Abstract

Objectives

The objective of this study was to evaluate the proficiency of Spanish laboratories with respect to accurate susceptibility testing and the detection and interpretation of quinolone resistance phenotypes in Enterobacteriaceae.

Methods

Thirteen strains of Enterobacteriaceae were sent to 62 participating centres throughout Spain; strains harboured GyrA/ParC modifications, reduced permeability and/or plasmid-mediated quinolone resistance genes. The centres were requested to evaluate nalidixic acid and five quinolones, provide raw/interpreted clinical categories and to detect/infer resistance mechanisms. Consensus results from reference centres were used to assign minor, major and very major errors (mEs, MEs and VMEs, respectively).

Results

Susceptibility testing in the participating centres was frequently performed using the MicroScan WalkAway, Vitek 2 and Wider systems (48%, 30% and 8%, respectively). CLSI/EUCAST breakpoints were used in 71%/29% of the determinations. The percentage of VMEs for all quinolones was well below 2%. Only ofloxacin and moxifloxacin showed higher values for raw VMEs (6.6%), which decreased to 0% and 2.9%, respectively, in the interpreted VMEs. These errors were particularly associated with the CC-03 strain [qnrS2 + aac(6′)-Ib-cr]. For MEs, percentages were always <10%, except in the case of ofloxacin and nalidixic acid. There was a significantly higher percentage of all types of errors for strains whose MICs were at the border of clinical breakpoints.

Conclusions

The use of different breakpoints and methods, the complexity of mutation-driven and transferable resistance mechanisms and the absence of specific tests for detecting low-level resistance lead to high variability and represent a challenge to accuracy in susceptibility testing, particularly in strains with MICs on the border of clinical breakpoints.

Introduction

Over the past three decades, quinolone resistance has increased in Enterobacteriaceae among human and veterinary isolates. Fluoroquinolone resistance mainly occurs as a result of mutations in chromosomal genes encoding the quinolone targets DNA gyrase and topoisomerase IV.1 Mutation-driven mechanisms also include loss of porins and overexpression of one or more of several efflux pumps (such as AcrAB-TolC), which can lead to resistance or decreased susceptibility to β-lactams and fluoroquinolones.2 Three plasmid-mediated quinolone resistance (PMQR) mechanisms have also been described: (i) Qnr proteins (members of the pentapeptide repeat protein family, which bind to DNA gyrase and topoisomerase IV and protect the quinolone targets); (ii) the Aac(6′)-Ib-cr enzyme (which acetylates not only aminoglycosides but also ciprofloxacin and norfloxacin); and (iii) QepA and OqxAB plasmid-mediated efflux pumps.3 Five types of qnr determinants encoded by the qnrA, qnrB, qnrS, qnrC and qnrD genes have been described in Enterobacteriaceae.3,4 Several of these mechanisms must be combined to achieve a clinical level of resistance.

Enterobacteriaceae, particularly Escherichia coli, Klebsiella spp. and Enterobacter spp., are among the most common organisms causing community, nosocomial and opportunistic infections. The emergence of a very close association between quinolone resistance and resistance to other antimicrobial agents, particularly β-lactams and aminoglycosides, is a critical problem when managing such infections. One of the most striking features of these species is their remarkable capacity for developing antibiotic resistance. The prevalence of infections caused by MDR Enterobacteriaceae strains is increasing globally (with resistance rates at around 30%; http://www.eurosurveillance.org/),5 often rendering antibiotic treatments useless and causing a significant rise in the morbidity and mortality of affected patients.

The task of clinical microbiologists to properly perform antimicrobial susceptibility testing and interpretation and carry out complementary phenotypic/genotypic tests proficient in detecting the resistance mechanisms involved is the necessary first step for implementing appropriate treatments and antibiotic use policies.6 In this regard, there are no reliable phenotypic tests available for detecting PMQR mechanisms, either for strains with borderline susceptibility to these determinants alone or for resistant strains harbouring additional chromosomal mutations.4,7 One disturbing consequence of this situation is the underestimation of the prevalence of acquired and/or mutational resistance mechanisms, which often leads to the application of erroneous antibiotic treatments and subsequent clinical failure. There is limited information about the proficiency of clinical microbiology laboratories in overcoming this challenge. Some previous studies have shown the importance of correctly evaluating resistance to other antimicrobial agents, such as β-lactams, in well-characterized strains of Enterobacteriaceae or Pseudomonas aeruginosa.8,9 The objective of this study was to assess the ability of Spanish clinical microbiology laboratories to properly carry out susceptibility testing and to infer underlying resistance mechanisms in a collection of well-characterized Enterobacteriaceae strains, including those with complex combinations of mutational and transferable quinolone resistance mechanisms causing low-level quinolone resistance (LLQR) phenotypes.

Methods

Bacterial strains, characterization of resistance mechanisms and susceptibility testing

Thirteen Enterobacteriaceae strains, coded CC-00 to CC-12, were selected for this study (Table 1). They included 12 clinical isolates or laboratory strains (CC-01 to CC-12) with different chromosomal and/or horizontally acquired resistance mechanisms and one reference strain: E. coli ATCC 25922 (CC-00). Identification of the strains, antimicrobial susceptibility testing and confirmation of the quinolone resistance mechanisms were verified independently at the two reference laboratories used for this study: the Department of Microbiology, School of Medicine, University of Seville, Seville, Spain, and the Service of Microbiology, University Hospital Marqués de Valdecilla, Santander, Spain. API 20-E (bioMérieux, Marcy-l'Étoile, France) was used for species identification. Testing for susceptibility to ciprofloxacin, levofloxacin, moxifloxacin, ofloxacin, norfloxacin and nalidixic acid was performed in duplicate at both the reference centres by disc diffusion, CLSI broth microdilution and Etest (bioMérieux, Madrid, Spain). The 2013 CLSI and EUCAST (http://www.eucast.org/clinical_breakpoints/) breakpoints were used to interpret clinical categories.10,11

Table 1.

