Integrating Susceptibility Data from Two Surveillance Programs with Unique Methodological Techniques: The Antibiogram Resistance Method or Isolate-Based Resistance Monitoring (ARMOR) Study Group

Abstract

Background Antimicrobial resistance can be tracked by several methods. The optimal application of surveillance strategies is an integral part in curbing the resistance epidemic. Objectives The Antimicrobial Resistance Management (ARM) and Meropenem Yearly Susceptibility Test Information Collection (MYSTIC) programs are ongoing antibiotic surveillance studies with different methodologies. Our aim was to compare susceptibility data generated by each program to determine if the two methods produce similar results. Methods ARM is a Web-based, antibiogram-driven surveillance system, whereas MYSTIC tests individual bacterial isolates in accordance with Clinical Laboratory Standards Institute methods. We compared national susceptibility rates during 1999 to 2004 from ARM and MYSTIC for Escherichia coli and Klebsiella pneumoniae. Results E. coli susceptibility was greater than 90% for all agents evaluated except ciprofloxacin and levofloxacin. Susceptibility rates in ARM and MYSTIC were similar (within 5%), with the exception of levofloxacin — for which susceptibility was reported as 92.3% in ARM and 83% in MYSTIC ( P < 0.0001). Amikacin, cefepime, ceftriaxone, imipenem, and meropenem were identified by both surveillance programs as the most active agents against E. coli and K. pneumoniae. Both systems detected increasing fluoroquinolone resistance ( P < 0.0001) and relatively stable activity among other agents tested. For K. pneumoniae, all agents displayed resistance rates less than 10%. Susceptibility rates in ARM and MYSTIC were remarkably similar. The largest difference was 2.9% for piperacillin/tazobactam (92.8% in ARM vs 95.7% in MYSTIC, P < 0.0001). Regionally, nonsusceptibility to E. coli and K. pneumoniae was similar, though for some comparisons MYSTIC reported higher rates of nonsusceptibility in the Northeast, Northwest, and South Central. Conclusion Surveillance of antibiotic activity patterns can be performed confidently with either method discussed above. Attempts should be made to integrate results from different surveillance programs to optimize surveillance.