Impact of changes in CLSI and EUCAST breakpoints for susceptibility in bloodstream infections due to extended-spectrum beta-lactamase-producing Escherichia coli

Impact of changes in CLSI and EUCAST breakpoints for susceptibility in bloodstream infections due to extended-spectrum b-lactamase- producing Escherichia coli J. Rodrı ´guez-Ban ˜o 1,2 , E. Pico ´n 1 , M. D. Navarro 1 ,L.Lo ´pez-Cerero 1 ,A ´. Pascual 1,3 and the ESBL-REIPI Group* 1) Unidad Clı ´nica de Enfermedades Infecciosas y Microbiologı ´a, Hospital Universitario Virgen Macarena, Sevilla, 2) Departamento de Medicina, Universidad de Sevilla, 3) Departamento de Microbiologı ´a, Universidad de Sevilla, Sevilla, Spain Abstract The impact of recent changes in and discrepancies between the breakpoints for cephalosporins and other antimicrobials, as determined by CLSI and European Committee on Antimicrobial Susceptibility Testing (EUCAST), was analysed in patients with bloodstream infections caused by extended-spectrum b-lactamase (ESBL) producing Escherichia coli in Spain, was analysed. We studied a cohort of 191 episodes of bloodstream infection caused by ESBL-producing E. coli in 13 Spanish hospitals; the susceptibility of isolates to different antimicrobials was investigated by microdilution and interpreted according to recommendations established in 2009 and 2010 by CLSI, and in 2011 by EUCAST. Overall, 58.6% and 14.7% of isolates were susceptible to ceftazidime, and 35.1% and 14.7% to cefepime using the CLSI-2010 and EUCAST-2009/2011 recommendations, respectively (all isolates would have been considered resistant using the previous guidelines). Discrepancies between the CLSI-2010 and the EUCAST-2011 recommendations were statistically significant for other antimicrobials only in the case of amikacin (98.4% versus 75.9% of susceptible isolates; p <0.01). The results varied depending on the ESBL produced. No significant differences were found in the percentage of patients classified as receiving appropriate therapy, following the different recommendations. Four out of 11 patients treated with active cephalosporins according to CLSI-2010 guidelines died (all had severe sepsis or shock); these cases would have been considered resistant according to EUCAST-2011. In conclusion, by using current breakpoints, extended-spectrum cephalosporins would be regarded as active agents for treating a significant proportion of patients with bloodstream infections caused by ESBL-producing E. coli. Keywords: Bloodstream infections, breakpoints, cephalosporins, Escherichia coli, extended-spectrum b-lactamases, therapy Original Submission: 10 July 2011; Revised Submission: 5 September 2011; Accepted: 6 September 2011 Editor: F. Allerberger Article published online: 14 September 2011 Clin Microbiol Infect 2012; 18: 894–900 10.1111/j.1469-0691.2011.03673.x Corresponding author: J. Rodrı ´guez-Ban ˜o, Seccio ´n de Enfermedades Infecciosas, Hospital Universitario Virgen Macarena, Avda Dr Fedriani 3, 41009 Seville, Spain E-mail: [email protected] *ESBL-REIPI Group members are in Appendix. Introduction The breakpoints for classifying organisms as susceptible or resistant to different antimicrobial agents, as determined by the CLSI and European Committee on Antimicrobial Susceptibility Testing (EUCAST), have an extraordinary influence on the surveillance of antimicrobial resistance as well as on the treatment of infections worldwide. Hence, it is important to analyse the impact of discrepancies and changes in breakpoints recommended by these organizations because any change or discrepancy could be significant for bacteria producing specific mechanisms of resistance, such as extended-spectrum b-lactamases (ESBLs), as the therapeutic options for infections caused by these isolates are limited. In 2010, the CLSI changed the susceptibility breakpoints for the Enterobacteriaceae from £8 mg/L to £1 in the case of cefotaxime, and from £8 mg/L to £4 in the case of ceftazidime; in addition, the interpretation of the breakpoint was to be reported as found, irrespective of whether there was ª2011 The Authors Clinical Microbiology and Infection ª2011 European Society of Clinical Microbiology and Infectious Diseases ORIGINAL ARTICLE BACTERIOLOGY ESBL production [1]. Previously, ESBL-producers should be reported as