Panel of Enterobacteriaceae strains sent to the 62 participating centres

Strain No. Species and characteristics (known quinolone resistance mechanisms) Source/reference 
E. coli ATCC 25922 CC-00 CLSI control strain laboratory collection 
Ea-C2653 CC-01 Enterobacter aerogenes clinical isolate harbouring Asp86Tyr substitution in GyrA and the aac(6′)-Ib-cr gene 27 
Ecl-6087442 CC-02 Enterobacter cloacae clinical isolate harbouring the qnrA1 gene 28 
Kox-6059371 CC-03 Klebsiella oxytoca clinical isolate harbouring qnrS2 and the aac(6′)-Ib-cr gene 28 
Cf-6067553 CC-04 Citrobacter freundii clinical isolate harbouring the qnrB48 gene 28 
Ec-C1984 CC-05 E. coli clinical isolate harbouring Ser83Leu and Asp87Asn substitutions in GyrA, the Ser80Ile substitution in ParC and the aac(6′)-Ib-cr and oqxAB genes 29 
Ec-C1550 CC-06 E. coli clinical isolate harbouring Ser83Leu and Asp87Asn substitutions in GyrA, the Ser80Ile substitution in ParC and the qepA1 gene 29 
Kpn-LB4 CC-07 Klebsiella pneumoniae clinical isolate harbouring Ser83Tyr and Val112Ile substitutions in GyrA, deficient in both OmpK35 and OmpK36, and expressing active efflux of fluoroquinolones 30,31 
Ec-25922 mut1 CC-08 E. coli ATCC 25922 harbouring the Ser83Leu substitution in GyrA and the Ser80Arg substitution in ParC (genetic replacement) 19 
Ec-25922 mut2 CC-09 E. coli ATCC 25922 harbouring the Ser83Leu substitution in GyrA, the Ser80Arg substitution in ParC (genetic replacement) and marR deficient (chromosomal inactivation) 19 
Ec-25922 mut3 CC-10 E. coli ATCC 25922 harbouring Ser83Leu and Asp87Asn substitutions in GyrA and the Ser80Arg substitution in ParC (genetic replacement) 19 
Ec-25922 mut4 CC-11 E. coli ATCC 25922 harbouring Ser83Leu and Asp87Asn substitutions in GyrA, the Ser80Arg substitution in ParC (genetic replacement) and marR deficient (chromosomal inactivation) 19 
Ec-38–27 CC-12 E. coli clinical isolate belonging to the ST131 clone harbouring the Ser83Leu substitution in GyrA 32 
Strain No. Species and characteristics (known quinolone resistance mechanisms) Source/reference 
E. coli ATCC 25922 CC-00 CLSI control strain laboratory collection 
Ea-C2653 CC-01 Enterobacter aerogenes clinical isolate harbouring Asp86Tyr substitution in GyrA and the aac(6′)-Ib-cr gene 27 
Ecl-6087442 CC-02 Enterobacter cloacae clinical isolate harbouring the qnrA1 gene 28 
Kox-6059371 CC-03 Klebsiella oxytoca clinical isolate harbouring qnrS2 and the aac(6′)-Ib-cr gene 28 
Cf-6067553 CC-04 Citrobacter freundii clinical isolate harbouring the qnrB48 gene 28 
Ec-C1984 CC-05 E. coli clinical isolate harbouring Ser83Leu and Asp87Asn substitutions in GyrA, the Ser80Ile substitution in ParC and the aac(6′)-Ib-cr and oqxAB genes 29 
Ec-C1550 CC-06 E. coli clinical isolate harbouring Ser83Leu and Asp87Asn substitutions in GyrA, the Ser80Ile substitution in ParC and the qepA1 gene 29 
Kpn-LB4 CC-07 Klebsiella pneumoniae clinical isolate harbouring Ser83Tyr and Val112Ile substitutions in GyrA, deficient in both OmpK35 and OmpK36, and expressing active efflux of fluoroquinolones 30,31 
Ec-25922 mut1 CC-08 E. coli ATCC 25922 harbouring the Ser83Leu substitution in GyrA and the Ser80Arg substitution in ParC (genetic replacement) 19 
Ec-25922 mut2 CC-09 E. coli ATCC 25922 harbouring the Ser83Leu substitution in GyrA, the Ser80Arg substitution in ParC (genetic replacement) and marR deficient (chromosomal inactivation) 19 
Ec-25922 mut3 CC-10 E. coli ATCC 25922 harbouring Ser83Leu and Asp87Asn substitutions in GyrA and the Ser80Arg substitution in ParC (genetic replacement) 19 
Ec-25922 mut4 CC-11 E. coli ATCC 25922 harbouring Ser83Leu and Asp87Asn substitutions in GyrA, the Ser80Arg substitution in ParC (genetic replacement) and marR deficient (chromosomal inactivation) 19 
Ec-38–27 CC-12 E. coli clinical isolate belonging to the ST131 clone harbouring the Ser83Leu substitution in GyrA 32 

Study design

The work was designed as a nationwide proficiency study. In March 2013, the 13 selected strains (Table 1) were sent to the 66 participating laboratories, together with detailed instructions (62 centres reported results). Participating centres were requested to implement the methods they routinely used to test antimicrobial susceptibility and to consider the 13 strains as blood culture isolates. For each strain, participants were requested to fill in an electronic form that included: (i) the results of the antibiogram [quantitatively, in terms of inhibition zone diameters and MIC values, and qualitatively, in terms of the derived ‘raw’ clinical categories (RCCs): susceptible (S), intermediate (I) and resistant (R)]; (ii) the breakpoints used (CLSI or EUCAST); (iii) the laboratory methods used (i.e. type of automatic system or manual instrumentation); (iv) the interpreted clinical categories (ICCs), when appropriate, according to internal laboratory criteria (also used in clinical practice); and (v) the inferred mechanism(s) that might be responsible for the quinolone resistance phenotype.

Data analysis

Analysis of results focused on three aspects: (i) a descriptive analysis of susceptibility testing methods, breakpoints applied, clinical categories (CCs) assigned and discrepancies between centres deriving from these; (ii) an analysis of errors detected in susceptibility testing results [minor errors (mEs), major errors (MEs) and very major (VMEs), defined according to standard criteria] compared with reference values;12 and (iii) an analysis of the ability of participating laboratories to accurately infer the underlying resistance mechanisms.

Results

Descriptive analysis of susceptibility testing methods, breakpoints applied, CCs assigned and derived discrepancies

Ninety percent of participating laboratories used automatic or semi-automatic devices for routine susceptibility testing. These included the MicroScan WalkAway (Dade MicroScan Inc., West Sacramento, CA, USA; n = 30, 48%), Vitek 2 (bioMérieux; n = 19, 30%), the Wider (Francisco Soria Melguizo, Madrid, Spain; n = 5, 8%) and Phoenix (BD Biosciences, Sparks, MD, USA; n = 2, 3%) systems. The remaining 10% used manual methods. Nevertheless, several automatic system users also performed complementary manual tests (up to 15%) to check certain borderline susceptibility values (mainly for CC-03, CC-07 and CC-09 strains). CLSI breakpoints were applied for 71% of determinations and EUCAST breakpoints for the remaining 29%.

The overall (all antibiotics/strain combinations) raw discrepancy rate due exclusively to the differential use of breakpoints was 21.8% (Table 2). Ofloxacin and norfloxacin were often involved in this kind of discrepancy, with particularly high percentages (38.4%–46.1%); CC-03 and CC-10 strains [qnrS2+aac(6′)-Ib-cr and GyrA Ser83Leu + GyrA Asp87Asn + ParC Ser80Arg + ΔmarR, respectively] showed the highest discrepancy rates (50%). High breakpoint-related discrepancies were also found for CC-07, CC-08, CC-09 and CC-11 (33.3%), all of which had chromosomally mediated mechanisms. On the other hand, CC-00, CC-04, CC-05 and CC-06 showed no breakpoint-related discrepancies (Table 2). When the RCC was modified to ICC, the discrepancy rate due exclusively to the differential use of breakpoints decreased considerably when some in the susceptible category were re-coded as I(r) (interpreted as intermediate susceptibility due to a less susceptible phenotype, e.g. with MICs ≥0.125 mg/L for ciprofloxacin; Table 2).

Table 2.

