Performance of EUCAST and CLSI approaches for co-amoxiclav susceptibility testing conditions for clinical categorization of a collection of Escherichia coli isolates with characterized resistance phenotypes

Performance of EUCAST and CLSI approaches for co-amoxiclav susceptibility testing conditions for clinical categorization of a collection of Escherichia coli isolates with characterized resistance phenotypes Marı ´aDı ´ez-Aguilar1,2, Marı ´a-Isabel Morosini1,2, Lorena Lo ´pez-Cerero2,3,A ´lvaro Pascual2,3, Jorge Calvo2,4, Luis Martı ´nez-Martı ´nez2,4,5, Francesc Marco2,6, Jordi Vila2,6, Adriana Ortega2,7, Jesu ´s Oteo2,7 and Rafael Canto ´n1,2* 1 Servicio de Microbiologı ´a, Hospital Universitario Ramo ´n y Cajal and Instituto Ramo ´n y Cajal de Investigacio ´n Sanitaria (IRYCIS), Madrid, Spain; 2 Red Espan˜ola de Investigacio ´n en Patologı ´a Infecciosa (REIPI) Instituto de Salud Carlos III, Madrid, Spain; 3 Servicio de Microbiologı ´a, Hospital Universitario Virgen Macarena and Departamento de Microbiologı ´a, Universidad de Sevilla, Sevilla, Spain; 4 Servicio de Microbiologı ´a, Hospital Universitario Marque ´s de Valdecilla-IDIVAL, Santander, Spain; 5 Departamento de Biologı ´a Molecular, Universidad de Cantabria, Santander, Spain; 6 Servicio de Microbiologı ´a Clı ´nica, Hospital Clı ´nic, Facultad de Medicina and Centro de Investigacio ´n en Salud Internacional de Barcelona (CRESIB), Barcelona, Spain; 7 Laboratorio de Antibio ´ticos, Bacteriologı ´a, Centro Nacional de Microbiologı ´a, Instituto de Salud Carlos III, Majadahonda, Madrid, Spain *Corresponding author. Servicio de Microbiologı ´a, Hospital Ramo ´n y Cajal, 28034 Madrid, Spain. Tel: +34-913368832; E-mail: ra[email protected]g Received 5 January 2015; returned 17 February 2015; revised 12 March 2015; accepted 18 March 2015 Objectives: There are different methodological recommendations for in vitro testing of the co-amoxiclav combination. Performance of co-amoxiclav MIC testing for Escherichia coli by the standard ISO microdilution method (ISO 20776-1) was compared using EUCAST (fixed 2 mg/L clavulanate concentration) and CLSI (2:1 ratio) interpretive criteria. Methods: MICs were determined by broth microdilution using a 2:1 ratio and fixed clavulanate concentrations (2 and 4 mg/L) for 160 clinical E. coli isolates with characterized resistance mechanisms. Essential agreements, categorical agreements and relative errors were determined. Results: For all isolates, essential agreement between microdilution using 2 mg/L clavulanate and a 2:1 ratio was 25.6%. For ESBL-producing isolates, considering EUCAST breakpoints, 55% of isolates tested with 2 mg/L clavulanate were classified as resistant; conversely, 95% of isolates tested with 4 mg/L clavulanate were susceptible. When using CLSI breakpoints and a 2:1 ratio, 90% of isolates were susceptible and 10% were intermediate. Conclusions: Variation in the clavulanate concentration gave different susceptibility testing results, particularly among ESBL-producing E. coli isolates. The in vitro concentration of clavulanate that better correlates with clinical outcome is still under debate and should be established. Keywords: MICs, amoxicillin/clavulanate, E. coli Introduction Discrepant methodological and interpretive approaches for co-amoxiclav in vitro testing create uncertainty at the time of comparing results if either CLSI or EUCAST standards are adopted. 1 Co-amoxiclav constitutes a widely used therapeutic option against an extensive range of bacteria and many Escherichia coli infections are frequently treated with this association. However, co-amoxiclav resistance in E. coli is increasing in many countries, reinforcing the necessity of reliable susceptibility testing data. 2,3 Two different recommendations exist for testing co-amoxiclav in Enterobacteriaceae: EUCAST indicates the use of a fixed 2 mg/L clavulanate concentration while CLSI recommends a 2:1 ratio. 