Rapid Detection of β-Lactamase-Hydrolyzing Extended-Spectrum Cephalosporins in Enterobacteriaceae by Use of the New Chromogenic βLacta Test

Abstract

The chromogenic βLacta test developed for the rapid detection of β-lactamase-hydrolyzing extended-spectrum cephalosporins in Enterobacteriaceae revealed good performance with extended-spectrum β-lactamase (ESBL) producers (97.5% true-positive results). However, false-negative results occurred with chromosomal AmpC hyperproducers and plasmid AmpC producers, whereas uninterpretable results were mostly due to VIM-1 carbapenemase producers and possibly low levels of expressed ESBLs.

TEXT

Detection of Enterobacteriaceae resistant to broad-spectrum cephalosporins, mostly due to production of extended-spectrum β-lactamases (ESBL), plasmid AmpC β-lactamases, and/or carbapenemases, has become a challenge in clinical microbiology laboratories because of important clinical consequences for infection control purposes and guidance of antimicrobial therapy (1–3). Methods routinely used to detect these organisms are primarily based on susceptibility testing results, either MICs or disk diffusion inhibition zones, as well as on ancillary testing using disk synergy tests with different β-lactamase inhibitors or MIC-gradient strips combining β-lactams and β-lactamase inhibitors (4). Molecular methods based on PCR or microarray hybridization techniques have been also developed (5, 6). In addition, mass spectrometry-based protocols using matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) and in-house colorimetric tests have been developed to detect the production of ESBLs in less than 4 h (7, 8). Increased interest in rapid colorimetric assays has been observed, because of their easy implementation in the routine workflow of clinical laboratories (8).

The βLacta test (Bio-Rad, Marnes la Coquette, France) is a new chromogenic method based on the use of a yellow substrate (HMRZ-86) that turns to red when hydrolyzed by ESBLs, AmpC β-lactamases, and most carbapenemases (9–11). According to the manufacturer, reading of the results can be performed visually in less than 15 min. In the present study, we assessed the performance of the βLacta test for rapid detection of β-lactamase-hydrolyzing extended-spectrum cephalosporins in two groups of Enterobacteriaceae clinical isolates. The first group (Table 1) consisted of 338 contemporary clinical isolates collected prospectively (in January to March 2012), and the second group (Table 2) included 106 clinical isolates with β-lactamase-mediated resistance mechanisms that were characterized at the molecular level and affected broad-spectrum cephalosporins (12–15). All isolates were recovered at the Ramón y Cajal University Hospital and were identified using both a MicroScan system (Siemens, West Sacramento, CA) and MALDI-TOF MS (Bruker Daltonics, Germany). In addition, the Escherichia coli ATCC 35218 strain (TEM-1 producer) was used as a negative control, whereas the Klebsiella pneumoniae ATCC 700603 strain (SHV-18 producer) was used as a positive control. Susceptibility testing of β-lactam antibiotics, including broad-spectrum cephalosporins (cefotaxime, ceftazidime, and cefepime) and carbapenems (imipenem and ertapenem), was performed using the MicroScan system (Siemens). EUCAST breakpoint criteria were used to define susceptible, intermediate, and resistant categories (16). β-Lactam MICs as well as in-house and commercial (Rosco-Diagnostica A/S, Taastrup, Denmark) ancillary test results (double-disk diffusion synergy techniques) using β-lactamase inhibitors (clavulanic acid, EDTA, dipicolinic acid, cloxacillin, and boronic acid) were used to infer phenotypes and resistance mechanisms affecting broad-spectrum cephalosporins (17). Isolates were classified according to these results.

TABLE 1.

