Reappraisal of the antimicrobial susceptibilities of Chryseobacterium and Flavobacterium species and methods for reliable susceptibility testing

Several Flavobacterium species, comprising a heterogeneous group of gram-negative bacilli that are capable of causing opportunistic infections in humans, have recently been reclassified as Chryseobacterium or Myroides species. Intrinsically resistant to a number of antibiotics, these organisms have been reported to be susceptible to vancomycin and certain other drugs that are normally active against gram-positive bacteria. By using the National Committee for Clinical Laboratory Standards (NCCLS) broth microdilution procedure, 58 clinical isolates of former flavobacteria (36 Chryseobacterium meningosepticum isolates, 11 C. indologenes isolates, 3 C. gleum isolates, 4 unspeciated former members of Flavobacterium group IIb, and 4 Myroides odoratum isolates) were tested with 23 antibiotics, including conventional and investigational agents. In addition, the broth micro- dilution results were compared to those generated by agar dilution, E-test, and disk diffusion for vancomycin and piperacillin-tazobactam. Compared to the NCCLS microdilution results, there were 7.1 and 17.9% very major errors with piperacillin-tazobactam by agar dilution and E-test, respectively. In addition, there were from 12.1 to 48.3% minor errors with both procedures with vancomycin and piperacillin-tazobactam. The very major and minor error rates were unacceptably high with disk testing of piperacillin-tazobactam; the use of enterococcal vancomycin disk breakpoints (zone diameter of ≤17 mm = susceptible) resulted in >20% minor errors but only one very major error. All of the isolates were susceptible to minocycline; over 90% were susceptible to sparfloxacin, levofloxacin, and clinafloxacin; and 88% were susceptible to rifampin. None was susceptible to vancomycin. When Chryseobacterium or Myroides species are isolated from serious infections, susceptibility testing by broth microdilution should be performed and therapy should be guided by those results.