UK clinical laboratories overestimate ciprofloxacin resistance amongst Pseudomonas aeroginosa isolates, relative to the MIC breakpoint of 1 mg/L. Most tests leading to this overestimation use 1 mu g discs and are by Stokes' method with the breakpoint taken as the zone radius for P. aeruginosa NCTC 10662 minus 3 mm. Aiming to reduce this error rate, we examined alternative disc breakpoints. Tests were performed for 100 P. aeruginosa isolates on three media, with breakpoints selected (i) as the zone for P. aeruginosa NCTC 10662 minus 7 mm, as recommended for ciprofloxacin by the BSAC; (ii) with reference to MIC/zone correlation lines; (iii) from natural divisions in zone distribution histograms; and (iv) so as to minimize categorization errors. Breakpoints from regression lines, and those optimized to the susceptibility distribution, reduced the proportion of susceptible organisms misreported as resistant, but the improvement was not significant (P. 0.05, chi(2) test). The breakpoint of the zone radius for P. aeruginosa NCTC 10662 minus 7 mm significantly reduced (P < 0.05) the number of susceptible organisms reported as resistant, but led to 50-75% of those with low level resistance (MIC 2-4 mg/L) and 4-10% of those with high-level resistance (MIC > 4 mg/L) being classed as susceptible. Irrespective of the medium and the basis of choosing breakpoints, 5 mu g ciprofloxacin discs gave a lower rate of susceptible organisms being reported as resistant than did 1 mu g discs; however, the improvement was not significant (P > 0.05, chi(2) test) and the 5 mu g discs had the disadvantages of forming very large zones for susceptible isolates and giving some-albeit small-zones for highly resistant organisms. In conclusion, the over-reporting of resistance could be reduced by use of zone breakpoints optimized to the MIC distribution and by the use of 5 mu g discs, but the case for these changes is not overwhelming; taking the breakpoint as the zone for NCTC 10662 minus 7 mm led to unacceptable numbers of resistant organisms being reported as susceptible. More fundamentally, ciprofloxacin zones and MICs are continuously distributed for P. aeruginosa isolates, so susceptibility tests cannot divide the species into discrete populations. In these circumstances, it is optimistic to expect disc and MIG categorizations to agree perfectly.