Contemporary clinical Enterobacteriaceae isolates (n = 338)

Microorganism (no. of isolates)	Phenotype or β-lactamase	No. with βLACTA test result of:
		Positive	Negative	Uninterpretable
Escherichia coli (231)	Wild type		65	1
	Penicillinase production		77
	Penicillinase hyperproduction		31
	ESBLa	50		3b
	ESBL + plasmid AmpC	1
	Plasmid AmpC		3c
Klebsiella spp. (67)	Wild type (penicillinase)		41
	Penicillinase hyperproduction		6
	ESBL	18
	ESBL + plasmid AmpC	1
	Carbapenemase			1d
Proteus mirabilis (12)	Wild type		7
	Penicillinase		5
Enterobacter spp. (13)	Wild-type (inducible AmpC)		11
	AmpC hyperproduction		1c
	ESBL	1
Serratia marcescens (6)	Wild type (inducible AmpC)		3
	AmpC hyperproduction	3
Morganella morganii (4)	Wild type (inducible AmpC)		2
	AmpC hyperproduction		1c
	ESBL	1
Citrobacter freundii (2)	Wild type (inducible AmpC)		1
	AmpC hyperproduction	1
Providencia stuartii (3)	Wild type (inducible AmpC)		2
	AmpC hyperproduction		1c
Total		76	257	5

^aESBL, extended-spectrum β-lactamase.

^bSHV-12 (n = 2) and SHV-2 (n = 1).

^cFalse-negative results.

^dVIM-1 producer.

TABLE 2.

Enterobacteriaceae isolates (n = 106) with characterized β-lactamases

Microorganism (no. of isolates)	Phenotype or β-lactamase	No. with βLACTA test result of:
		Positive	Negative	Uninterpretable
Escherichia coli (20)	ESBLa (1 CTX-M-9, 4 CTX-M-14, 5 CTX-M-15)	10
	KPC (1 KPC-1, 1 KPC-2)	2
	VIM-1			3
	VIM-1 + SHV-12	1
	CMY-2		3b
	OXA-48 + VIM-1 + CTX-M-15	1
Klebsiella spp. (65)	ESBL (2 TEM-4, 8 SHV-12, 1 CTX-M-9, 2 CTX-M-14, 3 CTX-M-10, 6 CTX-M-15)	22
	KPC (6 KPC-2, 13 KPC-3)	19
	KPC-3 + VIM-1	3
	VIM-1	7		2
	VIM-1 + SHV-12 + TEM-1	2
	OXA-48	3
	OXA-48 + CTX-M-15	3
	OXA-48 + VIM-1 + CTX-M-15	4
Enterobacter spp. (18)	ESBL (3 CTX-M-1, 1 SHV-2)	4
	KPC (1 KPC-2, 1 KPC-3)	2
	VIM-1	10
	OXA-48	1
	OXA-48 + CTX-M-15	1
Serratia marcescens (1)	VIM-1	1
Citrobacter freundii (2)	VIM-1	2
Total		98	3	5

^aESBL, extended-spectrum β-lactamase.

^bFalse-negative results.

Molecular characterization of bla genes was performed as described previously (12–15), for both contemporary clinical isolates (first group) with discrepant results in the βLacta test and isolates with well-characterized resistance mechanisms (second group). The βLacta test was performed according to the manufacturer's instructions. Briefly, several colonies from 18- to 20-h cultures were picked from blood agar medium, to yield a full 1-μl loop, and then were mixed with the βLacta test reagents (R1 and R2, one drop each) in plastic microtubes. The mixtures were left at room temperature, and color changes, when present, were read and interpreted after 2 and 15 min by following the manufacturer's instructions. Color changes were interpreted as follows: (i) red or purple, positive; (ii) no change (yellow), negative; (iii) orange, uninterpretable. The accuracy of the βLacta test was determined by using susceptibility test results and considering either a resistance or intermediate result for ceftazidime, cefotaxime, or cefepime as the reference value. Sensitivity, specificity, and likelihood ratios were determined. Confidence intervals (CIs) were estimated by Taylor's method.