Reference centres' susceptibility testing results, derived RCCs and ICCs based on resistance mechanisms

Strain (characteristics) Ciprofloxacin
 
Levofloxacin
 
Moxifloxacin
 
Ofloxacin
 
Norfloxacin
 
Nalidixic acid
 
MICa (mg/L) RCCb CLSI/EUCAST ICCc MICa (mg/L) RCCb CLSI/EUCAST ICCc MICa (mg/L) RCCb CLSI/EUCAST ICCc MICa (mg/L) RCCb CLSI/EUCAST ICCc MICa (mg/L) RCCb CLSI/EUCAST ICCc MICa (mg/L) RCCb CLSI/EUCAST ICCc 
CC-00 (WT) 0.004–0.008 0.005–0.015 0.008 NS/S NS/S 0.016–0.032 0.032 1–2 S/NS S/NS 
CC-01 [GyrA 1, aac(6′)-Ib-cr1-2 I(r) NS/I NS/I 2–4 S/R I(r)/R 128 128 R/NS R/NS 
CC-02 (qnrA10.125–0.25 I(r) 0.5 I(r) NS/I NS/I S/I I(r)/I 0.5–1 I(r) 16–32 S/NS S/NS 
CC-03 [qnrS2, aac(6′)-Ib-crI/R I/R 0.5 I(r) NS/R NS/R 1–2 S/I I(r)/I S/R I(r)/R 16 S/NS S/NS 
CC-04 (qnrB480.125 I(r) 0.25 I(r) NS/I NS/I 0.5–1 I(r) 0.25 I(r) 4–8 S/NS S/NS 
CC-05 [GyrA 2, ParC 1, aac(6′)-Ib-cr, oqxAB64 8–16 16–32 NS/R NS/R 32 256 >256 R/NS R/NS 
CC-06 (GyrA 2, ParC 1, qepA1>256 32 64 NS/R NS/R 64 >256 >256 R/NS R/NS 
CC-07 (GyrA 1, OmpK35, efflux) S/I I(r)/I 0.5 I(r) 0.5 NS/S NS/I(r) 4–8 4–8 S/R I(r)/R >256 R/NS R/NS 
CC-08 (GyrA 1, ParC 1) 0.25 I(r) 0.25 I(r) 0.25–0.5 NS/S NS/I(r) S/I I(r)/I S/R I(r)/R >256 R/NS R/NS 
CC-09 (GyrA 1, ParC 1, MarR0.5 I(r) 0.5 I(r) 0.5 NS/S NS/I(r) 1–2 S/I I(r)/I 2–4 S/R I(r)/R >256 R/NS R/NS 
CC-10 (GyrA 2, ParC 1) S/I I(r)/I 1–2 I(r) NS/R NS/R 2–4 I/R I/R S/R I(r)/R >256 R/NS R/NS 
CC-11 (GyrA 2, ParC 1, MarRI/R I/R NS/R NS/R 8–16 I/R I/R >256 R/NS R/NS 
CC-12 (GyrA 1) 0.25 I(r) 0.25 I(r) 0.25 NS/S NS/I(r) S/I I(r)/I 0.5 I(r) >256 R/NS R/NS 
Strain (characteristics) Ciprofloxacin
 
Levofloxacin
 
Moxifloxacin
 
Ofloxacin
 
Norfloxacin
 
Nalidixic acid
 
MICa (mg/L) RCCb CLSI/EUCAST ICCc MICa (mg/L) RCCb CLSI/EUCAST ICCc MICa (mg/L) RCCb CLSI/EUCAST ICCc MICa (mg/L) RCCb CLSI/EUCAST ICCc MICa (mg/L) RCCb CLSI/EUCAST ICCc MICa (mg/L) RCCb CLSI/EUCAST ICCc 
CC-00 (WT) 0.004–0.008 0.005–0.015 0.008 NS/S NS/S 0.016–0.032 0.032 1–2 S/NS S/NS 
CC-01 [GyrA 1, aac(6′)-Ib-cr1-2 I(r) NS/I NS/I 2–4 S/R I(r)/R 128 128 R/NS R/NS 
CC-02 (qnrA10.125–0.25 I(r) 0.5 I(r) NS/I NS/I S/I I(r)/I 0.5–1 I(r) 16–32 S/NS S/NS 
CC-03 [qnrS2, aac(6′)-Ib-crI/R I/R 0.5 I(r) NS/R NS/R 1–2 S/I I(r)/I S/R I(r)/R 16 S/NS S/NS 
CC-04 (qnrB480.125 I(r) 0.25 I(r) NS/I NS/I 0.5–1 I(r) 0.25 I(r) 4–8 S/NS S/NS 
CC-05 [GyrA 2, ParC 1, aac(6′)-Ib-cr, oqxAB64 8–16 16–32 NS/R NS/R 32 256 >256 R/NS R/NS 
CC-06 (GyrA 2, ParC 1, qepA1>256 32 64 NS/R NS/R 64 >256 >256 R/NS R/NS 
CC-07 (GyrA 1, OmpK35, efflux) S/I I(r)/I 0.5 I(r) 0.5 NS/S NS/I(r) 4–8 4–8 S/R I(r)/R >256 R/NS R/NS 
CC-08 (GyrA 1, ParC 1) 0.25 I(r) 0.25 I(r) 0.25–0.5 NS/S NS/I(r) S/I I(r)/I S/R I(r)/R >256 R/NS R/NS 
CC-09 (GyrA 1, ParC 1, MarR0.5 I(r) 0.5 I(r) 0.5 NS/S NS/I(r) 1–2 S/I I(r)/I 2–4 S/R I(r)/R >256 R/NS R/NS 
CC-10 (GyrA 2, ParC 1) S/I I(r)/I 1–2 I(r) NS/R NS/R 2–4 I/R I/R S/R I(r)/R >256 R/NS R/NS 
CC-11 (GyrA 2, ParC 1, MarRI/R I/R NS/R NS/R 8–16 I/R I/R >256 R/NS R/NS 
CC-12 (GyrA 1) 0.25 I(r) 0.25 I(r) 0.25 NS/S NS/I(r) S/I I(r)/I 0.5 I(r) >256 R/NS R/NS 

NS, not specified by CLSI or EUCAST committees; I(r), interpreted as intermediate susceptibility due to a less susceptible phenotype (MIC ≥0.125 mg/L for ciprofloxacin, 0.25 mg/L for levofloxacin, moxifloxacin and norfloxacin, and 0.5 mg/L for ofloxacin).

GyrA 1 and ParC 1 mean one substitution in the QRDR and GyrA 2 means two substitutions in the QRDR (see Table 1).

aDetermined in duplicate by the two reference centres using the CLSI broth microdilution method.

bAccording to CLSI/EUCAST criteria.

cInterpreted by the reference centres according to pre-characterized resistance mechanisms and interpretive reading of the antibiogram.

ICCs provided by participating centres and the discrepancies between them are shown in Figure 1. There was a high degree of consensus among centres for WT strains (CC-00) and for strains with high-level resistance patterns, such as CC-05, CC-06 and CC-11 [harbouring triple quinolone resistance-determining region (QRDR) modification + aac(6′)-Ib-cr+oqxAB, triple QRDR modification + qepA1 and triple QRDR modification + ΔmarR, respectively]. On the other hand, there was less consensus about some antimicrobial/strain combinations, such as ciprofloxacin and CC-03, CC-07 and CC-09 strains [harbouring qnrS2 + aac(6′)-Ib-cr, GyrA Ser83Leu + OmpK35 + efflux and double QRDR modification + ΔmarR, respectively, with MICs in the range of 0.5–2 mg/L], and levofloxacin and CC-01, CC-03, CC-07, CC-09 and CC-10 strains (MICs in the range of 0.5–2 mg/L).

Figure 1.

Overall ICC percentages provided by participating centres regarding quinolone antibiotics; ‘r’ values were re-coded as ‘I’. CIP, ciprofloxacin; LVX, levofloxacin; MXF, moxifloxacin; NAL, nalidixic acid; NOR, norfloxacin; OFX, ofloxacin.

Figure 1.

Overall ICC percentages provided by participating centres regarding quinolone antibiotics; ‘r’ values were re-coded as ‘I’. CIP, ciprofloxacin; LVX, levofloxacin; MXF, moxifloxacin; NAL, nalidixic acid; NOR, norfloxacin; OFX, ofloxacin.

