidime-clavulanic acid disks did not change significantly after 1 week at 4°C or after 12 weeks at −20°C (Table 5).

TABLE 5.

Effect of storage on the stability of laboratory-prepared ceftazidime- and cefotaxime-clavulanic acid disks

Storage (wk)	Zone diam (mm) for isolate, drug, and tempa
	K. pneumoniae ATCC 700603						K. pneumoniae 99-3-R251			K. pneumoniae 99-3-R256			E. coli 99-3-U609
	CAZ	CAZ-CLA		CTX	CTX-CLA		CAZ	CAZ-CLA		CAZ	CAZ-CLA		CAZ	CAZ-CLA
		4°C	−20°C		4°C	−20°C		4°C	−20°C		4°C	−20°C		4°C	−20°C
0	13	22	22	18	23	23	16	25	25	10	25	25	6	25	25
1	14	21	23	20	23	23	17	25	25	11	25	25	6	24	24
2	14	21	23	20	23	23	17	25	25	12	22	25	6	23	25
3	14	22	22	18	23	23	17	23	23	12	22	25	6	21	23
4	13	22	23	18	23	23	18	26	27	10	23	25	6	21	25
6	14	21	23	19	23	23	18	20	23	10	20	25	6	19	24
8	13	20	23	19	23	25	18	25	27	12	19	25	6	20	23
12	14	20	22	20	24	24	18	24	26	12	18	25	6	18	24
16	12	18	21	19	25	25	15	21	23	10	16	22	6	18	23
20	13	17	20	20	24	24	15	22	23	11	16	22	6	15	22

^aCeftazidime and cefotaxime disks were those used for routine susceptibility testing. Abbreviations: CAZ, ceftazidime; CAZ-CLA, ceftazidime-clavulanic acid; CTX, cefotaxime; CTX-CLA, cefotaxime-clavulanic acid. 

In the phase I study, 12.5% of E. coli isolates were cefpodoxime disk screening positive but ESBL negative. By reducing the screening breakpoint by 2 mm to ≤20 mm, the false-positive rate decreased by 59.4% in E. coli, but this did not influence the result significantly in K. pneumoniae (Table 1). In a U.S. study, 42.2% of cefpodoxime screening-positive E. coli isolates were ESBL nonproducers (F. C. Tenover, P. Raney, P. P. Williams, K. L. Brittain, C. D. Steward, S. K. Fridkin, R. P. Gaynes, and J. E. McGowan, Jr., Abstr. 40th Intersci. Conf. Antimicrob. Agents Chemother., abstr. 1606, 2000), which indicates the feasibility of eliminating unnecessary work by adjusting breakpoints in other countries, too.

CMY-1-producing strains continue to be prevalent in Korea (1), and AmpC β-lactamases have been reported worldwide, including the United States (3). It may not be significant to detect ESBL producers among cefoxitin-resistant isolates, as AmpC β-lactamase producers are resistant to all β-lactams except carbapenem, temocillin, and mecillinam (8). If so, 54.4% of cefpodoxime screening-positive E. coli isolates in the phase I study did not require a confirmatory test (Table 3). The reduction was only 10.2% in phase II hospitals, but such isolates may increase in the near future. In a U.S. study, 16.3% of cefpodoxime screening-positive E. coli isolates were ESBL nonproducers and were possibly AmpC β-lactamase producers (Tenover et al., 40th ICAAC).

Additional testing of cefpodoxime susceptibility for only E. coli and Klebsiella isolates may hinder laboratory workflow, as the species identification and susceptibility testing are usually performed simultaneously. If cefotaxime, ceftazidime, or aztreonam susceptibility is reliable for screening ESBL producers, its use can help streamline laboratory workflow, as the drugs are usually used for susceptibility testing of all aerobic gram-negative bacilli. In this study, some isolates were inhibited by 2 μg of cefotaxime or ceftazidime per ml, which is much lower than the >16 μg of ceftazidime per ml in a U.S. study (Tenover et al., 40th ICAAC), but all were screening test positive with cefotaxime and ceftazidime disks. However, two and four ESBL-producing E. coli isolates were just within the breakpoints of aztreonam and ceftazidime, respectively, suggesting that the current breakpoints may not be high enough in Korea.

In the past, preparation of screening disks has been a limiting factor in routine testing for ESBL-producing enteric organisms. The process was time-consuming for a busy laboratory, and the long-term stability of the disks was unknown. Recently, such disks have been reported to be stable for at least 14 days at −20°C (15), and in this study we have extended that observation for up to 12 weeks at −20°C. Although clavulanic acid-containing ceftazidime and cefotaxime disks have become commercially available, they may be too expensive for some laboratories to purchase and, hence, those laboratories may still have to prepare their own. The method described herein enables laboratories to prepare large batches of disks at one time and store them for several months. By using preprepared and stored disks, laboratories will have a cost- and time-saving alternative for the routine confirmation of ESBL-producing isolates.