Aims:  In this study, we compared different methods of colistin susceptibility testing, disc diffusion, agar dilution and Etest using a set of Enterobacteriaceae isolates that included colistin‐resistant strains. Methods and results:  Susceptibility of 200 clinical isolates of Enterobacteriaceae to colistin was tested to compare agar dilution (reference method), disc diffusion (50 and 10 μg) and Etest. MICs (minimum inhibitory concentrations) were interpreted using the criteria established by the European Committee on Antimicrobial Susceptibility Testing (EUCAST). Colistin exhibited excellent activity against Escherichia coli and E. cloacae (MIC90 = 0·5 mg l−1). In contrast, colistin was less active against Klebsiella pneumoniae (MIC90 = 16 mg l−1). Resistance rates varied from 0% in E. coli to 1·8% in E. cloacae and 13% in K. pneumoniae. High rates of very major errors were observed in the disc diffusion test using either the criteria of the Comité de l’antibiogramme de la Société Française de Microbiologie (CA‐SFM) or the criteria of the Clinical and Laboratory Standards Institute (CLSI), respectively, 3·5 and 2·5%. When the criteria of Gales et al. were applied, the number of very major errors was reduced to one (0·5%). The Etest showed good concordance with agar dilution method. Conclusion:  Disc susceptibility testing methods are unreliable on detecting colistin resistance. MIC should be determined to confirm the susceptibility results by disc diffusion. Significance and Impact of the study:  We recommend the determination of MIC by Etest for all multidrug‐resistant Enterobacteriaceae when colistin is required for the treatment.

We characterized 67 Escherichia coli isolates with reduced susceptibility to cefotaxime obtained from 136 samples of healthy broilers housed in 36 Tunisian farms. All these isolates harboured blaCTX‐M‐1 and/or blaCMY‐2 genes located mostly on self‐conjugative IncI1 plasmids. qnrS1, qnrA6 and aac(6′)‐Ib‐cr were detected in six isolates. Considerable genetic diversity was detected among isolates from different farms. To our knowledge, this is the first detailed documentation of a high occurrence of blaCTX‐M‐1 and blaCMY‐2 in E. coli at the poultry farm level in Tunisia as well as the first description of plasmid‐mediated quinolone resistance in food animals in Tunisia which may contribute to the dissemination of these genes throughout Tunisia.

This is the first report of an outbreak due to P. stuartii isolates carrying bla(OXA-48) in Tunisia. The simultaneous expression of various resistance genes (bla(OXA-48), bla(CMY-4), bla(PER-1), qnrA and aac-6'-Ib) by P. stuartii isolates is alarming.

Twenty-three strains of Staphylococcus aureus with borderline resistance to oxacillin were studied. These strains were not detected by the cefoxitin test, tests for penicillin-binding protein 2a (PBP2a), mecA , and mecA LGA251 were negative, and the strains were genetically unrelated. To detect all strains resistant to oxacillin, laboratories should routinely test for both cefoxitin and oxacillin.

19F and 14 were the most prevalent serotypes in the south of Tunisia. The inclusion of a PCV in the immunization program could be useful for reducing the burden of pneumococcal diseases. The high resistance rate to penicillin and macrolides is alarming. Prudent use of antibiotics is crucial to prevent the selection of multidrug-resistant pneumococci.