resistant to cephalosporins, regardless of the MIC [2]. A similar recommendation had previously been launched by EUCAST in 2009 and maintained in 2011, although breakpoints for cefotaxime, ceftazidime and cefepime were established at £1 mg/L [3]. The breakpoints for carbapenems, determined by the CLSI in June 2010, also changed from £4 mg/L to £1 for imipenem or meropenem, and from £2 mg/L to £0.25 for ertapenem [1,2]. The current breakpoints fixed by EUCAST are £2 for imipenem and meropenem, and £0.5 for ertapenem [3]. Finally, there are also differences in the breakpoints for piperacillin-tazobactam between CLSI and EUCAST (£16 mg/L and £8, respectively), ciprofloxacin (£1 and £0.5), gentamicin/tobramycin (£4 and £2) and amikacin (£16 versus £8) [1,3]. The objective of this study was to analyse the impact of recent changes to and discrepancies between the CLSI and EUCAST breakpoints in patients with bloodstream infections caused by ESBL-produc- ing Escherichia coli. Materials and Methods All 191 isolates from two concurrent, multicentre, prospective cohorts of community-onset and nosocomial bloodstream infections caused by ESBL-producing E. coli were studied; the methodology, epidemiology, and clinical and microbiological features of both cohorts have been previously published [4,5]. Briefly, the study included all cases of community-onset and nosocomial bloodstream infection caused by ESBL-producing E. coli detected in 13 tertiary Spanish hospitals belonging to the Spanish Network for Research in Infectious Diseases (REIPI) between October 2004 and January 2006. Patients were prospectively followed for 30 days after bloodstream infection. Demographics, predisposing factors, source, clinical severity, antimicrobial therapy and outcome were recorded. The first blood isolate of each case was sent to a reference laboratory (Servicio de Microbiologı ´a, Hospital Universitario Virgen Macarena, Sevilla) where identification to species level was confirmed using the API 20E system (bioMe ´rieux, Lyon, France), and susceptibility results, ESBL characterization (by PCR testing of bla genes, and sequencing) and clonal typing of isolates (by using the repetitive extragenic palindromic-PCR method) were performed. Susceptibility testing was performed by microdilution, following CLSI and EUCAST recommendations [1,3]; data on cefotaxime, ceftazidime, cefepime, amoxicillinclavulanic acid, piperacillin-tazobactam, imipenem, meropenem, ertapenem, ciprofloxacin, gentamicin, tobramycin and amikacin are shown. To further evaluate the clinical data, we also used the dosing-dependent pharmacokinetic–pharmacodynamic breakpoints proposed by Frei et al. [6]. The study was approved by the local Ethics committees of the participating hospitals. The ESBLs produced by the isolates were previously reported separately for community and nosocomial onset bloodstream infections [4,5]; combined data were as follows (seven isolates produced two ESBLs): CTX-M, 161 isolates (84%), including 122 isolates producing the cephotaxime CTX-M-9 group enzymes (97 CTX-M-14 and 25 CTX-M-9), 42 CTX-M-1 enzymes (23 CTX-M-15, 15 CTX-M-32, 3 cefotaxime-M-1, and 1 CTX-M-19); SHV, 33 isolates (17%), including 30 isolates producing SHV-12 and 3 SHV-2a; and 1 producing a TEM ESBL enzyme (TEM-52). Also, clonal typing of the isolates showed that 185 different repetitive extragenic palindromic-PCR profiles were found among the 191 isolates included. For the purposes of this analysis, isolates were considered susceptible, intermediate or resistant using the CLSI-2009 guidelines [2], the CLSI update in 2010 [1], and EUCAST- 2011 recommendations [3]. Percentages were compared using the chi-square test or the Fisher’s exact test, as appropriate, and MICs using the Mann–Whitney Utest. Results Antimicrobial susceptibility data for the 191 isolates according to the CLSI-2009, CLSI-2010 and EUCAST-2011 recommendations are shown in Table 1. Using the CLSI-2009 criteria, all isolates were classified as resistant to cefotaxime, ceftazidime and cefepime irrespective of the MIC, because an ESBL was produced. For cefotaxime, no isolate would continue to be regarded as susceptible using the CLSI-2011 and EUCAST-2011 recommendations, although two isolates (1%) producing a CTX-M-9 group enzyme and a CTX-M-1 