Considering all routine clinical isolates and excluding isolates with uninterpretable results (Table 1), 96.7% (327/338 isolates) gave expected results (251 isolates yielded true-negative results and 76 true-positive results), in accordance with the inferred phenotype. However, six isolates (1.8%) gave unexpected negative (false-negative) results, including 3 E. coli isolates expressing CMY-2 plasmid AmpC β-lactamase and one isolate each of Enterobacter cloacae, Providencia stuartii, and Morganella morganii expressing a hyperproduced AmpC phenotype (see Table S1 in the supplemental material). Moreover, 5 isolates (1.5%) had uninterpretable results, including one β-lactam-susceptible E. coli isolate, three ESBL-producing E. coli isolates expressing either SHV-2 (n = 1) or SHV-12 (n = 2), and one VIM-1-producing K. pneumoniae isolate. The MIC values, resistance phenotypes, and β-lactamase types of these strains are included in Table S2 in the supplemental material. Interpretive βLacta test results for the routine clinical isolates, considering both the resistance phenotype and the resistance mechanism as gold standards, were as follows: sensitivity, 92.7% (95% CI, 88.8 to 97.3%); specificity, 100% (95% CI, 98.5 to 100.0%); positive predictive value, 100% (95% CI, 95.3 to 100.0%); negative predictive value, 97.7% (95% CI, 94.9 to 99.2%). The negative likelihood ratio was 0.07.

These results prompted us to investigate the performance of the βLacta test with a collection of isolates with well-characterized resistance mechanisms. Within this collection, including isolates expressing ESBLs, plasmid AmpC β-lactamases, or carbapenemases (Table 2), 97.03% (98/101 isolates) yielded expected results (0 true-negative results and 98 true-positive results), 5 isolates yielded uninterpretable results (3 E. coli isolates and 2 K. pneumoniae isolates expressing VIM-1), and 3 isolates yielded false-negative results (3 CMY-2-producing E. coli isolates).

These results confirmed that the βLacta test is useful for the detection of ESBL-producing organisms (97.5% of all ESBL producers demonstrated true-positive results) but not AmpC producers (either plasmid-mediated or chromosomally mediated). This was also noted in a multicenter evaluation performed in French and Belgian hospitals, in which poor performance of the βLacta test with AmpC producers was observed (11). Unlike in our study, none of ESBL producers in the multicenter evaluation gave negative results, a situation that occurred with 3 E. coli isolates (two expressing SHV-12 and one expressing SHV-2) in our work. Extended-spectrum cephalosporins were variably affected, with MICs ranging from ≤1 to >16 mg/liter (see Table S2 in the supplemental material). Variable expression of ESBLs was described several years ago, particularly with SHV variants, affecting the hydrolysis of cephalosporins and MIC values (18). This could also eventually affect the hydrolysis of HMRZ-86, a fact that has been shown with nitrocefin and in vitro variants obtained by mutagenesis of bla_SHV-1 (19).

On the other hand, results for carbapenemase producers (n = 67) were mainly positive (92.5% [62/67 isolates]), including results for producers of OXA-48, an enzyme with minor hydrolytic activity against extended-spectrum cephalosporins (20). Uninterpretable results (7.5% [5/67 isolates]) were specifically associated with VIM-1 producers. The variable expression of resistance phenotypes in these isolates might affect extended-spectrum and carbapenem MICs, as we demonstrated previously (13).

The isolates with unexpected negative (yellow) or uninterpretable (orange) results were tested under different conditions in order to enhance enzymatic activity to favor a positive result (red). For this purpose, and to enhance β-lactamase release, a bacterial suspension of each tested isolate was frozen and thawed twice and then incubated with the R1 and R2 reagents at room temperature and at 37°C, in separate tubes. Cells disrupted by sonication were also tested. In all cases, the results remained unchanged (data not shown).

In summary, the chromogenic βLacta test was demonstrated to be a rapid and reliable assay for the detection of ESBL-producing Enterobacteriaceae and presumably KPC and class D (OXA-48) carbapenemase producers. However, results were negative for AmpC β-lactamase producers, including both chromosomal AmpC hyperproducers and plasmid AmpC producers. Moreover, the percentage of uninterpretable results was low, mostly due to VIM-1 carbapenemases and possibly low levels of expressed ESBLs. Use of the chromogenic βLacta test might provide useful early guidance for detection of ESBLs in Enterobacteriaceae in routine clinical microbiology laboratories.