Epidemic diffusion of KPC carbapenemase-producing Klebsiella pneumoniae in Italy: results of the first countrywide survey, 15 May to 30 June 2011

report on the results of a countrywide cross-sectional survey, carried out from 15 May to 30June 2011 to investigate the diffusion of CRE in Italy and to characterise the most prevalent resistance mechanisms and their dissemination patterns. CRE were reported from most (23 of 25) participating laboratories, with an overall proportion of 3.5% and 0.3% among consecutive non-duplicate clinical isolates of Enterobacteriaceae from inpatients (n=7,154) and outpatients (n=6,595), respectively.


Introduction
The increasing resistance to carbapenems among Enterobacteriaceae has become a public health problem of major concern [1,2].Carbapenem-resistant Enterobacteriaceae (CRE) usually exhibit complex multidrug resistance phenotypes that leave very few therapeutic options [3,4], and infections caused by CRE are associated with increased morbidity and mortality in comparison with those caused by carbapenem-susceptible strains (72% versus 22%) [5].
At least two mechanisms can be responsible for acquired carbapenem resistance in Enterobacteriaceae: (i) reduced outer membrane permeability by porin loss in combination with the production of an extendedspectrum beta-lactamase (ESBL) or of AmpC-type betalactamase; and (ii) production of beta-lactamases capable of hydrolysing carbapenems (carbapenemases) [2].While the former mechanism is a result of mutation and has a low overall propensity to disseminate, acquired carbapenemases are encoded by transferable genes that can disseminate among different strains and different species, and carbapenemase production is the leading carbapenem resistance mechanism in Enterobacteriaceae [2,6,7].Several types of acquired carbapenemases have been detected in CRE, with KPC-, VIM-, NDM-and OXA-48-type enzymes being the most prevalent, although with a notable geographical variability [2,[6][7][8][9].
In Europe, individual cases or outbreaks of CRE have been reported in several countries [6] but data from the EARS-NET database show that, until 2009, the proportion of CRE has remained overall low in most countries except Greece and Cyprus, where high-level endemicity of carbapenem-nonsuscetible Klebsiella pneumoniae has been reported since the mid-2000s [10].In Italy, sporadic cases or outbreaks caused by CRE of various species and with different resistance mechanisms have been reported since the early 2000s [11][12][13][14][15][16][17][18][19][20][21][22][23][24], but only since 2010 an abrupt and notable increase in the proportion of carbapenem-non-susceptible K. pneumoniae has ben reported by the EARS-NET surveillance system [10].This trend has recently been confirmed by data from the Micronet sentinel surveillance network [25].However, the resistance mechanisms responsible for this increase have not been investigated.
In this work we report the results of a countrywide cross-sectional survey promoted by the Italian Society of Clinical Microbiologists (AMCLI) and carried out in mid-2011, to investigate the diffusion of CRE in Italy and to characterise the most prevalent resistance mechanisms and their dissemination patterns.Results confirmed that CRE have reached epidemic dissemination in Italy, and revealed that this condition was mostly related with the clonal diffusion of K. pneumoniae producing KPC-type carbapenemases (KPC-KP) of clonal complex (CC) 258.

Study design
Twenty-five large clinical microbiology laboratories from 23 Italian cities, distributed across the national territory and covering most Italian regions, participated in the study (Figure 1).During the period from 15 May to 30 June 2011, each laboratory collected consecutive non-replicate clinical isolates of Enterobacteriaceae, from any site of infection, that exhibited minimum inhibitory concentrations (MICs) for imipenem and/or meropenem and/or ertapenem higher than 1 mg/L (for isolates of Proteeae, i. e. Morganella morganii, Proteus spp.and Providencia spp., only meropenem and ertapenem MICs were considered).The collected isolates were transferred to reference laboratories for confirmation of species identification and carbapenem MICs, and for characterisation of the carbapenem resistance mechanisms and analysis of clonal relatedness.For each isolate, information on the clinical specimen and type of ward (in case of isolates from inpatients) were provided.Moreover, each participating laboratory provided information on the total number of consecutive non-duplicate clinical isolates of Enterobacteriaceae observed during the collection period.

Characterisation of bacterial isolates and of resistance determinants
Bacterial identification and antimicrobial susceptibility testing were carried out by the collecting laboratories using either the Phoenix Automated Microbiology System (Becton Dickinson Diagnostic Systems, Sparks, United States) or the Vitek-2 System (bioMérieux, Marcy l'Etoile, France).Confirmatory identification was carried out by Matrix-assisted laser desorption ionisation -time of the flight (MALDI-TOF) mass spectrometry (Vitek-MS, bioMérieux).Confirmatory MIC testing for imipenem, meropenem and ertapenem was carried out by Etest (bioMérieux).All collected isolates confirmed to be non-susceptible to imipenem (not considered for Proteeae) and/or meropenem and/or ertapenem according to the EUCAST breakpoints [26] were considered as CRE for the purposes of this study.For KPC-KP, MICs of colistin, tigecycline and gentamicin were determined by the reference broth microdilution method [27], and results were interpreted according to the EUCAST breakpoints [26].CRE were evaluated for carbapenemase production by meropenem plus EDTA and meropenem plus phenylboronic acid using the disk diffusion method [28,29], and for the presence of the most common carbapenemase genes (bla KPC -, bla VIM -, bla NDM -, bla OXA-48 -type) by dotblot hybridisation [30].PCR amplification [13,[31][32][33] and sequencing of PCR amplicons were used to identify the carbapenemase genes detected by hybridisation.All CRE isolates testing negative in disk diffusion test and/or in hybridisation assays were further investigated for production of carbapenemase activity by modified Hodge test [34] and by spectrophotometric assay with crude extracts [35].No CRE detected

