Evaluation of the Haemophilus influenzae EUCAST and CLSI disc diffusion methods to recognize aminopenicillin and amoxicillin/clavulanate resistance S. A. Fernando 1 *, S. Pang 2,3 , G. L. McKew 1,4 , T. Phan 1 , J. Merlino 1 , G. W. Coombs 2,3 and T. Gottlieb 1,4 1 Department of Microbiology and Infectious Diseases, Concord Repatriation General Hospital, Sydney, Australia; 2 Antimicrobial Resistance and Infectious Diseases Laboratory, School of Veterinary Life Sciences, Murdoch University, Murdoch, Western Australia, Australia; 3 PathWest Laboratory Medicine WA, Fiona Stanley Hospital, Murdoch, Western Australia, Australia; 4 University of Sydney, Sydney, Australia *Corresponding author. E-mail: shelanah.fernando@health.nsw.gov.au Received 24 September 2019; returned 18 November 2019; revised 22 March 2020; accepted 30 April 2020 Objectives: Implementation of EUCAST susceptibility testing in an Australian hospital laboratory demonstrated higher rates of aminopenicillin and amoxicillin/clavulanate resistance in Haemophilus influenzae than previously recognized. This study aimed to better define the variability in the detection of b-lactam resistance based on EUCAST and CLSI disc diffusion (DD) methodology, by comparison with the recommended reference method, broth microdilution (BMD), and by concordance with genomic analysis. Methods: A total of 100 random H. influenzae isolates were assessed for ampicillin and amoxicillin/clavulanate susceptibility by EUCAST and CLSI DD and BMD. WGS was used to analyse the ftsI gene of a subset of isolates with b-lactam resistance, other than that due to isolated b-lactamase production. Results: Of the 100 isolates, 32 were categorized as either b-lactamase negative, ampicillin resistant (BLNAR) (n = 18) or b-lactamase positive, amoxicillin/clavulanate resistant (BLPACR) (n = 14) by EUCAST DD. All 18 EUCAST BLNAR isolates were genotypically confirmed by WGS. Five of 18 BLNAR isolates were concordant by CLSI DD, 12 by EUCAST BMD and 4 by CLSI BMD. Nine of 14 EUCAST BLPACR isolates were confirmed by WGS; the remaining 5 were 1mm below the EUCAST DD breakpoint. Only one isolate was detected as BLPACR by CLSI DD. Group III mutations associated with high-level ampicillin resistance were identified in 10/32 isolates. Conclusions: The EUCAST DD susceptibility method is more reliable than either CLSI or BMD for the detection of genotypically defined BLNAR resistance. However, accurate categorization of amoxicillin/clavulanate resistance remains problematic. Continuous and reproducible surveillance of resistance is needed; for this to be possible, robust susceptibility methods are required. Introduction Validating a ‘gold standard’ method for detecting amino- penicillin resistance and establishing a consensus definition of b-lactam resistance in Haemophilus influenzae has proved elusive. Numerous studies have emphasized the challenges of accurately detecting b-lactam resistance: poor correlation be- tween disc diffusion (DD) and MIC methods; clustering around breakpoints; and the inherent difficulty in phenotypically sepa- rating genotypically resistant strains from WT strains due to overlapping ampicillin MIC ranges. 1–3 To date, Australian epidemiology for H. influenzae has been largely derived from CLSI methodology; in this context, b-lacta- mase-negative, ampicillin-resistant (BLNAR) and b-lactamase positive, amoxicillin/clavulanate-resistant (BLPACR) strains were infrequently detected and were believed to be uncommon. 4–6 Earlier studies have identified the predominant non-b-lactamase PBP resistance mechanism for aminopenicillin resistance in Australia to be the Group II N526K mutation of the ftsI gene. 2,6 Such genotypic BLNAR (gBLNAR) strains are notable for variabil- ity in the phenotypic expression of aminopenicillin resistance (MICs = 0.5–16 mg/L), 2,7,8 with lower ampicillin MICs when compared with Group III strains containing the N526K mutation with additional S385T, L389F or M377I ftsI mutations. 9 Low-level aminopenicillin resistance has contributed to the challenge of phenotypically separating genotypic BLNAR strains from WT strains. 1,2 V C The Author(s) 2020. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For permissions, please email: journals.permissions@oup.com. 2594 J Antimicrob Chemother 2020; 75: 2594–2598 doi:10.1093/jac/dkaa229 Advance Access publication 25 June 2020 Downloaded from https://academic.oup.com/jac/article/75/9/2594/5862740 by guest on 04 January 2024