Analysis of errors (mEs, MEs and VMEs) in susceptibility testing results

Quinolone MICs, RCCs obtained after applying CLSI and EUCAST criteria and consensus ICCs obtained at the two reference centres are shown in Table 2. The distributions of discrepancies and categorical error rates by antibiotic and strain are shown in Tables 3 and 4, respectively. The percentage of VMEs (false susceptibility) for all quinolones was well below 2%, and only ofloxacin and moxifloxacin showed higher rates for raw VMEs (6.6%), which reduced to 0% and 2.9%, respectively, in the interpreted VMEs (Table 3) (these VMEs were associated with manual methods: disc diffusion for ofloxacin and agar dilution for moxifloxacin; data not shown). These errors were particularly associated with the CC-03 [qnrS2 + aac(6′)-Ib-cr] strain and were almost exclusively observed among MicroScan WalkAway automated systems users (50% of total raw VMEs with respect to this strain) (Table 4). The percentages of MEs (false resistance) were always <10%, except for the raw MEs for ofloxacin and nalidixic acid (22.2% and 19.1%, respectively) and the interpreted MEs for nalidixic acid (27.1%). Eighty-five percent of raw MEs for ofloxacin occurred in the CC-01 and CC-02 strains. The high percentage of mEs observed was not surprising, given the MIC values exhibited by several strains in this collection, which, in many cases, were on the borderline between two CCs (Tables 2 and 3).

Table 3.

Distribution of discrepancies and categorical error rates by antimicrobial agent tested

Quinolone (nPercentage of discrepancies in RCC/ICCb,c Percentage of errors in RCC/ICC (n)a
 
all methods and devices (2305)
 
MicroScan WalkAway (919)
 
Wider (162)
 
Vitek 2 (520)
 
mEc MEd VMEe mEc MEd VMEc mEc MEd VMEe mEc MEd VMEe 
CIP (805) 19.1/42.0 14.4/41.7 7.4/0 0.3/0.6 17.9/43.1 12.4/0 0.7/1.4 8.3/48.5 4.3/0 0/0 7.5/38.5 1.9/0 0/0 
LVX (243) 13.5/53.3 9.8/53.3 4.7/0 0/0 3.7/68.5 0/0 0/0 0/0 0/0 0/0 0/21.4 18.1/0 0/0 
MXF (77) 25.9/36.7 22.1/35.7 3.4/0 6.6/2.9 15.4/28 0/0 0/0 0/35.3 0/0 0/0 0/0 0/0 0/0 
OFX (354) 28.5/37.8 21.4/37.8 22.2/0 6.6/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 
NOR (42) 27.4/43.1 20/43.1 8.8/0 1.8/0 23/46.3 9.2/0 1.2/0 31.5/57.8 18.2/0 0/0 0/0 0/0 0/0 
All FQs (1521) 20.7/43.2 15.6/43.0 7.3/0 1.4/0.1 17.7/45.6 8.3/0 0.6/0.4 13.2/48.1 7.4/0 0/0 7.1/37.6 3.0/0 0/0 
NAL (784) 7/10.8 1.1/2.2 19.1/27.1 0/0.4 0/0.6 9.2/20.4 0/0.8 0/1.4 4.3/4.3 0/0 0/1.9 34.2/40.7 0/0 
All agents (2305) 16.1/31.5f 10.6/28.3f 9.6/18.1 0.7/0.5 11.3/29 9.7/13.8 0.4/1 7.4/29.3 6.5/2.8 0/0 3.6/19.8 12.8/26.7 0/0 
Quinolone (nPercentage of discrepancies in RCC/ICCb,c Percentage of errors in RCC/ICC (n)a
 
all methods and devices (2305)
 
MicroScan WalkAway (919)
 
Wider (162)
 
Vitek 2 (520)
 
mEc MEd VMEe mEc MEd VMEc mEc MEd VMEe mEc MEd VMEe 
CIP (805) 19.1/42.0 14.4/41.7 7.4/0 0.3/0.6 17.9/43.1 12.4/0 0.7/1.4 8.3/48.5 4.3/0 0/0 7.5/38.5 1.9/0 0/0 
LVX (243) 13.5/53.3 9.8/53.3 4.7/0 0/0 3.7/68.5 0/0 0/0 0/0 0/0 0/0 0/21.4 18.1/0 0/0 
MXF (77) 25.9/36.7 22.1/35.7 3.4/0 6.6/2.9 15.4/28 0/0 0/0 0/35.3 0/0 0/0 0/0 0/0 0/0 
OFX (354) 28.5/37.8 21.4/37.8 22.2/0 6.6/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 
NOR (42) 27.4/43.1 20/43.1 8.8/0 1.8/0 23/46.3 9.2/0 1.2/0 31.5/57.8 18.2/0 0/0 0/0 0/0 0/0 
All FQs (1521) 20.7/43.2 15.6/43.0 7.3/0 1.4/0.1 17.7/45.6 8.3/0 0.6/0.4 13.2/48.1 7.4/0 0/0 7.1/37.6 3.0/0 0/0 
NAL (784) 7/10.8 1.1/2.2 19.1/27.1 0/0.4 0/0.6 9.2/20.4 0/0.8 0/1.4 4.3/4.3 0/0 0/1.9 34.2/40.7 0/0 
All agents (2305) 16.1/31.5f 10.6/28.3f 9.6/18.1 0.7/0.5 11.3/29 9.7/13.8 0.4/1 7.4/29.3 6.5/2.8 0/0 3.6/19.8 12.8/26.7 0/0 

CIP, ciprofloxacin; LVX, levofloxacin; MXF, moxifloxacin; OFX, ofloxacin; NOR, norfloxacin; FQs, fluoroquinolones; NAL, nalidixic acid.

aRCC and ICC results from participating laboratories were compared with reference centre values and discrepancies were classified as mE, ME and VME, following standard criteria.33

bThe most favourable CC for the evaluated centre was chosen when the MIC value (obtained by the reference centres) gave rise to two different CCs according to EUCAST/CLSI breakpoints. In this case, the breakpoints chosen were those most in concordance with the CC issued by the evaluated centre.

cThe denominator is the number of susceptibility-testing determinations per antibiotic.

dThe denominator is the number of susceptible strains per antibiotic.

eThe denominator is the number of resistant strains per antibiotic.

fThe denominator is the total number of susceptibility determinations.

Table 4.

Distribution of categorical error rates by tested strain

Strain (characteristics) (nPercentage of errors in RCC/ICC (n)a,b
 
all methods and devices (2305)
 
MicroScan WalkAway (919)
 
Wider (162)
 
Vitek 2 (520)
 
mEc MEd VMEe mEc MEd VMEe mEc MEd VMEe mEc MEd VMEe 
CC-00 (WT) (175) 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 
CC-01 [GyrA 1, aac(6′)-Ib-cr] (178) 3.9/10.5 4.3/0 0/0 0/8.4 0/0 0/0 0/20 0/0 0/0 2.5/2.5 0/0 0/0 
CC-02 (qnrA1) (183) 11.4/47.9 21.9/57.6 0/0 11.6/52.2 16.4/43.4 0/0 8.3/53.3 0/0 0/0 0/31.7 48.7/48.7 0/0 
CC-03 [qnrS2, aac(6′)-Ib-cr] (181) 24.3/36.4 13.8/42.6 5.8/0 36.7/44.1 0/41.6 8.3/0 14.2/18.7 12.5/0 0/0 5.1/26.3 30/47.3 0/0 
CC-04 (qnrB48) (174) 5.1/60.9 1.2/8.2 0/0 0/60.8 0/0 0/0 0/50 7.7/16.6 0/0 0/52.6 0/11.1 0/0 
CC-05 [GyrA 2, ParC 1, aac(6′)-Ib-cr, oqxAB] (176) 0.5/0.6 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 
CC-06 (GyrA 2, ParC 1, qepA1) (177) 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 
CC-07 (GyrA 1, OmpK35, efflux) (175) 20/35 22.6/0 1.1/0 29.2/35.6 35.3/0 0/0 7.6/50 50/0 0/0 5/23.7 0/0 0/0 
CC-08 (GyrA 1, ParC 1) (181) 8.8/40.2 0/0 3.3/0 10.9/37.1 0/0 2.6/0 0/58.3 0/0 0/0 2.5/31.5 0/0 0/0 
CC-09 (GyrA 1, ParC 1, MarR) (178) 23/35.1 14/0 1.1/0 25/30.8 17.7/0 0/0 25/33.3 28/0 0/0 7.7/24.3 0/0 0/0 
CC-10 (GyrA 2, ParC 1) (176) 23.8/41.2 20.4/0 1/0 19.1/50.7 12/0 0/0 25/41.6 0/0 0/0 16.6/30.7 10.1/0 0/0 
CC-11 (GyrA 2, ParC 1, MarR) (175) 14.8/13.4 0/0 0/0.7 13.8/13.8 0/0 0/1.7 16.6/16.6 0/0 0/0 7.3/4.8 0/0 0/0 
CC-12 (GyrA 1) (176) 2.2/45.7 0.9/0 0/1.6 2.7/47.2 2.1/0 0/3.8 0/54.5 0/0 0/0 0/32.5 0/0 0/0 
Strain (characteristics) (nPercentage of errors in RCC/ICC (n)a,b
 