4,5 Differences between both methodologies and breakpoint values lead to dissimilar clinical interpretation. 1 EUCAST has recently established specific breakpoints for uncomplicated urinary tract infections [susceptible (S), ≤32 mg/L; resistant (R), .32 mg/L] that are different from those recommended for systemic infections (S, ≤8 mg/L; R, .8mg/L). 4 However, CLSI has the same breakpoints for all types of infections [S, ≤8 mg/L; intermediate (I), 16 mg/L; R, ≥32 mg/L]. These differences have generated debate about which are the most adequate interpretive criteria. 6 Our objective was to compare susceptibility testing results whenusinga2:1ratioofco-amoxiclavandfixedclavulanate #The Author 2015. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please e-mail: [email protected] J Antimicrob Chemother 2015; 70: 2306–2310 doi:10.1093/jac/dkv088 Advance Access publication 21 April 2015 2306 Downloaded from https://academic.oup.com/jac/article/70/8/2306/810536 by guest on 07 April 2025 concentrations of 2 and 4 mg/L by the standard ISO microdilution reference method. 7 Materials and methods A characterized collection of 160 E. coli clinical isolates was studied. 8 Twenty strains producing each of the following enzymatic co-amoxiclav resistance mechanisms were included: OXA-1, TEM-1 hyperproduction, OXA-1 plus CTX-M-15, plasmid AmpC (DHA-1and CMY-2), chromosomal AmpC hyperproduction and inhibitor-resistant TEM. A group of 20 isolates producing ESBLs (SHV-12, CTX-M-15, CTX-M-14, CTX-M-9 and CTX-M-32) and 20 b-lactam-susceptible isolates (WT phenotype) completed the study. 9 E. coli ATCC 25922 and ATCC 35218 were used as quality controls. Amoxicillin MICs (Sigma–Aldrich, St Louis, MO, USA) in the presence of lithium clavulanate (GlaxoSmithKline, Durham, UK) were determined by standard ISO broth microdilution (BBL-Mueller-Hinton-II cation-adjusted broth; Becton-Dickinson, Sparks, MD, USA) using a 2:1 co-amoxiclav ratio and amoxicillin in the presence of lithium clavulanate concentrations of 2 or 4 mg/L, respectively. Clavulanate MICs were also determined by the microdilution method. 7 Results were interpreted according to CLSI guidelines when the 2:1 co-amoxiclav ratio was used. EUCAST breakpoints were applied when 2 mg/L clavulanate concentrations were assayed and also when arbitrary 4 mg/L clavulanate concentrations were assayed. Data were also analysed according to the EUCAST urinary breakpoints for co-amoxiclav. 4 Essential agreement was calculated to compare performances of microdilution using 2 mg/L and 2:1 ratio against the total collection of isolates considering 2 mg/L clavulanate microdilution as the reference method.Comparisonbetweenmicrodilutionusing2and4mg/Lclavulanate and the 2:1 ratio, against ESBL-producing isolates, was performed considering 2 mg/L clavulanate microdilution as the reference. Comparison was also performed considering 4 mg/L clavulanate microdilution as a hypothetical reference method. Categorical agreement was calculated using EUCAST interpretive criteria. 4 Very major errors and major errors were recorded and validated according to published guidelines. 