Bacteroides spp. are the most predominant anaerobic gram-negative bacteria in the gut. Bacteroides fragilis accounts for only 0.5% of the gut microbiota, however it is the most commonly isolated anaerobic pathogen in intraabdominal infections and bacteremia with high mortality rate (more than 19%) [1]. Although often susceptible to metronidazole, B. fragilis infections could be challenging for treatment because physicians must face increasing resistance to antibiotics. For several years, an increase of resistance to β-lactams agents such as amoxicillin-clavulanic acid but also carbapenems has been observed [2]. For the latter, the resistance is mainly due to a carbapenemase enzyme encoded by the cfiA gene and belonging to the metallo-β-lactamase (MBL) class [3]. Once activated by an insertion sequence, this gene is able to confer high-level resistance to the carbapenems as well as to the penicillins, cephalosporins and to most β -lactamase inhibitors [4].

To tackle such antibiotic resistance in B. fragilis, an imipenem double-ended Etest ± EDTA is commonly used to rapidly detect MBL production [5]. This phenotypic test compares the resistance of strains to imipenem with and without EDTA. Since MBL can be experimentally inhibited with metal chelators such as EDTA, the CfiA production can be inferred with a MIC ratio ≥ 8 that indicates a reduction of imipenem MIC by at least 3 2-fold dilutions in the presence of EDTA [5].

A major issue with the imipenem ± EDTA Etest is the only detection of isolates with high-level resistance [6]. This is a significant drawback for clinical microbiology diagnosis because it may clearly underestimate the CfiA prevalence in B. fragilis [6]. Therefore, Schwensen et al. proposed to solve this problem by using preferentially the meropenem-EDTA double-ended Etest or the ROSCO KPC/MBL Confirm Kit [7]. This latter assay allows to compare the sensitivity to meropenem ± dipicolinic acid, ± boronic acid and ± cloxacillin in order to confirm the presence of MBL. The presence of CfiA is then evidenced by a restoration of inhibition diameter of the meropenem + dipicolinic acid disc only. However, the meropenem double-ended Etest is expensive and the ROSCO KPC/MBL Confirm kit could misclassify cfiA-negative and cfiA-positive isolates in case of small diameters of inhibition.

An alternative way to detect MBL production is to use carbapenem inactivation methods (CIM), based on the enzymatic hydrolysis of meropenem susceptibility-testing disc after its exposure to a carbapenemase producing strain, which allows subsequent uninhibited growth of a full susceptible indicator strain [8]. These phenotypic methods are well known for detecting carbapenemase production, mainly in Enterobacterales [9] but have also been proposed for the detection of carbapenemase-producing Acinetobacter species [10] and Pseudomonas aeruginosa [11]. Recently, an anaerobic CIM (Ana-CIM) has been described to improve and facilitate the detection of B. fragilis carbapenemase [12].

The aim of this study was to evaluate the practicability in everyday practice and to improve the performance of the Ana-CIM to detect cfiA carbapenemase production in B. fragilis group.