group enzyme, respectively, would be considered intermediate (MIC = 2 mg/L). For ceftazidime and cefepime, the CLSI- 2010 guidelines classified isolates as susceptible more often than the EUCAST-2011 recommendations; this is particularly relevant for isolates producing an ESBL from the CTX-M-9 group, in the case of ceftazidime and cefepime (88.8% and 21.1% were susceptible according to CLSI-2010, compared with 23.3% and 2.6% according to EUCAST-2011), and for isolates producing an ESBL in the SHV group, in the case of cefepime (89.7% susceptible according to CLSI-2010, against 65.5% according to EUCAST-2011). With regard to carbapenems, all isolates were considered susceptible to ertapenem using CLSI-2009 and EUCAST- 2011 recommendations; however, using the modified CLSICMI Rodrı ´guez-Ban ˜oet al. Breakpoints in ESBL-producing E. coli bacteraemia 895 ª2011 The Authors Clinical Microbiology and Infection ª2011 European Society of Clinical Microbiology and Infectious Diseases, CMI,18, 894–900 2010 guidelines, 187 isolates (97.7%) would continue to be considered susceptible to ertapenem, but four isolates would be classified as intermediate (two produced a CTX-M-9 group enzyme, two produced a CTX-M-1 group enzyme). All isolates were susceptible to imipenem and meropenem for all three sets of recommendations. CLSI breakpoints for amoxicillin-clavulanic acid, piperacillin-tazobactam, ciprofloxacin and aminoglycosides did not change in CLSI-2010 recommendations with respect to 2009. There were no significant differences of susceptibility to ciprofloxacin and gentamicin between CLSI-2010 and EUCAST-2011. The percentage of isolates susceptible to piperacillin-tazobactam, tobramycin and amikacin was somewhat higher using CLSI-2010 criteria rather than EUCAST-2011 (p values for comparisons: 0.07, 0.07, and 0.06, respectively). The MIC 50 and MIC 90 values of the antimicrobials studied, by ESBL group, are shown in Table 1. Overall, MICs for isolates producing enzymes from the CTX-M-9 group were lower than for the CTX-M-1 group in the case of ceftazidime (p <0.001), amoxicillin-clavulanic acid p 0.004), piperacillin-ta- zobactam (p <0.001), ciprofloxacin (p <0.001), tobramycin (p <0.001), and amikacin (p <0.001). There were no significant differences in the cases of cefotaxime, cefepime, ertapenem and gentamicin. In addition, the CTX-M-9 group showed higher MICs against ceftazidime than the SHV group (p <0.001), but a lower MIC against cefotaxime (p <0.001) and cefepime (p <0.001). MIC distributions for cefotaxime, ceftazidime, cefepime, piperacillin-tazobactam, amoxicillin-cla- vulanic acid and ertapenem, by group of ESBL produced, are shown in Fig.1. TABLE 1. Susceptibilities of extended-spectrum b-lactamase-producing Escherichia coli isolates to relevant antimicrobials, by CLSI-2009, CLSI-2010 and EUCAST-2011 criteria, and by ESBL group produced. Data are expressed as number of susceptible isolates (percentage in each row). All 191 isolates were included in the totals. For CTX-M-9, CTX-M-1, and SHV groups, only isolates producing one ESBL were included (116, 38, and 29 isolates, respectively) Antimicrobial Isolates MIC 50 –MIC 90 (mg/L) CLSI-2009 CLSI-2010 EUCAST- 2011 Cefotaxime Total 256–256 0 0 0 CTX-M-9 group 256–256 0 0 0 CTX-M-1 group 256–256 0 0 0 SHV group 32–256 0 0 0 Ceftazidime Total 2–8 0 112 (58.6) a 28 (14.7) d CTX-M-9 group 2–8 0 103 (88.8) a 27 (23.3) d CTX-M-1 group 32–128 0 8 (21.1) b 1 (2.6) e SHV group 64–128 0 0 0 Cefepime Total 64–64 0 67 (35.1) a 28 (14.7) d CTX-M-9 group 64–64 0 31 (26.7) a 3 (2.6) d CTX-M-1 group 64–64 0 8 (21.1) b 4 (10.5) SHV group 1–16 0 26 (89.7) a 19 (65.5) f Amoxicillin/clavulanic acid Total 8–16 NC 118 (61.8) 118 (61.8)* CTX-M-9 group 8–16 NC 41 (35.3) 41 (35.3)* CTX-M-1 group 16–32 NC 23 (60.5) 23 (60.5)* SHV group 8–16 NC 6 (20.7) 6 (20.7)* Piperacillin/tazobactam Total 2–8 NC 175 (91.6) 164 (85.9) g CTX-M-9 group 2–8 NC 110 (94.8) 108 (93.1) CTX-M-1 group 8–32 NC 31 (81.6) 23 (60.5) h SHV group 2–8 NC 27 (93.1) 27 (93.1) Ertapenem Total 0.03–0.125 191 (100) 187 (97.9) c 191 (100) i CTX-M-9 group 0.03–0.125 116 (100) 114 (98.3) 116 (100) CTX-M-1 group 0.03–0.125 38 (100) 36 (94.7) 38 (100) SHV group 0.015–0.125 29 (100) 29 (100) 29 (100) Ciprofloxacin Total 4–32 NC 62 (32.5) 59 (30.9) CTX-M-9 