10 Error levels were computed as percentages along with their corresponding exact 95% CIs. Results When comparing the 2:1 co-amoxiclav ratio microdilution results with those of microdilution using 2 mg/L clavulanate, accuracy was very low (essential agreement, 25.6%). This value reflects the discrepancies between MICs when performing susceptibility testing methods following either CLSI or EUCASTcriteria. According to CLSI, MICs for resistant isolates with mechanisms affecting the activity of b-lactam/b-lactamase inhibitor combinations were much lower than those obtained according to EUCAST, since the 2:1 ratio allows the b-lactamase inhibitor to act as an antimicrobial agent at high concentrations. Clavulanate shows antimicrobial activity by itself against many Enterobacteriaceae and, in this study, such MICs ranged from 8 to 128 mg/L, with a median value of 32 mg/L (Figure S1, available as Supplementary data at JAC Online). Consequently, reported MICs reflect clavulanate intrinsic activity, particularly in highly amoxicillin-resistant isolates. 11 This can be observed in Figure 1, where the co-amoxiclav MIC distribution for resistant isolates is displaced to lowerconcentrations when susceptibility testing is performed with the 2:1 ratio when compared with 2 mg/L clavulanate. With 2 mg/L clavulanate (EUCAST criteria), 20.6% of isolates were susceptible and 79.4% were resistant. Conversely, 26.2% of isolates were susceptible, 60.6% were resistant and 13.2% had intermediate susceptibility with the 2:1 ratio (CLSI criteria) (Table S1). Results were also compared when using the 2:1 ratio and 2 and 4 mg/L clavulanate concentrations in ESBL-producing isolates. The rationale for using 4 mg/L was that maximum achievable serum levels of clavulanate after standard intravenous and oral doses of co-amoxiclav, at conventional intervals, are usually .2 mg/L. 12,13 Interestingly, among these isolates, essential agreement between results obtained with the 2:1 ratio andthosewiththefixed4and2mg/Lconcentrationswas 95% (75.1%–99.8%). However, remarkable discrepancies were obtained when comparing results using fixed 2 mg/L clavulanate >102410241 02481632 MIC (mg/L) 64 128 256 512 10 20 30 Number of isolates 40 50 60 70 Microdilution 2:1 Microdilution 2 mg/L 80 Figure 1. Distribution of co-amoxiclav MICs for total E. coli tested: microdilution with a 2:1 co-amoxiclav ratio (CLSI) and microdilution with fixed 2 mg/L clavulanate (EUCAST). Amoxicillin/clavulanate testing for E. coli 2307 JA C Downloaded from https://academic.oup.com/jac/article/70/8/2306/810536 by guest on 07 April 2025 with those using 4 mg/L clavulanate: essential agreement was 75% (50.9%–91.3%) and categorical agreement was 60% (36%–80.9%). Considering the 2 mg/L microdilution as reference, a very high level of very major errors was recorded [88.9% (57%– 99.4%)]. When 4 mg/Lwas considered as the reference method, a high rate of major errors [42.1% (20.2%–66.5%)] was obtained. These results indicate that a lower clavulanate concentration resulted in a higher rate of resistance. This can be observed in Figure 2. With EUCAST breakpoints, 55% of isolates tested with 2 mg/L clavulanate remained in the ‘resistant side’ of the distribution; however, 95% of isolates tested with 4 mg/L were displaced to the ‘susceptible side’ of the figure. When using CLSI breakpoints and the corresponding 2:1 ratio, 90% of isolates were susceptible and 10% intermediate, with no resistant isolates. According to the EUCAST urinary breakpoints, 100% of ESBL- producing isolates were susceptible with all clavulanate concentrations. Discussion Methodologies and interpretive criteria to be adopted when performing co-amoxiclav susceptibility testing in Enterobacteriaceae are still a matter of debate. Recent studies alert about the consequences of switching from CLSI to EUCAST when reporting co-amoxiclav susceptibility rates. 14–16 Although an increase in resistance has been detected, a better clinical correlation between patient outcome and EUCAST-measured MICs has been observed. 