The spread of KPC-producing Enterobacteriaceae in Spain: WGS analysis of the emerging high-risk clones of Klebsiella pneumoniae ST11/KPC-2, ST101/KPC-2 and ST512/KPC-3

The authors employed whole-genome sequencing to show that 3 clones of KPC-producing Klebsiella pneumoniae were responsible for the spread of resistance throughout Spain.

Abstract

Objectives

We analysed the microbiological traits and population structure of KPC-producing Enterobacteriaceae isolates collected in Spain between 2012 and 2014. We also performed a comparative WGS analysis of the three major KPC-producing Klebsiella pneumoniae clones detected.

Methods

Carbapenemase and ESBL genes were sequenced. The Institut Pasteur MLST scheme was used. WGS data were used to construct phylogenetic trees, to identify the determinants of resistance and to de novo assemble the genome of one representative isolate of each of the three major K. pneumoniae clones.

Results

Of the 2443 carbapenemase-producing Enterobacteriaceae isolates identified during the study period, 111 (4.5%) produced KPC. Of these, 81 (73.0%) were K. pneumoniae and 13 (11.7%) were Enterobacter cloacae. Three major epidemic clones of K. pneumoniae were identified: ST11/KPC-2, ST101/KPC-2 and ST512/KPC-3. ST11/KPC-2 differed from ST101/KPC-2 and ST512/KPC-3 by 27 819 and 6924 SNPs, respectively. ST101/KPC-2 differed from ST512/KPC-3 by 28 345 SNPs. Nine acquired resistance genes were found in ST11/KPC-2, 11 in ST512/KPC-3 and 13 in ST101/KPC-2. ST101/KPC-2 had the highest number of virulence genes (20). An 11 bp deletion at the end of the mgrB sequence was the cause of colistin resistance in ST512/KPC-3.

Conclusions

KPC-producing Enterobacteriaceae are increasing in Spain. Most KPC-producing K. pneumoniae isolates belonged to only five clones: ST11 and ST512 caused interregional spread, ST101 caused regional spread and ST1961 and ST678 produced independent hospital outbreaks. ST101/KPC-2 had the highest number of resistance and virulence genes. ST101/KPC-2 and ST512/KPC-3 were recently implicated in the spread of KPC in Italy.

Introduction

The spread of carbapenemase-producing Enterobacteriaceae (CPE) has posed a challenge for health services worldwide.¹ Plasmid-mediated carbapenemases detected in Enterobacteriaceae are usually KPC, VIM, NDM, IMP or OXA-48 types.^2–4 KPC carbapenemases have proliferated worldwide and have become endemic in hospitals in the north-east USA, Israel, Greece and Italy.^2,4,5–8 Cases in which KPCs are involved have also been reported in other European countries; some of these cases are linked to people who have travelled to endemic areas.^2,9,10 The spread of KPC enzymes is a concern because of the limited number of therapeutic options available and the high rate of associated morbidity.¹¹

Recent studies in Spain have revealed an increase in the prevalence of CPE isolates,^9,12 mainly due to OXA-48 and VIM-1 producers.⁹ A smaller percentage of isolates (2.1%–3.4%) was found to produce KPC.^9,12 Nevertheless, over the last few years, large KPC-associated nosocomial outbreaks have been reported.^13,14

WGS can be applied to the study of MDR epidemic clones in a way that can improve our understanding of their evolution.^15,16

The aims of this study were: (i) to characterize the microbiological traits, population structure and evolutionary trends of KPC-producing Enterobacteriaceae in Spain; and (ii) to carry out a comparative WGS analysis of the three major KPC-producing epidemic clones in Klebsiella pneumoniae.

Materials and methods

Study design and bacterial isolates

Since 2009, our public health and reference institute has been operating an active and unrestricted national antibiotic resistance surveillance programme.¹² All Enterobacteriaceae isolates non-susceptible to carbapenems detected under the surveillance programme between January 2012 and December 2014 were included in this study. Only the first isolate per patient was analysed.