group 4–32 NC 40 (34.5) 38 (32.8) CTX-M-1 group 64–128 NC 6 (15.8) 6 (15.8) SHV group 4–32 NC 12 (41.4) 11 (37.9) Gentamicin Total 1–32 NC 152 (79.6) 145 (75.9) CTX-M-9 group 1–32 NC 96 (82.8) 89 (76.7) CTX-M-1 group 2–128 NC 27 (71.1) 27 (71.1) SHV group 1–64 NC 23 (79.3) 23 (79.3) Tobramycin Total 1–4 NC 175 (91.6) 164 (85.9) j CTX-M-9 group 1–4 NC 106 (91.4) 102 (87.9) CTX-M-1 group 8–64 NC 18 (47.4) 17 (44.7) SHV group 1–4 NC 27 (93.1) 24 (82.8) Amikacin Total 4–8 NC 188 (98.4) 171 (75.9) d CTX-M-9 group 4–8 NC 116 (100) 113 (97.4) CTX-M-1 group 4–16 NC 35 (92.1) 23 (60.5) d SHV group 4–8 NC 29 (100) 28 (96.6) *Not resistant. NC: no change in breakpoints in the 2010 version. The p values for comparison between CLSI-2010 and CLSI-2009 (chi-square except where specified). a £0.001; b 0.002 (Fisher’s test); c 0.06(Fisher’s test). The p values for comparing CLSI-2010 and EUCAST-2011 (chi-square except where specified): d £0.001; e 0.007; f 0.02; g 0.07; h 0.04; i 0.06 (Fisher’s test); j 0.07. Remaining comparisons >0.1. 896 Clinical Microbiology and Infection, Volume 18 Number 9, September 2012 CMI ª2011 The Authors Clinical Microbiology and Infection ª2011 European Society of Clinical Microbiology and Infectious Diseases, CMI,18, 894–900 As regards clinical impact, 132 patients received empirical monotherapy (35 with amoxicillin-clavulanic acid (18.3% of the whole series), 23 with cefotaxime (12%), 21 with imipenem or meropenem (10.9%), 17 with piperacillin-tazobactam (9.8%), 13 with ciprofloxacin or levofloxacin (6.8%), four with cefepime (2%), one with ceftazidime or amikacin (0.5%), and 17 with miscellaneous agents); 50 received combined therapy (18 with a b-lactam plus an aminoglycoside (9.4%), 14 with a b-lactam plus a fluoroquinolone (7.3%), and 18 with other combinations), and nine did not receive empirical therapy. Empirical therapy would be considered appropriate in 102 (53.4%) episodes by CLSI-2009 criteria, and in 106 (55.5%) by CLSI-2010 criteria (p 0.6). According to EUCAST-2011, empirical therapy was appropriate in 97 patients (50.8%) (p value for comparison with CLSI-2010; p 0.3). Taking only the 133 patients treated with antimicrobials who showed some difference between CLSI-2010 and EUCAST-2011 recommendations (ceftazidime, cefepime, piperacillin-tazobactam, carbapenems, ciprofloxacin, and aminoglycosides), 80 and 71 patients, respectively, received appropriate empirical therapy (60.2% versus 53.4%; p 0.2). There was no difference in mortality for patients receiving inappropriate empirical therapy according to CLSI-2010 or EUCAST-2011 breakpoints (16/53 (30.2%) versus 18/62 (29%), respectively; p 0.8). Fifteen patients had an isolate for which the MIC to the empirically used antibiotic showed a classification discrepancy depending on whether the CLSI-2009, CLSI-2010 or EUCAST-2011 recommendations were used. Data and characteristics are summarized in Table 2. Six patients (40%) died (five of these with severe sepsis or shock at presentation, plus one other). Among the cephalosporins, 11 patients received appropriate therapy according to CLSI-2010 recommendations (none would be considered appropriate by EUCAST-2011); four died (36.3%). As a reference, two out of the 29 patients empirically treated with a carbapenem died (9.5%); the p value for the comparison (Fisher test, two-tailed) is 0.1; of note, none of the eight patients presenting with severe sepsis or shock treated with a carbapenem died. If recently proposed pharmacokinetic–pharmacodynamic breakpoints (which are dependent on the dosing regimen) are applied, the mortality rate in the patients treated 0 10 20 30 40 50 60 70 80 90 100 <1 1 2 4 8 16 32 64 128 >128 0 5 10 15 20 25 30 35 40 45 50 <1 1 2 4 8 16 32 64 128 >128 0 10 20 30 40 50 60 70 <1 1 2 4 8 16 32 64 128 >128 Cefotaxime Ceftazidime Cefepime 0 5 10 15 20 25 <1 1 2 4 8 16 32 64 128 >128 0 10 20 30 40 50 60 70 <1 1 2 4 8 16 32 64 128 >128 0 5 10 15 20 25 30 35 40 45 50 <0.030 0.03 0.06 0.125 0.25 0.5 1 >1 Piperacillin/tazobactam Amoxicillin/clavulanic acid Ertapenem FIG. 1. Distribution of isolates according to minimal inhibitory concentrations (MIC) of cefotaxime, ceftazidime, cefepime, piperacillin/tazobactam, amoxicillin/clavulanic acid and ertapenem, and by extended-spectrum b-lactamase (ESBL) group