all methods and devices (2305)
 
MicroScan WalkAway (919)
 
Wider (162)
 
Vitek 2 (520)
 
mEc MEd VMEe mEc MEd VMEe mEc MEd VMEe mEc MEd VMEe 
CC-00 (WT) (175) 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 
CC-01 [GyrA 1, aac(6′)-Ib-cr] (178) 3.9/10.5 4.3/0 0/0 0/8.4 0/0 0/0 0/20 0/0 0/0 2.5/2.5 0/0 0/0 
CC-02 (qnrA1) (183) 11.4/47.9 21.9/57.6 0/0 11.6/52.2 16.4/43.4 0/0 8.3/53.3 0/0 0/0 0/31.7 48.7/48.7 0/0 
CC-03 [qnrS2, aac(6′)-Ib-cr] (181) 24.3/36.4 13.8/42.6 5.8/0 36.7/44.1 0/41.6 8.3/0 14.2/18.7 12.5/0 0/0 5.1/26.3 30/47.3 0/0 
CC-04 (qnrB48) (174) 5.1/60.9 1.2/8.2 0/0 0/60.8 0/0 0/0 0/50 7.7/16.6 0/0 0/52.6 0/11.1 0/0 
CC-05 [GyrA 2, ParC 1, aac(6′)-Ib-cr, oqxAB] (176) 0.5/0.6 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 
CC-06 (GyrA 2, ParC 1, qepA1) (177) 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0 
CC-07 (GyrA 1, OmpK35, efflux) (175) 20/35 22.6/0 1.1/0 29.2/35.6 35.3/0 0/0 7.6/50 50/0 0/0 5/23.7 0/0 0/0 
CC-08 (GyrA 1, ParC 1) (181) 8.8/40.2 0/0 3.3/0 10.9/37.1 0/0 2.6/0 0/58.3 0/0 0/0 2.5/31.5 0/0 0/0 
CC-09 (GyrA 1, ParC 1, MarR) (178) 23/35.1 14/0 1.1/0 25/30.8 17.7/0 0/0 25/33.3 28/0 0/0 7.7/24.3 0/0 0/0 
CC-10 (GyrA 2, ParC 1) (176) 23.8/41.2 20.4/0 1/0 19.1/50.7 12/0 0/0 25/41.6 0/0 0/0 16.6/30.7 10.1/0 0/0 
CC-11 (GyrA 2, ParC 1, MarR) (175) 14.8/13.4 0/0 0/0.7 13.8/13.8 0/0 0/1.7 16.6/16.6 0/0 0/0 7.3/4.8 0/0 0/0 
CC-12 (GyrA 1) (176) 2.2/45.7 0.9/0 0/1.6 2.7/47.2 2.1/0 0/3.8 0/54.5 0/0 0/0 0/32.5 0/0 0/0 

GyrA 1 and ParC 1 mean one substitution in the QRDR and GyrA 2 means two substitutions in the QRDR (see Table 1).

aRCC and ICC results from participating laboratories were compared with reference centre values and discrepancies were classified as mE, ME and VME, following standard criteria.33

bThe most favourable CC of the evaluating centre was chosen when an MIC value (obtained by reference centres) gave rise to two different CCs according to EUCAST/CLSI breakpoints. In this case, the breakpoints chosen were those showing most concordance with the CC issued by the evaluating centre.

cThe denominator is the number of susceptibility-testing determinations for each strain.

dThe denominator is the number of susceptibility-testing determinations for susceptible strains.

eThe denominator is the number of susceptibility-testing determinations for resistant strains.

The analysis of categorical errors according to strain (Table 4) revealed that those with the highest rates of raw VMEs were CC-03 [harbouring qnrS2 and aac(6′)-Ib-cr, 5.8%] and CC-08 (single GyrA and ParC mutation, 3.3%). These percentages dropped to 0% when interpreted categories were applied. VMEs in the CC-03 strain were linked to ciprofloxacin and moxifloxacin and those in CC-08 to norfloxacin. CC-03 was also a strain that gave a high rate of raw MEs (13.8%) linked with nalidixic acid (73% of these). Highest rates of raw MEs were observed for CC-02 strains (qnrA1) (21.9%) in association with nalidixic acid; CC-07 (single GyrA mutation with loss of porins and efflux) (22.6%) linked to norfloxacin; and CC-10 (double GyrA and single ParC mutation) (20.4%), linked to ciprofloxacin (71% of these) and levofloxacin (18%). A significant increase in interpreted ME values was observed for CC-02 and CC-03 strains (57.6% and 42.6%, respectively), due to the fact that some raw susceptible strains were modified to the intermediate category using ICC criteria, which obviously reduced the number of strains finally defined as susceptible. Additionally, interpretation of these results should take into account the fact that some of the quinolones (e.g. ciprofloxacin and nalidixic acid) were tested by most of the centres while others (e.g. moxifloxacin and norfloxacin) were tested only by a lower number of centres.

VMEs were not found for the Vitek 2 and Wider systems compared with the MicroScan WalkAway (0.4% and 1% for raw and interpreted data, respectively). Among all determinations performed using the MicroScan WalkAway, two raw VMEs and four interpreted VMEs were found (affecting ciprofloxacin and norfloxacin in the CC-03, CC-07 and CC-08 strains). The Wider system accounted for the lowest percentage of raw MEs overall, followed by the MicroScan WalkAway and Vitek 2 systems; after interpretation, the percentage of MEs was still in the same order. The fluoroquinolone giving the highest percentage of raw MEs was norfloxacin among Wider users, levofloxacin among Vitek 2 users and ciprofloxacin among MicroScan WalkAway users. The Wider system showed the lowest rate of raw mEs, followed by the MicroScan WalkAway and the Vitek 2.

Ability of centres to detect acquired PMQRs and to infer resistance mechanisms

An analysis of errors of inference is shown in Table 5, in which strains CC-01, CC-02, CC-03 and CC-04 (all with LLQR and harbouring PMQR mechanisms) are highlighted as those giving the highest percentage of errors, documented in 11.3%–54.8% of centres. One of the most important aspects was that acquired PMQR genes were not detected in certain strains. In such cases, laboratories attributed resistance exclusively to mutational mechanisms, with percentages of >40% in the strains specified. By contrast, PMQR genes were inferred in strains harbouring only chromosomal mechanisms (CC-07, CC-08, CC-09, CC-10, CC-11 and CC-12).

Table 5.