1 The inclusion of a large number of co-amoxiclav-resistant isolates in our study was justified by the increasing resistance to this combination among E. coli isolates in Spain, attributed mainly to selective antimicrobial pressure along with the spread of specific clones bearing resistance determinants such as OXA-1. 2,3,17 We observed a significant variation of resistance rates according to the adopted methodology. Discrepancies in ESBL-producing E. coli isolates’ susceptibility was remarkable since a high proportion of the isolates were classified as resistant using 2 mg/L clavulanate while most of the isolates were susceptible with a co-amoxiclav 2:1 ratio. Maximum clavulanate serum concentrations after oral and intravenous co-amoxiclav standard dosing regimens are usually .2 mg/L; therefore, the use of 2 mg/L in the in vitro studies may overestimate resistance rates, particularly in those ESBL- producing isolates that could be inhibited by clavulanate. 12,13 This fact was also partially addressed in a previous study with an E. coli collection, but with a lower number of resistant isolates. 18 To date, the pharmacodynamic target for clavulanate efficacy is unknown. However, according to pharmacokinetic/pharmacodynamic data, the efficacy of amoxicillin is determined by the percentage of time above the MIC of the unbound serum concentration [(fT .MIC )≥40]. The target attainment in the EUCAST rationale document for the standard dose of amoxicillin/clavulanate (1000/200 mg four times daily) is 100% when the MIC is ≤4 mg/L and 75% with MIC ≤8 mg/L (www.eucast.org). However, recent studies have reported lower target attainment values of 93% (MIC ≤4 mg/L) and 65% (MIC ≤8 mg/L), suggesting an increase in the standard dose to 1000/200 mg six times daily in severe Enterobacteriaceae sepsis or intra-abdominal infections. 19 b-Lactam/b-lactamase inhibitor combinations are an alternative to carbapenems for the treatment of infections caused by ESBL-producing E. coli isolates when interpreted as susceptible. 20 Although extensive clinical evaluation is needed, an insufficient clavulanate concentration in antimicrobial susceptibility testing studies could unnecessarily discard a suitable therapy. Consequently, susceptibility testing was also performed using 1 02481632 MIC (mg/L) 64 2 4 6 Number of isolates 8 10 12 Microdilution 2:1 EUCAST CLSI RIS Microdilution 2 mg/L 14 Microdilution 4 mg/L Figure 2. Co-amoxiclav MIC distribution for ESBL-producing E. coli tested according to microdilution susceptibility testing using a 2:1 co-amoxiclav ratio, fixed 2 mg/L clavulanate and fixed 4 mg/L clavulanate. The corresponding interpretive categories according to EUCAST and CLSI breakpoints are shown in the upper part of the figure. Dı ´ez-Aguilar et al. 2308 Downloaded from https://academic.oup.com/jac/article/70/8/2306/810536 by guest on 07 April 2025 4 mg/L clavulanate, a previously requested approach. 18 In our study, most ESBL-producing isolates became susceptible when using this concentration, resembling what happened when the 2:1 ratio was applied. In isolates with co-amoxiclav resistance mechanisms tested with the 2:1 ratio, relatively ‘unreal’ co-amoxiclav MIC values are reported since clavulanate is behaving as an antibiotic. This fact explains the very low percentage of essential agreement when using this ratio. EUCASTuncomplicated urinary tract infection breakpoints classified all the ESBL-producing isolates as susceptible, irrespective of the clavulanate concentration. This may support the use of co-amoxiclav for the treatment of urinary tract infections caused by ESBL-producing E. coli that are resistant to other oral agents. However, it is not currently possible to interpret urinary breakpoints through any automated susceptibility testing system, as their panels are not yet adapted to the required concentrations. Finally, susceptibility testing methods for co-amoxiclav should predict what might happen clinically. 