produced. Black bars, CTX-M-9 group; grey bars, CTX-M-1 group; white bars, SHV group. Only isolates producing one ESBL are included. The most frequent ESBL per group was: CTX-M- 14 in the CTX-M-9 group; CTX-M-15 in the CTX-M-1 group; SHV-12 in the SHV group. It should be noticed that the MIC scale differs for ertapenem. CMI Rodrı ´guez-Ban ˜oet al. Breakpoints in ESBL-producing E. coli bacteraemia 897 ª2011 The Authors Clinical Microbiology and Infection ª2011 European Society of Clinical Microbiology and Infectious Diseases, CMI,18, 894–900 with a cephalosporin and infected with organisms showing a MIC below or equal that pharmacokinetic–pharmacodynamic breakpoint was 14.2% (one out of seven patients), compared with 100% (four out of four) in those showing a higher MIC (p 0.01 by Fisher test). Four patients were treated with piperacillin-tazobactam, with an MIC of 16 mg/L (susceptible according to CLSI-2010, non-susceptible for EUCAST-2011); one died. Discussion Our results show that the CLSI-2010 recommendations for breakpoints for extended-spectrum cephalosporins and their interpretation have a significant impact on whether an invasive ESBL-producing E. coli is classified as susceptible to extendedspectrum cephalosporins. Hence, 58% and 35% of isolates would now be considered susceptible to ceftazidime and cefepime, and none to cefotaxime. In a collection of 21 ESBL-pro- ducing E. coli isolated from complicated intra-abdominal infections in the USA, 38%, 52% and 48% were considered susceptible to cefotaxime, ceftazidime and cefepime, respectively [7]; figures from the Asia-Pacific region were 0%, 19%, and 5%, respectively [8]. Our study shows that a change in breakpoint varies in impact depending on the ESBL group produced, so discrepancies between studies might simply be attributable to differences in regional prevalence of the ESBL type. The recommendation that interpretation should be reported according to MIC regardless of ESBL production has less impact using EUCAST-2011 breakpoints than CLSI- 2010, because the EUCAST-2011 breakpoints are lower for ceftazidime and cefepime. So, only 14.7% of isolates in our series would be considered susceptible to either ceftazidime or cefepime with EUCAST-2011 recommendations. In a multi-country European surveillance study for recently isolated ESBL-producing E. coli in complicated intra-abdominal infections, susceptibility to cefotaxime, ceftazidime and cefepime was 2.3%, 7.5% and 8.7%, respectively [9], with subtle differences depending on whether the infection source was hospital or community-onset [10]. Just as the susceptibility profiles for the isolates varied depending on which ESBL group was produced, so the impact of the different breakpoints varied accordingly. This was more evident in isolates producing CTX-M-9 group enzymes (mostly CTX-M-14) in the case of ceftazidime, and in isolates producing an SHV group ESBL (mostly SHV-12) in the case of cefepime. Furthermore, discrepancies between CLSI-2010 and EUCAST-2011 recommendations caused differences that were almost significant in the percentage of isolates susceptible to piperacillin-tazobactam (particularly TABLE 2. Features of patients whose isolates would be classified differently, according to CLSI-2009, CLSI-2010 and EUCAST-2011 recommendations Patient number Age (years), gender Underlying conditions Source of bloodstream infection Acquisition Severe sepsis/ shock Empirical therapy, dose (MIC, mg/L) MIC (mg/L), empirical antimicrobial Definitive therapy Outcome 1 83, female Gallstones Biliary tract Nosocomial Yes Ceftriaxone, 2 g/24 h 2 – Died (day 2) 2 63, female Cancer Urinary tract Nosocomial No Ceftazidime, 2 g/8 h 2 a,c Imipenem Cured 3 74, male None Urinary tract Community No Ceftazidime, 1 g/8 h 2 a,c Imipenem Cured 4 42, male Cancer Urinary tract Community No Cefepime, 2 g/12 h 1 a,b,c Amoxicillin/ clavulanic acid Cured 5 71, male Liver cirrhosis Cancer Primary Community Yes Cefepime, 1 g/12 h 2 a – Died (day 2) 6 53, male Cancer Primary Nosocomial No Cefepime, 2 g/12 h 2 a,c Imipenem Cured 7 61, male Cancer Pneumonia Nosocomial No Cefepime, 2 g/12 h 2 a,c Imipenem Cured 8 41, male Liver cirrhosis, HIV infection Primary Nosocomial No Cefepime, 2 g/12 h 2 a,c Imipenem Cured 9 74, male Liver cirrhosis, Cancer Biliary