Enterobacteriaceae isolates were identified using standard microbiological methods and the API 20E identification system (bioMérieux, France). When necessary, species identification was confirmed by 16S ribosomal DNA sequencing.

Antibiotic susceptibility and phenotypic characterization of carbapenemase production

Following EUCAST guidelines, the disc diffusion and microdilution methods were used to test antibiotic susceptibility.^17,18 EUCAST screening cut-off values were used to identify CPE.¹⁹ All of the isolates were analysed with both the Carba NP method²⁰ and a modified Hodge test. Carbapenemase activity was inhibited with EDTA, phenylboronic acid and cloxacillin, according to EUCAST guidelines.¹⁹

Characterization of resistance mechanisms and genetic environment

Genes encoding carbapenemases⁵ and ESBLs²¹ were detected using PCR and DNA sequencing assays.

The genetic environment of bla_KPC was studied by long-PCR mapping as previously described.²² All PCR products were sequenced and compared with sequences available in the GenBank database (http://www.ncbi.nlm.nih.gov/genbank/).

Molecular epidemiology

Carbapenemase-producing K. pneumoniae and Escherichia coli isolates were typed using the MLST schemes of the Institut Pasteur (http://bigsdb.pasteur.fr/klebsiella/klebsiella.html) and the University of Warwick (Warwick Medical School, Coventry, UK; http://mlst.warwick.ac.uk).

WGS, SNP tree and core genome MLST analysis

WGS was performed on the genomes of one representative isolate of each of the three major K. pneumoniae clones that had been revealed by MLST.

DNA was extracted using the QIAamp^® DNA Mini Kit (Qiagen, Hilden, Germany). Genomic DNA paired-end libraries were generated using the Nextera XT DNA Sample Preparation Kit (Illumina, San Diego, CA, USA). These libraries were sequenced using an Illumina MiSeq next-generation sequencer with 2 × 300 bp paired-end reads (Illumina). The short reads acquired were subsequently assembled de novo into contigs using Velvet²³ under parameters optimized to give the highest N50 value. The draft genomes were ordered using the Artemis Comparison Tool²⁴ and Abacas²⁵ and then uploaded to the Rapid Annotation using Subsystem Technology (RAST) server for subsystem classification and functional annotation.²⁶

Antimicrobial resistance genes were analysed using the ResFinder tool (https://cge.cbs.dtu.dk/services/ResFinder/) using an identity threshold of 98%. In order to reconstruct the plasmids of each genome, the recently developed methodology Plasmid Constellation Networks (PLACNET) was used.²⁷

Virulence genes were identified using the BIGSdb-Kp database (Institut Pasteur; http://bigsdb.web.pasteur.fr).²⁸ For comparison purposes, the whole-genome sequences of three previously described KPC-producing K. pneumoniae strains were included in the analysis of resistance and virulence genes. These were: KPN1H1/ST258, ST512-K30BO and HS11286/ST11 (GenBank accession numbers NZ_CP008827.1, HS11286 and CAJM00000000.2, respectively).^29–31 These strains had recently been correlated with outbreaks in countries outside Spain.

Sequence reads were mapped to the chromosome of K. pneumoniae NTUH-K2044³² (GenBank accession no. NC_012731.1). Subsequently, the software package Burrows-Wheeler Aligner (BWA) was used to identify SNPs.³³ Sequence reads from other bla_KPC-carrying K. pneumoniae isolates were included in order to establish phylogenetic relationships (GenBank accession numbers NZ_CP008827.1, CAJM00000000.2, AFXH01000001.1, AFQL01000001.1, AJJI01000001.1, AKAF01000001.1, AKAJ01000001.1, AKAN01000001.1, AMLO01000001.1, ATAK01000001.1, CANR01000001.1, NZ_JH930438.1, CP003200.1, CP003785.1, AFQK01000001.1, CP002910.1, NZ_JH930400.1, NZ_KI517376.1, CP000647.1, NZ_JH930419.1, NZ_JH930422.1, LGAB01000001.1 and CAJM02000001.1).

