Ability of laboratories to detect emerging antimicrobial resistance: proficiency testing and quality control results from the World Health Organization's external quality assurance system for antimicrobial susceptibility testing.

The accuracy of antimicrobial susceptibility data submitted by microbiology laboratories to national and international surveillance systems has been debated for a number of years. To assess the accuracy of data submitted to the World Health Organization by users of the WHONET software, the Centers for Disease Control and Prevention distributed six bacterial isolates representing key antimicrobial-resistance phenotypes to approximately 130 laboratories, all but one of which were outside of the United States, for antimicrobial susceptibility testing as part of the World Health Organization's External Quality Assurance System for Antimicrobial Susceptibility Testing. Each laboratory also was asked to submit 10 consecutive quality control values for several key organism-drug combinations. Most laboratories were able to detect methicillin (oxacillin) resistance in Staphylococcus aureus, high-level vancomycin resistance in Enterococcus faecium, and resistance to extended-spectrum cephalosporins in Klebsiella pneumoniae. Many laboratories, particularly those using disk diffusion tests, had difficulty in recognizing reduced susceptibility to penicillin in an isolate of Streptococcus pneumoniae. The most difficult phenotype for laboratories to detect was reduced susceptibility to vancomycin in an isolate of Staphylococcus epidermidis. The proficiency testing challenge also included a request for biochemical identification of a gram-negative bacillus, which most laboratories recognized as Enterobacter cloacae. Although only a small subset of laboratories have submitted their quality control data, it is clear that many of these laboratories generate disk diffusion results for oxacillin when testing S. aureus ATCC 25923 and S. pneumoniae ATCC 49619 that are outside of the acceptable quality control range. The narrow quality control range for vancomycin also proved to be a challenge for many of the laboratories submitting data; approximately 27% of results were out of range. Thus, it is important to establish the proficiency of laboratories submitting data to surveillance systems in which the organisms are tested locally, particularly for penicillin resistance in pneumococci and glycopeptide resistance in staphylococci.