tract Nosocomial Yes Cefepime, 2 g/12 h 4 a, Died (day 1) 10 73, male Diabetes, Heart failure Urinary tract Nosocomial Yes Cefepime, 2 g/12 h 4 a Imipenem Died (day 8) 11 80, male None Pneumonia Nosocomial (ICU) Yes Cefepime 1 g/12 h (renal insufficiency) 1 a,b,c – Died (day 1) 12 57, male Liver cirrhosis, Cancer Primary peritonitis Community No Piperacillin/tazobactam 4 g/6 h 16 a Imipenem Died (day 11) 13 41, male Cerebrovascular disease Urinary tract Nosocomial No Piperacillin/tazobactam 4 g/6 h 16 a Imipenem Cured 14 87, male Heart failure, Cancer Biliary tract Community No Piperacillin/tazobactam 4 g/8 h 16 a Ertapenem Cured 15 57, male Cancer, Liver cirrhosis Primary peritonitis Community No Piperacillin/tazobactam 4 g/6 h 16 a Ertapenem Cured Isolate susceptible according to: a CLSI-2010, b EUCAST-2011, and c PK/PD breakpoints proposed in reference 6. 898 Clinical Microbiology and Infection, Volume 18 Number 9, September 2012 CMI ª2011 The Authors Clinical Microbiology and Infection ª2011 European Society of Clinical Microbiology and Infectious Diseases, CMI,18, 894–900 among CTX-M-1 group producers, which were more frequently resistant using EUCAST-2011 breakpoints), tobramycin and amikacin. No differences were found for ciprofloxacin or gentamicin. Overall, the MICs of most antimicrobials against isolates producing an ESBL in the CTX-M-1 group (mainly CTX-M-15) were the same as or higher than those for isolates producing an ESBL in the CTX-M-9 or SHV groups. Johnson et al. [11] also found that CTX-M-15-producing E. coli isolated in the USA showed a higher resistance score than those producing other ESBLs. This could be related to the fact that mobile genetic elements harbouring bla CTX-M-15 genes frequently harbour other resistance genes too, such as bla OXA-1 (which confers resistance to b-lactamase inhibitors) or aac(6¢)-Ib-cr (which affects aminoglycosides and fluoroquinolones) in the ST 131 clonal group, which are associated with the presence of chromosomal mutations affecting fluoroquinolones [11–13]. So, the type of ESBL produced should be considered when comparing the percentage of susceptible isolates in studies from different areas or using different recommendations. Differences in breakpoint may also have important clinical implications. Although the use of extended-spectrum cephalosporins had been discouraged for some time regardless of MIC [14], these antimicrobials would now be considered appropriate by current CLSI and EUCAST recommendations provided that the MIC is low enough, according to some pharmacokinetic–pharmacodynamic data [15,16] and small case series [17]. Based on those recommendations, our data indicate that cephalosporins would be considered appropriate in a significant number of cases. However, there is still controversy about the efficacy of cephalosporins, even for isolates with a very low MIC because of the marked inoculum effect in vitro [18], and because the MIC could increase as the result of b-lactamase hyperproduction [19] or porin loss [20]. Whether these aspects are associated with an increased risk of failure during cephalosporin therapy for ESBL-producing isolates with a low MIC has not been adequately studied. In our study, the number of patients treated with a low-MIC cephalosporin was low, probably reflecting adherence to previous recommendations. However, we think it is important to report all such patients, so that cumulative experience can be analysed in greater detail in the future. We found a high mortality rate in patients with a bloodstream infection caused by ESBL-producing E. coli and treated with a cephalosporin which would be considered as susceptible to ceftazidime or cefepime according to present CLSI criteria. Most deaths occurred in patients with a severe presentation. The importance of the dosing regimen is to be taken into account. Although the dosing regimens were correct according to the approved labels of the drugs except in one patient, they might have been insufficient for some of these organisms with borderline MICs, as suggested by the data considering the dose-dependent pharmacokinetic–pharmacodynamic breakpoints [6]. Hence, even though the data should be interpreted with prudence because the number of cases is low, we would suggest caution in the use of