Description of errors in the inferred resistance mechanisms

Strain (characteristics) Wrong inferred mechanisms
 
no. (%) description of mistakes 
CC-00 (WT) 0 (0) no mistakes 
CC-01 [GyrA 1, aac(6′)-Ib-cr34 (54.8) multiple QRDR modifications (12); PMQR not detected (22) 
CC-02 (qnrA115 (24.2) single QRDR modification at GyrA and PMQR not detected (15) 
CC-03 [qnrS2, aac(6′)-Ib-cr7 (11.3) single QRDR modification at GyrA and PMQR not detected (4); chromosomal efflux pump phenotype and PMQR not detected (3) 
CC-04 (qnrB4810 (16.1) single QRDR modification at GyrA and PMQR not detected (6); chromosomal efflux pump phenotype and PMQR not detected (4) 
CC-05 [GyrA 2, ParC 1, aac(6′)-Ib-cr, oqxAB0 (0) no mistakes 
CC-06 (GyrA 2, ParC 1, qepA10 (0) no mistakes 
CC-07 (GyrA 1, OmpK35, efflux) 6 (9.6) PMQR (6) 
CC-08 (GyrA 1, ParC 1) 4 (6.5) PMQR (4) 
CC-09 (GyrA 1, ParC 1, MarR5 (8) PMQR (5) 
CC-10 (GyrA 2, ParC 1) 3 (4.8) PMQR (3) 
CC-11 (GyrA 2, ParC 1, MarR2 (3.2) PMQR (2) 
CC-12 (GyrA 1) 1 (1.6) PMQR (1) 
Strain (characteristics) Wrong inferred mechanisms
 
no. (%) description of mistakes 
CC-00 (WT) 0 (0) no mistakes 
CC-01 [GyrA 1, aac(6′)-Ib-cr34 (54.8) multiple QRDR modifications (12); PMQR not detected (22) 
CC-02 (qnrA115 (24.2) single QRDR modification at GyrA and PMQR not detected (15) 
CC-03 [qnrS2, aac(6′)-Ib-cr7 (11.3) single QRDR modification at GyrA and PMQR not detected (4); chromosomal efflux pump phenotype and PMQR not detected (3) 
CC-04 (qnrB4810 (16.1) single QRDR modification at GyrA and PMQR not detected (6); chromosomal efflux pump phenotype and PMQR not detected (4) 
CC-05 [GyrA 2, ParC 1, aac(6′)-Ib-cr, oqxAB0 (0) no mistakes 
CC-06 (GyrA 2, ParC 1, qepA10 (0) no mistakes 
CC-07 (GyrA 1, OmpK35, efflux) 6 (9.6) PMQR (6) 
CC-08 (GyrA 1, ParC 1) 4 (6.5) PMQR (4) 
CC-09 (GyrA 1, ParC 1, MarR5 (8) PMQR (5) 
CC-10 (GyrA 2, ParC 1) 3 (4.8) PMQR (3) 
CC-11 (GyrA 2, ParC 1, MarR2 (3.2) PMQR (2) 
CC-12 (GyrA 1) 1 (1.6) PMQR (1) 

Discussion

The growing prevalence of quinolone resistance in Enterobacteriaceae severely compromises the therapeutic arsenal available.1,13,14 The complexity of the mechanisms involved also affects the ability of the clinical microbiology laboratory to perform proper standardized antimicrobial susceptibility tests. This situation may cause greater concern with borderline resistance phenotypes. The detection of specific resistance mechanisms is particularly important for guiding therapy and infection control strategies.8,9 It was this clinical challenge that motivated this large multicentre study, in which we evaluated the proficiency of 62 Spanish laboratories with respect to accurate susceptibility testing and detection of quinolone resistance phenotypes in a collection of well-characterized Enterobacteriaceae strains showing different combinations of mutational and transferable quinolone resistance mechanisms.

Of the various aspects that had a significant impact on discrepancies in the susceptibility testing results provided by participating centres, the differential application of the currently non-harmonized CLSI or EUCAST breakpoints should be mentioned. EUCAST breakpoints were used in only 29% of the participating centres, although this represents a significant increase when compared with similar Spanish studies performed in 2001 and 2011, in which 100% and 86% of the centres, respectively, applied CLSI breakpoints.8,9 Furthermore, this value is expected to increase significantly in the near future once the current institutional recommendations in Europe are followed. The application of European breakpoints has increased significantly in Spain, due particularly to the 2012 constitution of the Spanish Antibiogram Committee (Coesant), whose main commitment is to endorse EUCAST breakpoints.15

Regarding the discrepancies that derived from breakpoints, ofloxacin and norfloxacin were the antibiotics most affected. These discrepancies may be particularly important in PMQR-producing strains (such as CC-03) or those with combinations of two or three mutational modifications (such as CC-07, CC-08, CC-09 and CC-10), since several strains that showed mutational resistance mechanisms or relevant acquired PMQR genes, such as qnrS2 and aac(6′)-Ib-cr, were reported as susceptible according to CLSI breakpoints but resistant according to EUCAST. Less relevant are discrepancies deriving from different ciprofloxacin or levofloxacin breakpoints, where discrepancies were mostly reported as intermediate according to CLSI breakpoints and resistant according to EUCAST breakpoints.10,11

The method used to carry out susceptibility testing was another source of discrepancy in the results between centres. Most centres (90%) used semi-automatic or automatic devices, mainly the MicroScan WalkAway, Vitek 2 and Wider systems. Vitek 2 and Wider devices did not show VMEs, although they were used considerably less than the MicroScan WalkAway. Regarding MEs, both raw and after interpretation, the lowest percentage overall corresponded to the Wider, followed by the MicroScan WalkAway and then the Vitek 2 systems. Among the antibiotics tested, moxifloxacin and ofloxacin had the highest recorded VME rates. Among the studied strains, CC-03 [a combination of qnrS2 plus aac(6′)-Ib-cr] and CC-08 (single mutation in GyrA and ParC) were those associated with a higher rate of VMEs overall. These data indicate that borderline phenotypes may imply a source of discrepancy for susceptibility testing in clinical microbiology and to have consequences for treatment outcome.16,17

A further source of divergent results concerns the interpretive reading of susceptibility testing data, and hence the reporting of ICCs. This ICC parameter is important because the in vivo activity of quinolones in the clinical setting is related to MIC values and AUC/MIC or Cmax/MIC, which are good indicators of a favourable outcome for quinolone treatment.18 Consequently, it may be assumed from these indicators that an increase in the MIC of a quinolone (of even just a few dilution steps) will have a negative impact on its therapeutic efficacy. In our study, 15% of all susceptibility determinations involved a change in the final CC, with percentages of >40% for some antibiotic and strain combinations. A high percentage of discrepancies in ICCs involved a change to intermediate susceptibility [I(r)] when interpreted at the reference centres; e.g. an MIC of ≥0.125 mg/L for ciprofloxacin was modified to intermediate susceptibility, depending on the associated resistant molecular mechanism (although it should be noted that PMQR mechanisms could be present in isolates with lower MICs values, as reported).19 We suggest that the current breakpoints defined by EUCAST/CLSI be re-evaluated, and that the already agreed general principles for the clinical categorization of antibiogram results for Salmonella could also be applied to other enterobacteria when considering severe infections.10,11

