Suggested modifications for disk diffusion susceptibility testing criteria for levofloxacin and sparfloxacin following tests with a predictor panel of ciprofloxacin-resistant clinical isolates.

A predictor panel of 300 clinical bacterial isolates (200 resistant to ciprofloxacin) was used to compare 5-micrograms disk diffusion test results with the MICs of ofloxacin (control), levofloxacin, and sparfloxacin. Regression analysis demonstrated high correlations between the methods for all three fluoroquinolones (r > or = 0.95). In order to minimize disk diffusion testing errors among the fluoroquinolone-resistant strains, the following modifications to previously proposed or published interpretive criteria were suggested: for levofloxacin, susceptible at > or = 17 mm (< or = 2 micrograms/ml) and resistant at < or = 13 mm (> or = 8 micrograms/ml); for sparfloxacin, susceptible at > or = 20 mm (< or = 1 microgram/ml) and resistant at < or = 16 mm (> or = 4 micrograms/ml). The study control drug, ofloxacin, did not appear to possess a significant error rate (5% minor error) when fluoroquinolone-resistant strains were tested, and no modifications were proposed. Under these proposed interpretive criteria, the absolute categorical agreements between standardized susceptibility testing methods for levofloxacin and sparfloxacin results were 91.3 and 94.0%, respectively (< or = 0.3% major errors and nil very major errors).