cephalosporins, at least in patients with a severe presentation or with infections associated with a high bacterial load [21], and, if cephalosporins are used for ESBL-producers, a dosing regimen that maximizes the probability of attaining the pharmacokinetic–pharmacodynamic target considering the MIC should be prescribed. Our data did not enable us to analyse the clinical impact of discrepancies between CLSI-2010 and EUCAST-2011 for breakpoints of other antimicrobials. In conclusion, extended-spectrum cephalosporins would be considered active agents for treating a significant proportion of patients with bloodstream infections caused by ESBL- producing E. coli using the present breakpoints launched by EUCAST and, particularly, CLSI. The proportion of susceptible isolates varies by type of ESBL produced. More clinical data are necessary to support the present EUCAST and CLSI recommendations for cephalosporin susceptibility for ESBL- producers in different types of infections, but at present, EUCAST breakpoints seem safer in terms of clinical application. There are also some significant discrepancies between CLSI and EUCAST when applied to other antimicrobials, although there are insufficient clinical data to evaluate the impact of such differences. Ackowledgements This study was funded by the Ministerio de Ciencia e Innovacio ´n, Instituto de Salud Carlos III—co-financed by European Development Regional Fund ‘A way to achieve Europe’ ERDF, Spanish Network for Research in Infectious Diseases (REIPI RD06/0008), Fondo de Investigacio ´n Sanitaria (grants 070190, 10/02021, and 10/01955), and Junta de Andalucı ´a (grants 0063/2006, 0048/2008, and CTS-5259). The funders had no role in the design, analysis, writing the manuscript and decision to publish. We thank Rafael Canto ´n for sharing some ideas about the objective of this manuscript. Transparency Declarations J. Rodrı ´guez-Ban ˜o has been a consultant for Wyeth, Merck, and Pfizer, has served as speaker for Wyeth, Merck, Pfizer, Astra-Zeneca and GlaxoSmithKline, and has received CMI Rodrı ´guez-Ban ˜oet al. Breakpoints in ESBL-producing E. coli bacteraemia 899 ª2011 The Authors Clinical Microbiology and Infection ª2011 European Society of Clinical Microbiology and Infectious Diseases, CMI,18, 894–900 research support from Merck and Wyeth. A. Pascual has been a consultant for Merck and Pfizer, has served as speaker for Wyeth, Astra-Zeneca, Merck and Pfizer and has received research support from Merck and Pfizer and Wyeth. M.D. Navarro, L. Lo ´pez-Cerero, and E. Pico ´n had no conflict of interest. Appendix *Other participants from the ESBL-REIPI/GEIH group are: Paloma Gijo ´n (Hospital Universitario Gregorio Maran ˜o ´n, Madrid), Jose ´Ramo ´n Herna ´ndez (Hospital Universitario Virgen Macarena, Sevilla), Jose M. Cisneros (Hospital Universitario Virgen del Rocı ´o, Sevilla), Carmen Pen ˜a (Hospital Universitario de Bellvitge, Barcelona), Manuel Almela (Hospital Clinic, Barcelona), Benito Almirante (Hospital Universitario Vall d’Hebro ´n, Barcelona), Fabio Grill (Hospital Universitario Ramo ´n y Cajal, Madrid; present address, Hospital Universitario La Paz, Madrid), Javier Colomina (Hospital de la Ribera, Alzira, Valencia), Monserrat Gime ´nez (Hospital Germans Trias i Pujol, Badalona), Antonio Oliver (Hospital Son Espases, Palma de Mallorca), Juan Pablo Horcajada (Hospital Universitario Marque ´s de Valdecilla, Santander; present address, Hospital del Mar, Barcelona), Gemma Navarro (Corporacio Sanitaria Parc Taulı ´, Sabadell), Ana Coloma (Hospital Santa Creu i San Pau, Barcelona). References 1. Clinical and Laboratory Standard Institute (CLSI). Performance Standards for Antimicrobial Susceptibility Testing; 20th Informational Supplement (June 2010 Update) CLSI document M100-S20-U. Wayne, PA, USA: CLSI, 2010. 2. Clinical and Laboratory Standard Institute (CLSI). Performance Standards for Antimicrobial Susceptibility Testing; 19th Informational Supplement CLSI document M100-S19. Wayne, PA, USA: CLSI, 2009. 3. European Committee on Antimicrobial Susceptibility Testing (EUCAST). http://www.eucast.org. Accessed 5 September 2011. 4. Rodrı ´guez-Ban ˜o J, Pico ´n E, Gijo ´nPet al. Community-onset bacteremia due to extended-spectrum b-lactamase-producing Escherichia coli: risk factors and prognosis. Clin Infect Dis 2010; 50: 40–48. 