Evaluation of accuracy and reproducibility of E test for susceptibility testing of Streptococcus pneumoniae to penicillin, cefotaxime, and ceftriaxone.

We evaluated the reproducibility with which technologists perform and interpret the E test (AB Biodisk, North America, Inc., Piscataway, N.J.) for determining the susceptibility of Streptococcus pneumoniae to penicillin, cefotaxime, and ceftriaxone. Four technologists prepared E test assays to test 124 isolates of S. pneumoniae. Each technologist then interpreted the results of the E test blinded to the interpretation of the other technologists. In addition, E test results were compared with the reference method of broth microdilution. Intraobserver and interobserver agreement were assessed by use of the kappa statistic. Interpretation of the E test and broth microdilution results showed substantial to excellent agreement, with kappa values ranging from 0.878 to 0.987. Compared with broth microdilution, no very major errors and only four major errors were made with the E test. Most minor errors with penicillin and ceftriaxone occurred for isolates with intermediate or high-level resistance, whereas for cefotaxime the minor errors were more evenly distributed between susceptible and intermediate resistance and between intermediate and high-level resistance. These results indicate that there is good agreement between technologists for the interpretation of the E test when testing the susceptibility of S. pneumoniae to penicillin, cefotaxime, and ceftriaxone and that the results of the E test agree with those of broth microdilution.