20 However, variation in the clavulanate concentration when combined with amoxicillin has shown different results, particularly among ESBL-producing E. coli. Establishment of the optimal concentration of clavulanate when testing co-amoxiclav in vitro is still needed. Moreover, the corresponding correlation with clinical outcomes is not yet well documented. Our approach should help in the selection of the correct b-lactam inhibitor concentration and in the better definition of clinical breakpoints. Acknowledgements We are grateful to the ‘Spanish Group for the Study of Co-amoxiclav Resistance in Escherichia coli’ for providing the bacterial isolates and supporting the study: Marı ´a Merino-Carballeira and Germa ´n Bou (Hospital Universitario A Corun˜a), Beatriz Mirelis and Ferran Navarro (Hospital de la Santa Creu i Sant Pau, Barcelona), Juan Jose ´Gonza ´lez-Lo ´pez (Hospital Vall d’Hebron, Barcelona), Emilia Cercenado Mansilla (Hospital Universitario Gregorio Maran˜o ´n, Madrid), Antonio Oliver (Hospital Universitario Son Espases, Palma de Mallorca), Jorge Calvo and Luis Martı ´nez-Martı ´nez (Hospital Universitario Marque ´s de Valdecilla, Santander), Lorena Lo ´pez-Cerero and A ´lvaro Pascual (Hospital Universitario Virgen Macarena, Sevilla) and Jesu ´s Oteo and Jose ´Campos (Instituto de Salud Carlos III, Majadahonda, Madrid). We thank Ana Royuela (Unidad de Bioestadı ´stica Clı ´nica, Hospital Universitario Ramo ´n y Cajal) for statistical advice during manuscript preparation and COESANT (Comite ´Espan˜ol del Antibiograma) for critical reading of the manuscript prior to submission. Funding This study was supported by the Instituto de Salud Carlos III, Subdireccio ´n General de Redes y Centros de Investigacio ´n Cooperativa, Ministerio de Economı ´a y Competitividad, Spanish Network for Research in Infectious Diseases (REIPI, RD12/0015) co-financed by the European Development Regional Fund (ERDF; A way to achieve Europe) and Fondo de Investigacio ´n Sanitaria (grants PI09/0917 and PI11/01117) and Fundacio ´nFranciscoSoria Melguizo. Transparency declarations R.C. is currently chairman of EUCAST. All other authors: none to declare. Supplementary data Figure S1 and Table S1 are available as Supplementary data at JAC Online (http://jac.oxfordjournals.org/). References 1Leverstein-van Hall MA, Waar K, Muilwijk J et al. Consequences of switching from a fixed 2:1 ratio of amoxicillin/clavulanate (CLSI) to a fixed concentration of clavulanate (EUCAST) for susceptibility testing of Escherichia coli.J Antimicrob Chemother 2013; 68: 2636–40. 2Mimica Matanovic S, Bergman U, Vukovic D et al. Impact of restricted amoxicillin/clavulanic acid use on Escherichia coli resistance–antibiotic DU90% profiles with bacterial resistance rates: a visual presentation. Int J Antimicrob Agents 2010; 36: 369–73. 3Oteo J, Campos J, La ´zaro E et al. Increased amoxicillin-clavulanic acid resistance in Escherichia coli blood isolates, Spain. Emerg Infect Dis 2008; 14: 1259–62. 4EUCAST. Breakpoint Tables for Interpretation of MICs and Zone Diameters, Version 4.0. www.eucast.org/clinical_breakpoints. 5Clinical and Laboratory Standards Institute. Performance Standards for Antimicrobial Susceptibility Testing: Twenty-fourth Informational Supplement M100-S24. CLSI, Wayne, PA, USA, 2014. 