Comparison of various in vitro susceptibility methods for testing Stenotrophomonas maltophilia

A total of 57 clinical isolates were screened by disk diffusion for a relat ed pharmacodynamic study. Testing was performed using National Committee fo r Clinical Laboratory Standards guidelines, except that results were interp reted at 16 to 18 h and 48 h. Of the 57 isolates, 19 were randomly chosen f or additional comparative susceptibility testing of five methods (disk diff usion, Etest, Alamar colorimetric broth microdilution, Vitek, and MicroScan ) and an in-house broth microdilution method. The two, diffusion methods (d isk and Etest) had the closest correlation. The commercial broth microdilut ion methods and the in-house,microdilution method generated inconsistent re sults for all agents except trimethoprim-sulfamethoxazole. Vitek compared p oorly with both diffusion and microbroth dilution methods. The most signifi cant discrepancies were evident with all methods when the incubation period was extended to 48 h. When results mere interpreted at 48 h, the incidence of resistance for all bactericidal agents was approximately double the res istance observed at 16 to 18 h. The bacteriostatic agents, trimethoprim-sul famethoxazole and doxycycline, demonstrated the greatest in vitro activity and were least influenced by extended incubation with diffusion methods. Be cause correlative in vivo and in vitro studies have not revealed an effecti ve therapeutic regimen for serious S. maltophilia infections, susceptibilit y results with all testing methods should be interpreted with caution when choosing therapy for patients with life-threatening infections. Susceptibil ity testing for this heterogeneous group remains controversial and routine testing, with the possible exception of doxycycline (Or minocycline) and tr imethoprim-sulfamethoxazole, should be avoided. Our data support that if te sting is done with bactericidal agents, consideration should be given to in terpretation after 48-h incubation. (C) 1998 Elsevier Science Inc.