Analysis of clonal relatedness
Genotyping of K. pneumoniae isolates by pulsed-field gel electrophoresis (PFGE) profiling of genomic DNA was carried out after digestion with XbaI with a CHEF-DRIII apparatus (Bio-Rad, Hemel Hempstead, United Kingdom) [36], and results interpreted as recommended by Van Belkum et al. [37].Multi-locus sequence typing (MLST) of K. pneumoniae isolates was performed as previously described [38], and sequence types (STs) were assigned using the MLST web site [39].

Proportions of carbapenem-resistant Enterobacteriaceae from inpatients and outpatients in Italy
During the study period (15  CRE were reported from 23 of the 25 participating laboratories (Figure 1).Carbapenem-nonsusceptible isolates of K. pneumoniae were detected at any of these 23 laboratories, with proportions ranging from 1.2 to 32.7% (mean: 11.5%) (Figure 1).

Carbapenem resistance mechanisms in carbapenem-resistant Enterobacteriaceae
Carbapenemase production was detected in the majority of CRE (85%) (Table 2).KPC-type enzymes were by far the most common (89.5% of carbapenemase producers).Other types of carbapenemases included VIM-1 and OXA-48 (9.2% and 1.3% of carbapenemase producers, respectively).KPC-type enzymes were only detected in K. pneumoniae and in one E. coli.VIM-1, although much less prevalent, was detected in a wider variety of bacteria: K. pneumoniae, Klebsiella oxytoca, E. coli and Enterobacter cloacae.OXA-48 was only detected in three K. pneumoniae isolates (Table 2).KPC-producers were detected in 21 of 25 centres, showing a countrywide distribution.VIM-1-producers were detected in six centres, while the three OXA-48producers were from three different centres (Figure 1).Other types of carbapenemases, including NDM, were not detected.All collected isolates confirmed to be non-susceptible to imipenem (not considered for Proteeae) and/or meropenem and/or ertapenem according to the EUCAST breakpoints [26] were considered as CRE for the purposes of this study.

Carbapenem-non-susceptible Klebsiella pneumoniae: proportion in clinical specimens, distribution in hospital wards, and carbapenem MICs
The overall proportion of carbapenem non-susceptibility was approximately seven-fold higher in K. pneumoniae isolates from inpatients than in those from outpatients (16.3% versus 2.4%, Table 1).Considering isolates from inpatients, the proportion of carbapenem non-susceptibility was higher among bloodstream isolates than among isolates from other specimens (Table 3), revealing that carbapenem-non-susceptible K. pneumoniae strains circulating in Italy retained a remarkable potential for causing invasive infections.In the case of outpatients, carbapenem-non-susceptible isolates of K. pneumoniae were obtained only from urine, which was by far the most common specimen, and most of them were KPC-producers (Table 3).
Concerning the in-hospital distribution, 42.5% of the 219 carbapenem-nonsusceptible K. pneumoniae from inpatients were from intensive care units (ICUs), while 32.4% were from medical wards, 21.5% from surgical wards, and 3.6% from other areas.
Carbapenem MICs of the 234 carbapenem-nonsusceptible K. pneumoniae are reported in Table 4. Virtually all isolates were resistant to ertapenem, while some were intermediate or susceptible to the other carbapenems.A susceptible or intermediate phenotype to imipenem and/or meropenem was observed with most of the VIM producers, with the OXA-48 producers and with the non-carbapenemase producers, while most of the KPC producers were resistant also to these drugs.