6Nguyen HM, Labreche MJ, Grabrer CJ. Making sense of cephalosporin and amoxicillin/clavulanate susceptibility testing for uropathogens. Clin Infect Dis 2014; 59: 1349–50. 7International Organization for Standardization (ISO). Clinical Laboratory Testing and In Vitro Diagnostic Test Systems—Susceptibility Testing of Infectious Agents and Evaluation of Performance of Antimicrobial Susceptibility Test Devices. Part 1: Reference Method for Testing the In Vitro Activity of Antimicrobial Agents against Rapidly Growing Aerobic Bacteria Involved in Infectious Diseases. International Standard 20776-1. Geneva: ISO, 2006. 8Oteo J, Gonza ´lez-Lo ´pez JJ, Ortega A et al. Inhibitor-resistant TEM- and OXA-1-producing Escherichia coli isolates resistant to amoxicillinclavulanate are more clonal and possess lower virulence gene content than susceptible clinical isolates. Antimicrob Agents Chemother 2014; 58: 3874–81. 9Dı ´az MA, Herna ´ndez-Bello JR, Rodrı ´guez-Ban˜o J et al.Diversityof Escherichia coli strains producing extended-spectrum b-lactamases in Spain: second nationwide study. J Clin Microbiol 2010; 48: 2840–5. 10 Clark RB, Lewinski MA, Loeffelholz MJ et al.Cumitech 31A, Verification and Validation of Procedures in the Clinical Microbiology Laboratory. Sharp SE, coordinating ed. Washington, DC: ASM Press, 2009. 11 Greenwood D, O’Grady F, Baker P. An in vitro evaluation of clavulanic acid,apotent,broad-spectrumb-lactamase inhibitor. J Antimicrob Chemother 1979; 5: 539–47. 12 JacksonD,CooperDL,HardyTLet al. Pharmacokinetic, toxicological and metabolic studies with Augmentin. In: Rolinson GN, Watson A, eds. Augmentin: Clavulanate-Potentiated Amoxicillin.Proceedingsofthe First Symposium, 3 and 4 July 1980. Oxford: Excerpta Medica, 1980; 87–111. 13 Todd PA, Benfield P. Amoxicillin/clavulanic acid. An update of its antibacterial activity, pharmacokinetic properties and therapeutic use. Drugs 1990; 39: 264–307. 14 VanderBijAK,vanDijkK,MuilwijkJet al.Clinicalbreakpointchanges and their impact on surveillance of antimicrobial resistance in Escherichia coli causing bacteraemia. Clin Microbiol Infect 2012; 18: E466–72. Amoxicillin/clavulanate testing for E. coli 2309 JA C Downloaded from https://academic.oup.com/jac/article/70/8/2306/810536 by guest on 07 April 2025 15 Hombach M, Bloemberg GV, Bo ¨ttger EC. Effects of clinical breakpoint changes in CLSI guidelines 2010/2011 and EUCAST guidelines 2011 on antibiotic susceptibility test reporting of Gram-negative bacilli. J Antimicrob Chemother 2012; 67: 622–32. 16 Brown D, Canto ´n R, Dubreuil L et al. Widespread implementation of EUCAST breakpoints for antibacterial susceptibility testing in Europe. Euro Surveill 2015; 20: pii¼21008. 17 Ortega A, Oteo J, Aranzamendi-Zaldumbide M et al. Spanish multicenter study of the epidemiology and mechanisms of amoxicillin-clavulanate resistance in Escherichia coli.Antimicrob Agents Chemother 2012; 56: 3576–81. 18 Oliver A, Pe ´rez-Va ´zquez M, Martı ´nez-Ferrer M et al.Ampicillinsulbactam and amoxicillin-clavulanate susceptibility testing of Escherichia coli isolates with different b-lactam resistance phenotypes. Antimicrob Agents Chemother 1999; 43:862–7. 19 Haeseker M, Havenith T, Stolk L et al. Is the standard dose of amoxicillin-clavulanic acid sufficient? BMC Pharmacol Toxicol 2014; 15: 38. 20 Rodrı ´guez-Ban˜o J, Navarro MD, Retamar P et al.b-Lactam/b-lactam inhibitor combinations for the treatment of bacteremia due to extendedspectrum b-lactamase-producing Escherichia coli: a post hoc analysis of prospective cohorts. Clin Infect Dis 2012; 54: 167–74. Dı ´ez-Aguilar et al. 2310 Downloaded from https://academic.oup.com/jac/article/70/8/2306/810536 by guest on 07 April 2025