Sequence reads mapped to an average of 92% of the reference genome, with a mean depth of 75x in mapped regions across sequenced isolates. SNPs were identified as described by Harris et al.³⁴ and pairwise comparisons were carried out. SNPs called in integrative and conjugative elements and phage sequences that had been identified using a prophage locus prediction tool³⁵ were excluded. Sequences were suspected of having resulted from homologous recombination when an elevated SNP density was observed using Gubbins software.³⁶

A maximum-likelihood phylogenetic tree was reconstructed as previously described.^16,37 A set of 694 genes was considered as the core genome MLST (cgMLST)²⁸ and cgMLST-based minimum spanning trees were calculated and drawn using Ridom SeqSphere+ software (Ridom, Munster, Germany).

Results and discussion

Bacterial isolates and carbapenemase types

Between January 2012 and December 2014, 2443 CPE isolates were identified, of which 111 (4.5%) were KPC-producing Enterobacteriaceae: 8 of these were isolated in 2012 (3.4% of the 237 CPE isolates studied), 34 in 2013 (4.4% of 775) and 69 in 2014 (4.8% of 1431). The 111 KPC isolates were collected in 29 Spanish hospitals located in 13 geographical areas.

Of the bacterial species carrying bla_KPC, the most prevalent was KPC-2-producing K. pneumoniae (59.5%) (Table 1). In Spain, OXA-48 is the most common carbapenemase identified in Enterobacteriaceae;^9,38 however, the number of isolates producing KPC enzymes has significantly increased over the last few years.⁹

Table 1.

Bacterial species carrying bla_KPC genes in Spain according to the Spanish Antibiotic Resistance Surveillance Program (2012–14)

Species	Number of isolates (%)	Carbapenemase types (n)
Klebsiella pneumoniae	81 (73.0)	KPC-2 (66)a, KPC-3 (15)
Enterobacter cloacae	13 (11.7)	KPC-2 (12), KPC-3 (1)
Citrobacter freundii	9 (8.1)	KPC-2 (9)
Escherichia coli	4 (3.6)	KPC-2 (4)
Klebsiella oxytoca	2 (1.8)	KPC-2 (2)b
Serratia marcescens	1 (0.9)	KPC-2 (1)
Enterobacter aerogenes	1 (0.9)	KPC-2 (1)

^aTwo isolates coproduced both KPC-2 and OXA-48 carbapenemases.

^bOne isolate coproduced both KPC-2 and VIM-1 carbapenemases.

Population structure of KPC-producing Enterobacteriaceae isolates

The 81 KPC-producing K. pneumoniae isolates were grouped into three primary STs: ST11 (46 isolates), ST101 (13 isolates) and ST512 (10 isolates) (Table 2). ST11 and ST512 were involved in the interregional spread of K. pneumoniae and ST101 in the regional spread. ST101 predominated in 2012 (57.1%), ST512 appeared in 2013 and ST11 widely disseminated in 2013 and 2014.

Table 2.

Distribution of K. pneumoniae MLST clones producing KPC carbapenemases in Spain according to the Spanish Antibiotic Resistance Surveillance Program (2012–14)

ST	Carbapenemase	Number of cases	Number of hospitals	Provinces
11	KPC-2	44	13	Madrid, Toledo, Ciudad Real, Guadalajara and Cuenca
	KPC-3	2	2	Madrid
101	KPC-2	13	5	Madrid
512	KPC-3	10	7	Jaén, Málaga, Badajoz, Murcia, Toledo and Madrid
1961	KPC-2	5	1	Pontevedra
678	KPC-3	3	1	Madrid
359	KPC-2	1	1	Madrid
565	KPC-2	1	1	Madrid
1962	KPC-2	1	1	Madrid
1963	KPC-2	1	1	Almeria

The KPC-3-producing ST512 caused a severe outbreak in Cordoba,¹³ Andalucía, Spain, whose index case was a patient transferred from Italy. ST512 is one of the most frequent STs of KPC-producing K. pneumoniae isolates in Italy.^4,39,40 ST11 is one of the STs most commonly associated with the production of various types of carbapenemases.^4,9 Both ST11 and ST512 are single-locus variants of the globally disseminated high-risk clone ST258,⁴¹ which was not detected in this study. However, ST258 was shown to be a hybrid derivative of ST11, in which a substantial part of the genome has been replaced.⁴² ST101 is an emerging clone that is unrelated to ST258, previously described in carbapenem-resistant K. pneumoniae isolates producing OXA-48,⁴³ KPC-2^12,40,44 and CTX-M-15 plus OmpK35 porin loss.⁴⁵