Various well-known factors that limit the efficiency of interpretive readings include a lack of knowledge about the basis of resistance in a significant proportion of medical microbiologists; unknown factors affecting resistance in clinical strains; low-level resistance mechanisms; and certain mechanisms that may be camouflaged as a result of the complex interplay of intrinsic and acquired resistance mechanisms. For example, it is not possible to infer PMQR mechanisms directly when they are associated with QRDR modifications that confer a high level of nalidixic acid resistance; meanwhile the PMQR phenotype is generally characterized by reduced susceptibility to fluoroquinolones, with susceptibility to nalidixic acid in the absence of QRDR modifications.20–22 The implementation of expert systems in most automatic/semi-automatic systems may help in this task.23,24 Recommendations for interpretive reading, inferring and detecting resistance mechanisms have never been issued by CLSI or EUCAST, despite the fact that the breakpoints are certainly not adapted to overcoming PMQR mechanisms. As observed in our study, a combination of one to three relevant mechanisms of mutational resistance or PMQR determinants frequently yields a borderline MIC. This, therefore, is a major source of conflicting results in the absence of an interpretive reading of the antibiogram. The pharmacokinetic parameter for the clinical efficacy of quinolones is the AUC/MIC. Small variations in MIC values may decrease this parameter and affect the treatment outcome. In this way, these low-level resistant phenotypes (breakpoint borderlines) could have clinical relevance leading to in vivo resistance. In particular, it has been reported that PMQR mechanisms can reduce fluoroquinolone activity in vivo.25,26 In this sense, the epidemiological cut-off for ciprofloxacin in E. coli, established as 0.032 mg/L according to EUCAST (www.eucast.org), seems to be useful for LLQR detection.

Taken together, the data obtained in this work clearly suggest that the use of different breakpoints and systems, as well as the complexity of the mechanisms that lead to reduced susceptibility and produce borderline phenotypes, may have important consequences for the treatment and control of infections caused by these microorganisms.

Funding

This work was supported by the Ministerio de Sanidad y Consumo, Instituto de Salud Carlos III (projects PI11-00934 and PI11/01117) and the Consejería de Innovación Ciencia y Empresa, Junta de Andalucía (P11-CTS-7730), Spain, by the Plan Nacional de I+D+i 2008-2011 and the Instituto de Salud Carlos III, Subdirección General de Redes y Centros de Investigación Cooperativa, Ministerio de Economía y Competitividad, the Spanish Network for Research in Infectious Diseases (REIPI RD12/0015)—co-financed by European Development Regional Fund ‘A way to achieve Europe’ ERDF.

Transparency declarations

None to declare.

Acknowledgements

This study was performed under the auspices of the Study Group of Antimicrobial Mechanisms of Action and Resistance (GEMARA) from the Spanish Society of Clinical Microbiology and Infectious Diseases (SEIMC) and the Spanish Network for Research in Infectious Diseases (REIPI). We are grateful to the 62 participating centres and the SEIMC Quality Control Program for their essential contribution to making this study possible. The complete list of the 62 participating hospitals is as follows: Hospital Universitario Virgen de la Macarena (Sevilla), Hospital Universitario de Valme (Sevilla), Hospital Universitario de Tarragona Joan XXIII (Tarragona), Hospital General de Gran Canaria Dr Negrín (Las Palmas Gran Canaria, Las Palmas), Hospital Costa del Sol (Marbella, Málaga), Hospital Universitario Miguel Servet (Zaragoza), Hospital Sierrallana (Torrelavega, Cantabria), Hospital Provincial de Castellón (Castellón), Hospital Santa Barbara (Complejo Hosp. de Soria, Soria), Hospital Clínico Universitario de Salamanca (Salamanca), Hospital Clínico Universitario de Valladolid (Valladolid), Hospital Universitario Rio Hortega (Valladolid), Complejo Asistencial de Avila (Ávila), Hospital General Universitario de Albacete (Albacete), Hospital General de Ciudad Real (Ciudad Real), Hospital Virgen de la Salud (Toledo), Hospital Universitario de Girona Dr Josep Trueta (Girona), Hospital Clínic (Barcelona), Hospital Virgen de Puerto (Plasencia, Cáceres), Complejo Hospitalario Universitario de Vigo (Vigo, Pontevedra), Complejo Hospitalario Universitario A Coruña (A Coruña), Hospital Infantil Niño Jesús (Madrid), Hospital General U. Gregorio Marañón (Madrid), Hospital Universitario Puerta de Hierro Majadahonda (Majadahonda, Madrid), Hospital Carlos III (Madrid), Complejo Hospitalario de Navarra (Pamplona, Navarra), Hospital de Txagorritxu (Vitoria, Álava), Complejo Hospitalario Donostia (Donosti-San Sebastián, Gipuzkoa), Hospital de Cruces (Barakaldo, Vizcaya), Hospital de Galdakao (Galdakao Vizcaya), Consorcio Hospital General Universitario de Valencia (Valencia), Hospital Universitario La Fe (Valencia), Hospital General Universitario de Elche (Elche, Alicante), Complejo Asistencial Universitario de Burgos (Burgos), Hospital Universitario Puerta del Mar (Cádiz), Hospital Nª Sra de la Candelaria (Santa Cruz de Tenerife, Tenerife), Hospital Severo Ochoa (Leganés, Madrid), Complejo Hospitalario de Pontevedra (Pontevedra), Hospital Universitario Reina Sofía (Córdoba), Hospital Universitario La Paz (Madrid), Hospital Universitario Son Espases (Palma de Mallorca, Baleares), Hospital Ramón y Cajal (Madrid), Hospital Universitario Insular de Gran Canaria (Las Palmas de Gran Canaria, Las Palmas), Hospital Virgen de Altagracia (Manzanares, Ciudad Real), Hospital Son Llatzer (Palma de Mallorca, Baleares), Hospital de Jerez (Jerez de la Frontera, Cádiz), Hospital de la Ribera (Alzira, Valencia), Hospital Universitari Germans Trias i Pujol (Badalona, Barcelona), Hospital San Pedro de Alcantara (Cáceres), Hospital Universitario Vall d′Hebron (Barcelona), Hospital General Reina Sofia (Murcia), Hospital Universitario Virgen del Rocío (Sevilla), Hospital Universitario San Cecilio (Granada), Hospital Universitario Santa Cristina (Madrid), Hospital de Cabueñes (Gijón, Asturias), Hospital Universitario Marqués de Valdecilla (Santander, Cantabria), Hospital Universitario de Bellvitge (L'Hospitalet de Llobregat, Barcelona), Hospital de la Princesa Madrid (Madrid), Hospital Clínico Universitario San Carlos (Madrid), Hospital Clínico Universitario Lozano Blesa (Zaragoza), Hospital Universitario Fundación Alcorcón (Alcorcón, Madrid), Hospital Universitario Virgen de la Arrixaca (El Palmar, Murcia).

References

1
Drlica
K
Hiasa
H
Kerns
R
et al
Quinolones: action and resistance updated
.
Curr Top Med Chem
 
2009
;
9
:
981
98
.
2
Singh
R
Swick
MC
Ledesma
KR
et al
Temporal interplay between efflux pumps and target mutations in development of antibiotic resistance in Escherichia coli
.
Antimicrob Agents Chemother
 
2012
;
56
:
1680
5
.
3
Strahilevitz
J
Jacoby
GA
Hooper
DC
et al
Plasmid-mediated quinolone resistance: a multifaceted threat
.
Clin Microbiol Rev
 
2009
;
22
:
664
89
.
4
Rodriguez-Martinez
JM
Cano
ME
Velasco
C
et al
Plasmid-mediated quinolone resistance: an update
.
J Infect Chemother
 
2011
;
17
:
149
82
.
5
Fraimow
H
Nahra
R
.
Resistant Gram-negative infections
.
Crit Care Clin
 
2013
;
29
:
895
921
.
6
Paterson
DL
.
Impact of antibiotic resistance in Gram-negative bacilli on empirical and definitive antibiotic therapy
.
Clin Infect Dis
 