Molecular epidemiology and susceptibility to non-betalactam agents of KPC-producing Klebsiella pneumoniae
To gather information on the molecular epidemiology of KPC-KP circulating in Italy, the 204 KPC-KP isolates were characterised by MLST and by PFGE genotyping, and their KPC allelic variants were determined by sequencing.MLST revealed that most isolates belonged in CC 258, either ST-258 or ST-512 (a single locus variant of ST-258), while a minority were ST-101.ST-258 and ST-512 isolates were detected in many centres, with an overall countrywide distribution, while ST-101 isolates were detected in only two centres, where ST-512 was also present (Figure 2).Isolates of ST-258 carried either bla KPC-2 or bla KPC-3 alleles, while those of ST-512 and of ST-101 carried bla KPC-3 and bla KPC-2 , respectively (Figure 2).PFGE genotyping revealed an oligoclonal population of KPC-KP, with a prevalent PFGE profile, named A, consisting of several variants (A0 to A6) and including the isolates of CC-258.In particular, the A0, A4 and A5 variants included isolates of ST-512 producing KPC-3, while the A1 and A2 variants included isolates of ST-258 producing KPC-3, and the A3 and A6 variants included isolates of ST-258 producing KPC-2 (Figure 2).Each variant was detected in multiple centres (Figure 2), confirming the epidemic propensity of this lineage.An additional PFGE profile, named B and consisting of only a few variants (B0 to B2), included the isolates of ST-101 producing KPC-2 and was detected in two centres (Figure 2).
Susceptibility testing of the 204 KPC-KP against non-beta-lactam agents which often retain activity against these strains, revealed that 22.4% were resistant to colistin, 20.9% were non-susceptible to tigecycline, and 15.8% were non-susceptible to gentamicin.Concerning co-resistances, 1.5% of the KPC-KP were non-susceptible to all three drugs, while 6.4% were non-susceptible to tigecycline and colistin, 1% to gentamicin and colistin, and 2.7% to tigecycline and gentamicin.No association was detected between ST and non-susceptibility to colistin or tigecycline, while non-susceptibility to gentamicin was more frequent in isolates of ST-101 (85.8% versus 12.6% in ST-258 or 10% in ST-512, p<0.05).

Discussion
A rapid dissemination of carbapenem-non-susceptible K. pneumoniae has been reported in Italy since 2010 [10,24].Results of this cross-sectional survey, carried out in mid-2011, confirmed that CRE have undergone an epidemic diffusion in Italy, with a widespread distribution across the national territory (although with variable proportions in different areas), and that the phenomenon was mostly related with rapid dissemination of carbapenem-non-susceptible K. pneumoniae, while the contribution by CRE of other enterobacterial species was much more limited.
Production of KPC-type carbapenemases was the most prevalent carbapenem resistance mechanism.Production of other carbapenemases and non-carbapenemase-mediated mechanisms were also detected, but remained in the background.However, continuous surveillance should be enforced, considering a recent outbreak caused by NDM-1-producing K. pneumoniae [18].phenomenon was also observed in other countries and further underscores the propensity for dissemination of multi-resistant K. pneumoniae strains belonging to ST-258 and ST-512 [40][41][42][43][44][45][46][47].Emergence of KPC-KP belonging to ST-101 was also observed, although only in two of the participating centres.This finding, along with recent reports of ST-101 isolates of KPC-KP from Italy, Brazil and the United States [21,42,48,49], emphasises the emerging role in dissemination of KPC of this clonal lineage which is also involved in the dissemination of other carbapenemases, such as OXA-48 and OXA-181 [50,51] as well as extended-spectrum beta-lactamases [52,53].
Since aggressive infection control was shown to be effective in controlling the dissemination of KPC-KP [54][55][56], present results mandate for strong and prompt intervention in Italy.Implementation of infection control measures on a countrywide scale appears now to be necessary, in addition to the actions that have already been taken at local and regional level (the Italian public healthcare system has a typically regional organisation) with positive results.In a hospital in Catania, Sicily, it was possible to control the spread of a KPC-3-producing K. pneumoniae clone without closing the ICU, by applying a multimodal infection control programme [16].The Emilia-Romagna region issued in July 2011 guidelines and protocols to monitor and control the spread of carbapemenase-producing Enterobacteriaceae in all the healthcare structures the region.The increasing trend of KPC-producing K. pneumoniae slowed down in the second half of 2011 and early 2012, and in hospitals that had been able to implement all control activities the number of cases had a remained stable or showed a sustained decrease [57].Dissemination of carbapenem-non-susceptible K. pneumoniae was not restricted to ICU settings, but affected all major hospital sectors.Although at least some of these patients could have been transferred from ICU settings, this propensity to dissemination on  multiple wards should be considered when planning infection control strategies.
Although cases and outbreaks of CRE have been reported in many European countries [6], only Greece, Cyprus and Italy have experienced such an extensive CRE epidemic to date.However, the phenomenon observed in these countries deserves considerable attention by public health authorities in Europe, considering the high mobility of people (tourists, workers, patients) within Europe, and the difficulties in containing the epidemic diffusion of CRE.

table 4
[26]nsion of KPC-KP strains belonging to variants (ST-258 and ST-512) of the hyperepidemic CC-258, detected for the first time in Italy in late 2008[13], was responsible for a major part of the CRE epidemic in Italy.A similar Susceptibility of carbapenem-non-susceptible Klebsiella pneumoniae to various carbapenems, Italy, 15 May-30 June 2011 (n=234) MIC: minimum inhibitory concentration; R: resistant; S: susceptible.MICs in mg/L, interpreted according to EUCAST breakpoints[26].Percentage of isolates with intermediate susceptibility not shown in the table.