The four KPC-producing E. coli isolates were unrelated and belonged to four different STs: ST131, ST90, ST355 and ST569. The high-risk E. coli clone ST131 had previously been described as producing, albeit infrequently, different types of carbapenemases.^9,46

The genetic relationship between Enterobacter cloacae and Citrobacter freundii isolates, as determined by PFGE after digestion of chromosomal DNA with XbaI, is shown in Figure S1 (available as Supplementary data at JAC Online).

Analysis of the major KPC-producing K. pneumoniae clones

ST11, ST101 and ST1961 isolates produced KPC-2, whereas ST512 and ST678 isolates produced KPC-3 (Table 2). Isolates of the ST101/KPC-2, ST1961/KPC-2 and ST678/KPC-3 clones coproduced the ESBLs CTX-M-15, CTX-M-14 and SHV-41, respectively. ESBL genes were not found in isolates of ST11/KPC-2 and ST512/KPC-3 clones. KPC-producing K. pneumoniae generally coproduce ESBLs with low frequency,⁴⁶ but this assertion is mainly based on observations on clonal complex 258. New KPC-producing clones that also produce ESBLs are emerging.⁴⁷

The ST11/KPC-2 isolates were resistant to almost all of the antibiotics tested, but were fully susceptible to tigecycline, colistin, tetracycline and fosfomycin; 88.9% were susceptible to amikacin and 22.2% to gentamicin and tobramycin. All ST512/KPC-3 isolates were susceptible to tigecycline and gentamicin, while only 20% were susceptible to colistin. ST101 isolates were fully susceptible to tigecycline and colistin; one isolate (7.7%) was also susceptible to co-trimoxazole.

WGS of ST11/KPC-2, ST101/KPC-2 and ST512/KPC-3 clones

According to WGS data, ST11/KPC-2 differed from ST101/KPC-2 and ST512/KPC-3 by 27 819 and 6924 SNPs, respectively. ST101/KPC-2 differed from ST512/KPC-3 by 28 345 SNPs.

The metrics of all assemblies and the contig annotation results obtained with RAST are laid out in Table S1. Using this approach, the presence of 103, 109 and 128 genes related to antibiotic and toxic compound resistance was detected in ST11/KPC-2, ST101/KPC-2 and ST512/KPC-3 clones, respectively.

To show to what extent ST11/KPC-2, ST101/KPC-2 and ST512/KPC-3 clones are genetically related to previously sequenced KPC-producing K. pneumoniae isolates from other countries, we identified 83 521 high-quality SNPs with respect to the whole-genome sequence of K. pneumoniae strain NTUH-K2044³² that were used to construct a maximum-likelihood phylogenetic tree (Figure 1a). The ST11/KPC-2 and ST512/KPC-3 isolates were grouped together with isolates belonging to ST258. In a pairwise comparison, the Italian isolate ST512-K30BO³⁰ was shown to be closely related to the Spanish ST512/KPC-3 representative isolate, since their sequences differed from each other by 75 SNPs (Figure 1a) and only had two allelic mismatches (Figure 1b).

Figure 1.

Population structure of K. pneumoniae expressing bla_KPC genes. (a) Maximum-likelihood tree showing the relationship between the isolates sequenced in this study and similar strains whose sequences were obtained from the GenBank database. Branch length is indicative of the number of SNPs. (b) Minimum spanning tree of K. pneumoniae genomes as determined on the basis of the 694 core genome multilocus ST. Identification of isolates includes the ST; Spanish sequenced isolates are highlighted in boxes (a) or black circles (b).

