EUCAST warnings concerning antimicrobial susceptibility testing products or procedures.
The EUCAST disk diffusion development laboratories, a network of laboratories coordinated from the EUCAST development laboratory in Växjö, Sweden, from time to time discover products (disks, media batches, gradient tests or procedures) which are not performing to the expected standard. When this is the case we inform the manufacturer and publish a warning on this page.
We do not systematically test all products so the lack of a warning does not imply that there is no problem with the product in question.
Laboratories which experience problems with a susceptibility test method, and suspect that this may be related to a particular product, may contact EUCAST for advice.
1. Problems with piperacillin-tazobactam gradient tests from two manufacturers - the warning issued 2015 will be removed when problem is resolved.
2. Wide variation in disk quality in 16 disks from nine manufacturers - the warning issued in 2015 and reiterated 2016 will be removed when issue is resolved.
3. Problems with colistin gradient tests from both bioMérieux and Liofilchem - the warning initially issued July 2016 was finalised 28 November, 2016.
1. Problems with piperacillin-tazobactam gradient tests from two manufacturers.
This warning, issued 2015, will be removed when problem is solved.
Following reports of problems related to piperacillin-tazobactam gradient tests, several batches of tests from bioMérieux and Liofilchem (from both single packs and multipacks) were evaluated at the EUCAST Development Laboratory, Växjö, Sweden (January 2016). Enterobacteriaceae and Pseudomonas aeruginosa (type strains and strains with resistance mechanisms) and broth micro dilution were used.
Gradient tests from both manufacturers gave variable and unreliable results. The difference between gradient tests and broth micro dilution was not systematic, some values were two dilutions too high, others too low. There was also considerable variation between batches from the same manufacturer.
We urge manufacturers to seriously consider these problems and users to introduce internal quality control of each procedure. We have detected no reason not to trust the disk diffusion (30+6 µg disk) results but currently you can not confirm your disk diffusion results with a gradient test.
2. Wide variation in disk quality in 16 disks from 9 manufacturers.
The warning was issued 2015 and reiterated in 2016. It will be removed when problem is solved.
During 2014 and 2015, EUCAST has evaluated the disk potency of 16 strategically important antibiotic disks from nine manufacturers of disks for antimicrobial susceptibility testing. Results from the first study were released at ECCMID 2014. Manufacturers were offered an opportunity to supply new disk lots to possibly improve their results. The second round of testing was presented at ECCMID in Amsterdam. Some manufacturers have asked that EUCAST performs further checks on new lots produced to try to right poor quality in previous lots. The current presentation (released 25 June, 2016) contains both old data and the new data when available.
The study disclosed some good and some poor quality among disks and manufacturers. Disks from a few of them were consistently of a high quality whereas the opposite was true for some manufacturers. Laboratories are warned against the use of some disks and manufacturers are urged to improve their manufacturing process.
A follow-up studies are being performed and are subsequently included in the presentation.
3. Antimicrobial susceptibility testing of colistin - problems detected with several commercially available products.
The warning was issued July 2016 and an addendum 26 August 2016. The current text is from 28 November, 2016, and is a summary of the previous warnings and includes the results following further testing of more broth microdilution tests and more organisms. Warnings will be modified or removed when issues have been resolved.
Antimicrobial susceptibility testing of colistin has been fraught with difficulties. A joint EUCAST and CLSI subcommittee recently issued recommendations confirming that broth microdilution (BMD) is so far the only valid method and that disk diffusion does not work because of the poor diffusion of the large colistin molecule. The report did not evaluate gradient tests of colistin, but reports in the literature have questioned the validity of MICs obtained with gradient tests.
We used a collection of clinical isolates (n=75) obtained from international contacts (acknowledge: SENTRY collection (P. Rhomberg, JMI Laboratories, USA), S. Gatermann, Bochum, Germany, R. Henriksen, DTU, Copenhagen, Denmark, Ö. Samuelsen, Tromsö, Norway, J. Vila, Barcelona and L. Martinez-Martinez, Santander, Spain) of Escherichia coli (n= 14), Klebsiella pneumoniae, (n=18 ) Pseudomonas aeruginosa (n=21) and Acinetobacter spp (n=22) without and with various resistance mechanisms and with different levels of resistance to colistin. Several Enterobacteriaceae had the mcr-1 gene. Broth microdilution MIC values of the four species were 0.25 – 128 mg/L.
For colistin broth microdilution, colistin frozen panels from TREK Diagnostics (Thermo Fisher Scientific) were compared with standard freeze-dried panels (Sensititre, Thermo Fisher cientific™ ) and two different Micronaut panels from Merlin™. All tested colistin broth microdilution panels worked well and only few errors occurred with any of the panels.
Colistin gradient tests can be obtained from bioMérieux and from Liofilchem. The EUCAST Development Laboratory has evaluated the two available gradient tests.
Results were similar for both gradient tests on Mueller-Hinton agar from two manufacturers. Etests were also tested on MH-E (the MH produced by and recommended by bioMérieux). Both gradient tests underestimated MIC values by one or more twofold dilutions, especially for concentrations on or above the breakpoint of 2 mg/L, leading to false susceptible results (VME). Results on Etest were marginally better when the test was performed on the MH recommended by the manufacturer but VMEs were still a significant problem, and even more so with Pseudomonas and Acinetobacter.
Both manufacturers have been informed and are working on improving their products. It is not known whether or not this issue can be resolved by recalibrating the gradient tests or whether it is related to the same difficulties as those seen with disk diffusion. The currently available gradient tests should until further notice be withdrawn from use in the laboratory!
Internal quality control. To detect the problems we have described it is not enough to use the recommended QC strains for E. coli (ATCC 25922) and P. aeruginosa (ATCC 27853). Most failures occurred where broth microdilution MICs were 2, 4 and 8 mg/L and the MICs of the QC strains are lower. The EDL is currently developing a suitable QC strain in the MIC range 4 - 8 mg/L.
Disk diffusion. The EUCAST Development Laboratory performed disk diffusion tests (disk potencies 10, 25 and 50 µg) in parallel with BMD and gradient tests. Our results confirmed that disk diffusion can not be used for AST of colistin using any of the currently available disks.
Semi-automated AST devices. We have not had the opportunity to systematically test either of Vitek2, Phoenix or MicroScan on the 75 isolates with varying colistin susceptibility, so we cannot recommend the use or non-use of either for colistin susceptibility testing. However, we would recommend that users until further notice include the quality control strains to better control their instrument’s performance.
Quality control of colistin must be performed with both a susceptible QC strain (E. coli ATCC 25922 or P. aeruginosa ATCC 27853) and the colistin resistant E. coli NCTC 13846 (mcr-1 positive). For E. coli NCTC 13846, the colistin MIC target value is 4 mg/L and should only occasionally be 2 or 8 mg/L
Summary and conclusion: Disk diffusion cannot be used at present for susceptibility testing of colistin. Currently available gradient tests underestimate colistin MIC values and undercall resistance, , and should be avoided. The broth microdilution tests we have managed to evaluate all seem to give correct results both for susceptible and non-susceptible isolates. We have not had the opportunity to systematically evaluate semi-automated colistin methods but by sending isolates with MIC values in the non-susceptible range to colleagues around the world we have disclosed several Very Major Errors. Our conclusion is that all colistin susceptibility testing should be quality-controlled by the inclusion of a sensitive and a resistant QC strain. The latter has recently been defined (see above).
Warnings will be removed when the issues have been solved.