2008
;
47
Suppl 1
:
S14
20
.
7
Woodford
N
Eastaway
AT
Ford
M
et al
Comparison of BD Phoenix, Vitek 2, and MicroScan automated systems for detection and inference of mechanisms responsible for carbapenem resistance in Enterobacteriaceae
.
J Clin Microbiol
 
2010
;
48
:
2999
3002
.
8
Canton
R
Loza
E
Del Carmen Conejo
M
et al
Quality control for β-lactam susceptibility testing with a well-defined collection of Enterobacteriaceae and Pseudomonas aeruginosa strains in Spain
.
J Clin Microbiol
 
2003
;
41
:
1912
8
.
9
Juan
C
Conejo
MC
Tormo
N
et al
Challenges for accurate susceptibility testing, detection and interpretation of β-lactam resistance phenotypes in Pseudomonas aeruginosa: results from a Spanish multicentre study
.
J Antimicrob Chemother
 
2013
;
68
:
619
30
.
10
Clinical and Laboratory Standards Institute
.
Performance Standards for Antimicrobial Susceptibility Testing: Twenty-fourth Informational Supplement M100-S24
 .
CLSI
,
Wayne, PA, USA
,
2014
.
11
EUCAST
.
Clinical Breakpoints and Epidemiological Cut-off Values
 .
2014
. .
12
Camaro-Sala
ML
Martinez-Garcia
R
Olmos-Martinez
P
et al
Validation and verification of microbiology methods
.
Enferm Infecc Microbiol Clin
 
2014
; .
13
Hooper
DC
.
Emerging mechanisms of fluoroquinolone resistance
.
Emerg Infect Dis
 
2001
;
7
:
337
41
.
14
Jacoby
GA
.
Mechanisms of resistance to quinolones
.
Clin Infect Dis
 
2005
;
41
Suppl 2
:
S120
6
.
15
Martinez-Martinez
L
Pascual
A
Canton
R
.
The Spanish antibiogram committee (COESANT), in step with EUCAST
.
Enferm Infecc Microbiol Clin
 
2013
;
31
:
639
40
.
16
Rodriguez-Martinez
JM
Briales
A
Velasco
C
et al
Discrepancies in fluoroquinolone clinical categories between the European Committee on Antimicrobial Susceptibility Testing (EUCAST) and CLSI for Escherichia coli harbouring qnr genes and mutations in gyrA and parC
.
J Antimicrob Chemother
 
2011
;
66
:
1405
7
.
17
Poirel
L
Pitout
JD
Calvo
L
et al
In vivo selection of fluoroquinolone-resistant Escherichia coli isolates expressing plasmid-mediated quinolone resistance and expanded-spectrum β-lactamase
.
Antimicrob Agents Chemother
 
2006
;
50
:
1525
7
.
18
Schentag
JJ
Meagher
AK
Forrest
A
.
Fluoroquinolone AUIC break points and the link to bacterial killing rates. Part 2: human trials
.
Ann Pharmacother
 
2003
;
37
:
1478
88
.
19
Machuca
J
Briales
A
Labrador
G
et al
Interplay between plasmid-mediated and chromosomal-mediated fluoroquinolone resistance and bacterial fitness in Escherichia coli
.
J Antimicrob Chemother
 
2014
;
69
:
3203
15
.
20
Cavaco
LM
Aarestrup
FM
.
Evaluation of quinolones for use in detection of determinants of acquired quinolone resistance, including the new transmissible resistance mechanisms qnrA, qnrB, qnrS, and aac(6′)Ib-cr, in Escherichia coli and Salmonella enterica and determinations of wild-type distributions
.
J Clin Microbiol
 
2009
;
47
:
2751
8
.
21
Corvec
S
Cremet
L
Reynaud
A
et al
Vitek2 system: a reliable tool to detect qnr determinants in Enterobacteriaceae without quinolone resistance-determining region modifications
.
Diagn Microbiol Infect Dis
 
2009
;
64
:
455
7
.
22
Lupo
A
Papp-Wallace
KM
Sendi
P
et al
Non-phenotypic tests to detect and characterize antibiotic resistance mechanisms in Enterobacteriaceae
.
Diagn Microbiol Infect Dis
 
2013
;
77
:
179
94
.
23
Calvo
J
Cano
ME
Pitart
C
et al
Evaluation of three automated systems for susceptibility testing of Enterobacteria containing qnrB, qnrS, and/or aac(6′)-Ib-cr
.
J Clin Microbiol
 
2011
;
49
:
3343
5
.
24
Rodriguez-Martinez
JM
Conejo
MC
Martinez-Martinez
L
et al
Evaluation of antimicrobial susceptibility of bacteria containing the qnr gene and FOX-5 β-lactamase by four automated systems
.
Clin Microbiol Infect
 
2005
;
11
:
402
4
.
25
Rodriguez-Martinez
JM
Pichardo
C
Garcia
I
et al
Activity of ciprofloxacin and levofloxacin in experimental pneumonia caused by Klebsiella pneumoniae deficient in porins, expressing active efflux and producing QnrA1
.
Clin Microbiol Infect
 
2008
;
14
:
691
7
.
26
Guillard
T
Cambau
E
Chau
F
et al
Ciprofloxacin treatment failure in a murine model of pyelonephritis due to an AAC(6′)-Ib-cr-producing Escherichia coli strain susceptible to ciprofloxacin in vitro
.
Antimicrob Agents Chemother
 
2013
;
57
:
5830
5
.
27
Ruiz
E
Ocampo-Sosa
AA
Alcoba-Florez
J
et al
Changes in ciprofloxacin resistance levels in Enterobacter aerogenes isolates associated with variable expression of the aac(6′)-Ib-cr gene
.
Antimicrob Agents Chemother
 
2012
;
56
:
1097
100
.
28
Cano
ME
Calvo
J
Agüero
J
et al
Plasmid-mediated quinolone resistance among enterobacteria with reduced susceptibility or resistance to ciprofloxacin but susceptible to nalidixic acid
. In:
Abstracts of the Forty-seventh Interscience Conference on Antimicrobial Agents and Chemotherapy
,
Chicago, IL
,
2007
.
Abstract C2–152
,
p. 103
.
American Society for Microbiology
,
Washington, DC, USA
.
29
Ruiz
E
Saenz
Y
Zarazaga
M
et al
qnr, aac(6′)-Ib-cr and qepA genes in Escherichia coli and Klebsiella spp.: Genetic environments and plasmid and chromosomal location
.
J Antimicrob Chemother
 
2012
;
67
:
886
97
.
30
Martinez-Martinez
L
Hernandez-Alles
S
Alberti
S
et al
In vivo selection of porin-deficient mutants of Klebsiella pneumoniae with increased resistance to cefoxitin and expanded-spectrum-cephalosporins
.
Antimicrob Agents Chemother
 
1996
;
40
:
342
8
.
31
Martinez-Martinez
L
Garcia
I
Ballesta
S
et al
Energy-dependent accumulation of fluoroquinolones in quinolone-resistant Klebsiella pneumoniae strains
.
Antimicrob Agents Chemother
 
1998
;
42
:
1850
2
.
32
Lopez-Cerero
L
Bellido
MM
Serrano
L
et al
Escherichia coli O25b:H4/ST131 are prevalent in Spain and are often not associated with ESBL or quinolone resistance
.
Enferm Infecc Microbiol Clin
 
2013
;
31
:
385
8
.
33
Elder
BL
Hansen
SA
Kellogg
JA
et al
Verification and Validation of Procedures in the Clinical Microbiology Laboratory
 .
In: McCurdy BW, ed. CUMITECH: Cumulative Techniques and Procedures in Clinical Microbiology. Washington, DC: American Society for Microbiology, 1997
.