ST101/KPC-2 was not grouped with the other KPC isolates analysed. There were ≥435 allelic mismatches between the sequences of ST101/KPC-2 and those of the other K. pneumoniae isolates. The main cluster was formed by ST258 and ST512 isolates (Figure 1b), which had an average of 10 allelic mismatches. There were 50 allelic mismatches between this cluster and ST11 isolates (Figure 1b), showing that ST512 and ST258 are more closely related to each other than ST11 as previously described by Bialek-Davenet et al.²⁸

Profile of antibiotic resistance genes of ST11/KPC-2, ST101/KPC-2 and ST512/KPC-3 by WGS

The genetic profile of antibiotic resistance genes (resistome) differed in the ST11/KPC-2, ST101/KPC-2 and ST512/KPC-3 representative isolates, although all of them carried many accessory resistance genes: 9 in ST11/KPC-2, 11 in ST512/KPC-3 and 13 in ST101/KPC-2 (Figure 2). Only ST101/KPC-2 produced an ESBL (bla_CTX-M-15). A good correlation was observed between the genetic and phenotypic profiles of resistance for β-lactam and aminoglycoside antibiotics.

Figure 2.

Distribution of multidrug resistance loci and virulence genes in the genomes of bla_KPCK. pneumoniae clones sequenced in Spain in comparison with strains isolated in other countries. The names of genes responsible for drug resistance (resistome) and virulence are found at the top of a heat map in which black rectangles indicate the presence of a gene and white rectangles indicate its absence. KPN1H1/ST258 was detected during an outbreak in the USA,²⁹ ST512-K30BO was isolated during an outbreak in Italy³⁰ and HS11286/ST11 was described in China.³¹

These results were compared with the previously sequenced genomes of KPC-producing K. pneumoniae (Figure 2).^29–31 The ST101/KPC-2 Spanish representative isolate had the highest number of resistance genes (n = 13), while the Italian ST512-K30BO³⁰ isolate had the lowest number (n = 7). Remarkably, the resistance gene profile of the ST512/KPC-3 Spanish representative isolate differed from the ST512 K. pneumoniae isolate detected during an outbreak in Italy;³⁰ however, it was very similar to the KPN1H1/ST258 isolate detected during an outbreak in the USA.²⁹

The following topoisomerase mutations were identified: gyrA S83I, found in both ST11/KPC-2 and ST512/KPC-3; gyrA S83Y and D87G, found in ST101/KPC-2 only; and parC S80I, observed in all three isolates. An OqxAB efflux pump was also detected. Only ST101/KPC-2 had the aac(6′)-Ib-cr gene implicated in aminoglycoside and ciprofloxacin resistance.

All three Spanish representative isolates had the WT ompK35 gene (GenBank accession no. AJ011501). ST101/KPC-2 had an ompK36 gene with a 9 nt deletion at positions Δnt549–557 and an additional deletion at position 713. The latter deletion caused a frameshift that introduced a premature TGA stop codon after nine amino acids. The 9 nt deletion corresponded to a loss of Leu184, Ser185 and Pro186. These amino acids form part of the L4 loop, which is found on the cell surface.⁴⁸ ST11/KPC-2 and ST512/KPC-3 had the WT ompK36 gene (GenBank accession no. NC_016845.1). Rapid characterization of ompK36 genotypes by WGS might be a useful method for predicting the efficacy of combination therapies that include carbapenems.⁴⁹

The colistin-resistant ST512/KPC-3 isolate had an mgrB coding sequence with a deletion of 11 bp (Δnt109–119) that resulted in a frameshift leading to a premature termination codon (TAA) associated with colistin resistance.⁵⁰ The clonal spread of a colistin-resistant ST512 strain with the same deletion in mgrB has been reported in a recent Italian study.³⁹ The colistin-susceptible isolates ST11/KPC-2 and ST101/KPC-2 had a WT mgrB gene (GenBank accession no. KF852760.1). The plasmid-carried colistin resistance gene mcr-1 was not detected in any of these isolates.

