Reporting antimicrobial susceptibilities and resistance phenotypes in Acinetobacter spp: a nationwide proficiency study.

Abstract

To evaluate the proficiency of Spanish microbiology laboratories with respect to the antimicrobial susceptibility testing (AST) of Acinetobacter spp. Eight Acinetobacter spp. with different resistance mechanisms were sent to 48 Spanish centres which were asked to report: (i) the AST system used; (ii) MICs; (iii) breakpoints used (CLSI versus EUCAST); (iv) clinical category; and (v) resistance mechanisms inferred. Minor, major and very major errors (mE, ME and VME, respectively) were determined. The greatest percentages of discrepancies were: (i) by AST method: 18.5% Etest, 14.3% Vitek 2 and Sensititre; (ii) by breakpoints: 20.5% (CLSI) and 10.8% (EUCAST); and (iii) by antimicrobial agent: ampicillin/sulbactam (56.2% CLSI), minocycline (40.7% CLSI), tobramycin (38.7% CLSI, 16.8% EUCAST), imipenem (27.8% CLSI, 30.0% EUCAST) and meropenem (25.4% CLSI, 20.8% EUCAST). Categorical error rates: (i) by AST method ranged from 30.0% (Phoenix) to 100% (Sensititre and disc diffusion) for mE, 0.0% (Etest, Sensititre, disc diffusion) to 40% (Phoenix) for ME, and 0.0% (Sensititre and disc diffusion) to 30% (Phoenix) for VME; (ii) by breakpoints: mE (80.1% CLSI, 58.4% EUCAST), ME (3.5% CLSI, 12.4% EUCAST) and VME (16.4% CLSI, 29.2% EUCAST); and (iii) by antimicrobial agent: mE (100% levofloxacin/CLSI, 100% levofloxacin and meropenem/EUCAST), ME (35.3% colistin/CLSI, 25.0% colistin/EUCAST) and VME (64.7% colistin/CLSI, 86.7% gentamicin/EUCAST). Clinical microbiology laboratories must improve their ability to determine antimicrobial susceptibilities of Acinetobacter spp. isolates. Higher discrepancies using CLSI when compared with EUCAST are mainly due to mE and to a much lesser extent to ME or VME.