28 Mar 2021

EUCAST aphorisms

  • The manufacturer of antimicrobial susceptibility testing material is responsible for the quality of the material provided to microbiological laboratories. The laboratory is responsible for the result provided to doctors and patients.
  • To perform AST where there is intrinsic resistance is wasting the time of doctors, patients and laboratory staff. 
  • Expert rules are not the absolute truth, they may change over time.
  • The MIC is a relative measurement which in a standardised system can place the organism inside or outside the wild type distribution, but not quantitate the level of sensitivity inside the wild type1.
  • Two levels of resistant, I and R, were replaced by two levels of susceptible, S and I, when EUCAST changed the definitions of S, I and R, and the breakpoints to match.
  • The EUCAST "increased exposure" is a function of dosing, mode of administration and the pharmacokinetics of the agent. Dosing can be affected by changing the individual dose and/or the interval between doses. Exposure can be affected by changing the dose, or the mode or length of administration, or by the pharmacokinetic properties of the agent.
  • The ECOFF is determined on phenotypic criteria only - it is the highest MIC of wild type organisms, e.g. organisms without phenotypically detectable acquired resistance.
  • The area of technical uncertainty (ATU) is a warning in the laboratory which predicts interpretative difficulties but does not interfere with categorisation to S, I or R.
  • EUCAST RAST (rapid AST) directly from positive blood culture bottles can provide susceptibility results in 4 - 6 hours for pathogens important in blood stream infections - why wait for another 18h only because it requires re-organising the work flow.
  • EUCAST disk diffusion screening tests and algorithms reliably predict susceptibility by excluding resistance.

    • Cefoxitin 30 µg to exclude methicillin resistance in Staphylococci.
    • Oxacillin 1 µg in S. pneumoniae to exclude all mechanisms of beta-lactam resistance.
    • Benzylpenicillin 1U in H. influenzae to exclude all mechanisms of beta-lactam resistance.
    • Norfloxacin 10 µg to exclude fluoroquinolone resistance in grampositive bacteria (Staphylococci, Pneumococci, Enterococci).
    • Pefloxacin 5 µg to exclude fluoroquinolone resistance in Salmonella enterica.
    • Nalidixic acid 30 µg to exclude fluoroquinolone resistance in H. influenzae and M. catarrhalis.

1The variation in the wild type MIC distribution has been a source of debate for quite some time. There are publications indicating that 70% of the variation in a typical wild type distribution is down to technical variation and 30 % to biological variation. Without doubt, a true wild type for any species/agent (where there are not grave technical difficulties) span 3 – 5 twofold dilutions. Look at the EUCAST MIC distribution website at two typical extremes (gentamicin vs. E.coli (0.125 – 2 mg/L) and vancomycin vs. S.aureus (0.5 – 2 mg/L). 

It is also a known fact that if someone who is expert at performing broth microdilution determine the MIC of a specific isolate over and over again (not changing the material) a mean value plus minus one dilution step will cover 97-99% of the values obtained. If you change batch, operator, atmosphere or systematically change the incubation time from let us say 20 h to 16h (or vice versa) you widen the distribution significantly compared to the original one. If you shorten or extend incubation time even further, your MIC-values will be significantly different – which is why MIC is not an absolute measurement but a relative one.

So subtracting three dilution steps from 3 – 5 dilution steps leaves very little room, for variation between organisms inside the wild type. The representative examples of gentamicin and vancomycin, are based on very large numbers of results from very many investigators using standardised but not coordinated methods. Adding the variation between them, there is very little room for true, and above all no room at all for USEFUL AND PREDICTABLE, biological variation. 

This is why EUCAST since many years considers the wild type to be a “unity” which should not be divided by a breakpoint. And indeed we manage to achieve this. 

If we cannot have breakpoints divide wild type distributions (always defined by a species and an agent) it follows that you cannot determine the MIC of an organism inside the wild type to predict outcome of therapy beyond the prediction valid for the entire wild type distribution of isolates..

This is the basis for the EUCAST aphorism: "The MIC is a relative measurement which in a standardised system can place the organism inside or outside the wild type distribution, but not quantitate the level of sensitivity inside the wild type".

See also: J Antimicrob Chemother. 2018 Mar 1;73(3):564-568. doi: 10.1093/jac/dkx427.