Genetic environment of the bla_KPC genes

In the ST512 clone, the bla_KPC-3 gene was carried by the classical transposon Tn4401 ‘isoform’ (Figure S2). This transposon contains a 100 bp deletion between the sequences of istB and bla_KPC. The sequence detected was tnpR-tnpA-ISKpn7(istA-istB)-bla_KPC-3-ISKpn6-tnpA-IRrTn4401-tnpR-tnpA (total size 20 338 bp), which was identical to that of the pNYC plasmid (GenBank accession no. EU176011).⁵¹

The genetic environment of bla_KPC-2 in ST11 and ST101 was indistinguishable between them, but different from bla_KPC-3 of ST512; it had ISKpn8 and a truncated bla_TEM gene that was 570 bp in length. The sequence was tnpA-tnpR-ISKpn8-(trun)bla_TEM-bla_KPC-2-ISKpn6-tnpA-korC-klcA (total size 7704 bp) (Figure S2). This sequence was very similar to Variant 1 described in the pKP048 plasmid from a Chinese K. pneumoniae isolate⁵² (accession number not provided). A similar genetic environment for bla_KPC-2 has been described in other countries such as Argentina.⁵³

Reconstructing the bla_KPC plasmids of ST11/KPC-2, ST101/KPC-2 and ST512/KPC-3 using the WGS data

In ST512/KPC-3 and ST101/KPC-2 isolates, the bla_KPC genes were located in constructed IncF plasmids in which the RepFIA and FrepB replicons were detected. In the case of the ST11/KPC-2 isolate, the bla_KPC-2 gene was detected in a non-typeable plasmid. The predominant bla_KPC plasmids described in the literature are IncF plasmids, mainly associated with ST258.⁴ As shown in this and previous studies, IncF is also associated with the KPC-3-producing clone ST512 of K. pneumoniae.⁴⁴ In addition, bla_KPC-2 has been found in plasmids of different incompatibility groups (IncP, IncR and IncA/C2) and non-typeable plasmids.^14,54,55

All three sequenced strains had the IncW plasmid containing the resistance genes aad, sul and dfrA. In the ST101/KPC-2 isolate, bla_CTX-M-15 appeared exclusively in the chromosome and was directly linked to an upstream ISEcp1-like element, which was known to affect expression and mobilization of this gene. Chromosomally encoded CTX-M alleles have been previously reported,^56–58 including ST101 K. pneumoniae.⁵⁹ An IncFII plasmid harbouring bla_KPC-2 and bla_{CTX-M-1-group} has recently been detected in an Italian ST101 isolate.⁴⁴

Profile of the virulence genes of ST11/KPC-2, ST101/KPC-2 and ST512/KPC-3 by WGS

Five virulence genome features were detected out of the nine documented in the BIGSdb-Kp database²⁸ (Figure 2). These were the siderophore gene irp2, the yersiniabactin siderophore cluster (ybtAEPQSTUX), the mannose-resistant Klebsiella-like (type III) fimbriae cluster (mrkABCDFHIJ), the ferric uptake system (kfuABC) and the yersiniabactin receptor gene fyuA. Considering all together the different genes included in these virulence genome features, a total of 21 virulence genes were detected: 20 in ST101/KPC-2, 8 in ST512/KPC-3 and 7 in ST11/KPC-2. ST512/KPC-3 and ST11/KPC-2 only had virulence genes from the fimbriae cluster (Figure 2). The Spanish ST101/KPC-2 clone had the highest number of virulence genes, even when compared with the three KPC-producing isolates that had previously been sequenced (Figure 2).

The yersiniabactin siderophore cluster, only detected in the Spanish ST101/KPC-2 isolate in this study, has already been associated with ST258 isolates.⁶⁰ The presence of yersiniabactin in K. pneumoniae KPC producers is worrisome, because it is strongly linked with colonization of the respiratory tract in humans.⁶⁰ In our study, three out of the seven (42.9%) respiratory tract infections were caused by the ST101/KPC-2 K. pneumoniae clone.

