Single-strain regression analysis for quality control of cephalothin-susceptibility testing and determination of interpretive breakpoints.

Abstract

Histogram analysis of inhibition zone diameters around the 30 micrograms cephalothin disk for E. coli, P. mirabilis, and K. pneumoniae in samples from 1975 to 1982 showed a marked reproducibility of the disk-diffusion antibiotic-susceptibility test in the routine laboratory. A comparison of interpretive breakpoints with histograms for E. coli, P. mirabilis, K. pneumoniae, S. aureus, coagulase-negative staphylococci, and S. faecalis showed a higher proportion of possible misinterpretations using the breakpoints of the Swedish Reference Group, SRG, as compared to international (NCCLS) breakpoints. Further analysis using single-strain regression analysis revealed two major causes of interpretive errors. Firstly, the laboratory-related regression line for a bacterial species can be different from the general regression line of the reference laboratory. This difference has to be corrected by using species-related breakpoints. For E. coli, a species-specific breakpoint was determined to R = greater than 13 mm. Secondly, MIC limits recommended for the susceptibility categories of cephalothin by SRG are lower than the international limits and close to the true MIC values of many bacterial isolates, leading to misinterpretations due to the methodological variation. These studies suggest an adoption of international MIC limits for the susceptibility categories of cephalothin in Scandinavia. The "I" category should denote an indeterminate zone. A multi-laboratory quality control assessment using histogram analysis is recommended with optional single-strain regression analysis to determine breakpoints for problem combinations of bacterial species and antibiotics.