5. Rodrı ´guez-Ban ˜o J, Pico ´n E, Gijo ´nPet al. Risk factors and prognosis of nosocomial bloodstream infections caused by extended-spectrum b-lactamase-producing Escherichia coli.J Clin Microbiol 2010; 48: 1726– 1731. 6. Frei CR, Wiederhold NP, Burgess DS. Antimicrobial breakpoints for Gram-negative aerobic bacteria based on pharmacokinetic-pharmaco- dynamic models with Monte Carlo simulation. J Antimicrob Chemother 2008; 61: 621–628. 7. Hoban DJ, Bouchillon SK, Hawser SP, Badal RE, LaBombardi VJ, Di Persio J. Susceptibility of gram-negative pathogens isolated from patients with complicated intra-abdominal infections in the United States, 2007-2008: results of the Study for Monitoring Antimicrobial Resistance Trends (SMART). Antimicrob Agents Chemother 2010; 54: 3031–3034. 8. Chen YH, Hsueh PR, Badal RE et al. Antimicrobial susceptibility profiles of aerobic and facultative Gram-negative bacilli isolated from patients with intra-abdominal infections in the Asia–Pacific region according to currently established susceptibility interpretive criteria. J Infect 2011; 62: 280–291. 9. Hawser S, Hoban D, Bouchillon S, Badal R, Carmeli Y, Hawkey P. Antimicrobial susceptibility of intra-abdominal Gram-negative bacilli from Europe: SMART Europe 2008. Eur J Clin Microbiol Infect Dis 2011; 30: 173–179. 10. Hawser SP, Bouchillon SK, Hoban DJ, Badal RE, Canto ´n R, Baquero F. Incidence and antimicrobial susceptibility of Escherichia coli and Klebsiella pneumoniae with extended-spectrum b-lactamases in community- and hospital-associated intra-abdominal infections in Europe: results of the 2008 Study for Monitoring Antimicrobial Resistance Trends (SMART). Antimicrob Agents Chemother 2010; 54: 3043–3046. 11. Johnson JR, Johnston B, Clabots C, Kuskowski MA, Castanheira M. Escherichia coli sequence type ST131 as the major cause of serious multidrug-resistant E. coli infections in the United States. Clin Infect Dis 2010; 51: 286–294. 12. Coque TM, Novais A, Carattoli A et al. Dissemination of clonally related Escherichia coli strains expressing extended-spectrum b-lac- tamase CTX-M-15. Emerg Infect Dis 2008; 14: 195–200. 13. Pitout JD, Gregson DB, Campbell L, Laupland KB. Molecular characteristics of extended-spectrum-b-lactamase-producing Escherichia coli isolates causing bacteremia in the Calgary Health Region from 2000 to 2007: emergence of clone ST131 as a cause of communityacquired infections. Antimicrob Agents Chemother 2009; 53: 2846–2851. 14. Paterson DL, Bonomo RA. Extended-spectrum b-lactamases: a clinical update. Clin Microbiol Rev 2005; 18: 657–686. 15. Andes D, Craig WA. Treatment of infections with ESBL-producing organisms: pharmacokinetic and pharmacodynamic considerations. Clin Microbiol Infect 2005; 11 (Suppl. 6): 10–17. 16. MacGowan A. Breakpoints for extended-spectrum b-lactamase-pro- ducing Enterobacteriaceae: pharmacokinetic/pharmacodynamic considerations. Clin Microbiol Infect 2008; 14 (Suppl. 1): 166–168. 17. Bin C, Hui W, Renyuan Z et al. Outcome of cephalosporin treatment of bacteremia due to CTX-M type extended-spectrum b-lactamase- producing Escherichia coli.Diag Microbiol Infect Dis 2006; 56: 351–357. 18. Queenan AM, Foleno B, Gownley C, Wira E, Bush K. Effects of inoculum and b-lactamase activity in AmpC- and extended-spectrum b-lactamase (ESBL)-producing Escherichia coli and Klebsiella pneumoniae clinical isolates tested by using NCCLS ESBL methodology. J Clin Microbiol 2004; 42: 269–275. 19. Pałucha A, Mikiewicz B, Gniadkowski M. Diversification of Escherichia coli expressing an SHV-type extended-spectrum b-lactamase (ESBL) during a hospital outbreak: emergence of an ESBL-hyperproducing strain resistant to expanded-spectrum cephalosporins. Antimicrob Agents Chemother 1999; 43: 393–396. 20. Martı ´nez-Martı ´nez L. Extended-spectrum b-lactamases and the permeability barrier. Clin Microbiol Infect 2008; 14 (Suppl. 1): 82–89. 21. Rodrı ´guez-Ban ˜o J, Pascual A. Clinical significance of extended-spec- trum b-lactamases. Expert Rev Anti Infect Therapy 2008; 6: 671–683. 900 Clinical Microbiology and Infection, Volume 18 Number 9, September 2012 CMI ª2011 The Authors Clinical Microbiology and Infection ª2011 European Society of Clinical Microbiology and Infectious Diseases, CMI,18, 894–900