The Quality of Antimicrobial Susceptibility Testing in Europe

Abstract & Commentary

Synopsis: The apparent differences in resistance rates between countries might simply reflect technical variation rather than true divergence of susceptibility.

Source: Snell JJ, Brown DF. External quality assessment of antimicrobial susceptibility testing in Europe. J Antimicrob Chemother. 2001;47:801-810.

The Quality Assurance Laboratory of the Central Public Health Laboratory has been running a quality assurance scheme for antimicrobial susceptibility testing in England and Wales since 1975 and branched out into Europe some 10 years later. Beginning in 1985 with Haemophilus influenzae that did not express ß-lactamase, but were resistant to ampicillin and co-amoxiclav, and in 1990 with stains that were resistant to chloramphenicol, they added enterococci expressing van B vancomycin resistance or high-level gentamicin resistance in 1991; methicillin-resistant coagulase-negative staphylococci and methicillin-resistant Staphylococcus aureus (MRSA) in 1992; and, finally, moderately penicillin-resistant Streptococcus pneumoniae to their distribution. By this time, a total of 716 laboratories in Switzerland and 9 countries of the European Union (EU) had become involved. Each participant was asked to use their routine method for antimicrobial susceptibility testing and report the results as either "susceptible" or "resistant." As is customary in these exercises, the true identity of the countries was concealed.

Considerable variation was noted between countries. The number of times the laboratory performance for the test fell below average (ie, getting the right answer) ranged from 12% for country I to 70% for country F (see Table 1).

Table 1: Performance by Country
Country # tests
# times the lab performance for the test fell below average % Rank 

A 58 9 16 2
B 58 32 55 7
C 50 21 42 5
D 53 25 47 6
E 49 15 31 3
F 27 19 70 10
G 28 11 39 4
H 45 26 58 8
I 26 3 12 1
J 38 23 61 9

The poorest performance was seen for van B enterococci with resistance being detected by a mean of only 77% of participants and the best performance was seen for MRSA with all participants arriving at the correct answer (see Table 2).

Table 2: Average Performance for Different Bacterial Strains on the Last Round of Tests
Haemophilus influenzae that do not express ß-lactamase but were resistant to ampicillin 92%
Haemophilus influenzae that do not express ß-lactamase but were resistant to co-amoxiclav 82%
Haemophilus influenzae resistant to chloramphenicol 82%
Enterococci expressing van B vancomycin resistance 77%
Enterococci high-level gentamicin resistance 88%
Methicillin-resistant coagulase-negative staphylococci 71%
Methicillin-resistant S aureus 100%
Moderately penicillin-resistant Streptococcus pneumoniae 85%

Snell and Brown cautioned against reading too much into the divergence in results because the organism-antimicrobial combinations were known to be the most challenging to even the best laboratories. Also, some of the participants had joined the program only recently, although their apparent inexperience was not always reflected in poorer performance. Snell and Brown were unable to offer an explanation for the divergent results but surmised that it was almost certainly due to different methodologies that would be the subject of future investigations. Whatever the explanation, the differences observed would almost certainly distort any comparisons of resistance rates and would make estimates of incidence rates of resistance unreliable.

Comment by J. Peter Donnelly, PhD

Whatever else these results show, it is clear that different laboratories arrive at different answers and that this is to some extent determined by local conditions. Unlike the United States, which has adopted a single national standard for antimicrobial susceptibility testing, we still await a single standard for the EU. In fact, countries like Germany are the exception, rather than the rule, in having a national standard. As a result, there are a variety of methods used many of which are, in fact, variants of the National Committee for Clinical and Laboratory Standards (NCCLS) protocols. The lack of a common standard here in Europe motivated the European Society of Clinical Microbiology and Infectious Diseases to set up a Standing Committee called the European Committee on Antimicrobial Susceptibility Testing of EUCAST "to try to ensure that susceptibility testing in Europe produces comparable results and interpretations."1 Besides the main committee there are specific subcommittees for Breakpoints, Dilution methods, Diffusion methods, Quality control, Fungi (of which I am currently secretary), Intracellular pathogens, and Automation. EUCAST already liases with the NCCLS, the European Medicines Evaluation Agency, and other agencies for setting standards. As with other areas of life, this initiative does not enjoy unanimous support within the EU nor across Europe. But it has gotten further than before, and fulfills an obvious need to have standards. The current trend toward drawing up resistance league tables requires them, the greater public interest in antibiotic resistance demands them, and the results of the NEQAS external quality assessment program illustrate the need for them.

Dr. Donnelly, Clinical Microbiologist, University Hospital, Nijmegen, The Netherlands, is Associate Editor of Infectious Disease Alert.