Analysis of the capsular polysaccharide cps locus, which contained galF, orf2, wzi, wza, wzb and wzc, showed that each sequenced isolate had a different wzi allele. According to the BIGSdb-Kp database, these wzi alleles corresponded to a different K type: K13 corresponded to ST11/KPC-2, K17 corresponded to ST101/KPC-2 and no K type was associated with ST512/KPC-3. The cps gene cluster detected in ST512/KPC-3 is identical to the gene cluster cps_BO-4 already described in other Italian isolates of ST512; this CPS type is different from any other studied.⁶¹

The plasmid-associated heavy metal resistance clusters pcoABCDERS and silCERS (responsible for copper and silver resistance, respectively) were only detected in the ST512/KPC-3 isolate and in the previously reported KPN1H1/ST258 clone.²⁹ These clusters had the following allele assignments: 13,2,1,2,1,1,2 and 3,3,2,2, respectively; both clusters were widely distributed.²⁸

Conclusions

KPC-producing isolates are increasing and spreading throughout many geographical areas of Spain, mainly due to the non-ST258 high-risk clones of K. pneumoniae: ST11/KPC-2, ST101/KPC-2 and ST512/KPC-3. The ST101/KPC-2 clone was highly resistant and hypervirulent. Of concern is the detection in Spain of K. pneumoniae clones similar to those recently implicated in the spread of KPC isolates in Italy.

Other members of the Spanish Antibiotic Resistance Surveillance Program Collaborating Group

José Antonio Rodríguez-Polo (H. Virgen de la Salud, Toledo, Spain), Esteban Aznar and Carolina Campelo (Laboratorio Br Salud, San Sebastián de los Reyes, Madrid), Rocío Martínez-Ruiz and Isabel Sánchez-Romero (H. Universitario Puerta de Hierro Majadahonda, Majadahonda, Madrid), Silvia Jimenez-Álvarez (Hospital General Universitario de Ciudad Real, Ciudad Real), Luisa García-Picazo (H. de El Escorial, Madrid), Alberto Delgado-Iribarren (Hospital Universitario Fundación Alcorcón, Alcorcón, Madrid), Patricia Álvarez-García (Complejo Hospitalario Universitario de Pontevedra, Pontevedra), Yolanda Gil and María Almagro (H. Universitario de Móstoles, Móstoles, Madrid), Emilia Cercenado (H. General Universitario Gregorio Marañón, Madrid), Eva M. González Barberá and Salvador Giner Almaraz (H. Universitario y Politécnico La Fe, Valencia), Inmaculada Carazo and Inocente Cuesta (H. Médico Quirúrgico de Jaén), Waldo Sánchez-Yebra (Complejo Hospitalario Torrecárdenas, Almería), David Molina Arana and Juan-Ignacio Alós (H. Universitario de Getafe, Getafe, Madrid), Juan Manuel Hernández Molina (H. Universitario Carlos Haya, Málaga), Alejandro Gonzalez Praetorius and Sonia Solís del Baño (H. Universitario de Guadalajara, Guadalajara), José Luis Hernández (H. de Cruces, Barakaldo, Vizcaya), Concepción Baladón (Lab. Dr Echevarne), Rafael Carranza (H. General La Mancha Centro, Ciudad Real), Isabel Wilhelmi (H. Severo Ochoa, Leganés, Madrid), Eugenio Garduño (Complejo Hospitalario Universitario de Badajoz, Badajoz), Genoveva Yagüe (H. Virgen de la Arrixaca, Murcia), Cristina Frau (H. de Santa Bárbara, Ciudad Real), Mª José Rodriguez-Escudero (H. General Virgen de la Luz, Cuenca) and Ángel Somodevilla (Megalab).

Funding

This work was supported by grant PI12/01242 from the Fondo de Investigación Sanitaria and grant PI15CIII/00017 from Instituto de Salud Carlos III. This research was also supported by the Antibiotic Resistance Surveillance Program of the Centro Nacional de Microbiología (Instituto de Salud Carlos III, Ministerio de Economía y Competitividad) of Spain. It was also supported by grant REIPI RD12/0015 awarded by the Spanish Network for Research in Infectious Diseases, Instituto de Salud Carlos III, Subdirección General de Redes y Centros de Investigación Cooperativa, Ministerio de Economía y Competitividad.

Transparency declarations

None to declare.