Variability of β-lactam susceptibility testing for Streptococcus pneumoniae using 4 commercial test methods and broth microdilution

Limited data are available that verify the performance of commercial susceptibility methods for Streptococcus pneumoniae following the 2008 Clinical and Laboratory Standards Institute revision of the β-lactam breakpoints. We compared the performance of Etest, M.I.C. Evaluator (M.I.C.E), Vitek, and Sensititre systems to broth microdilution for S. pneumoniae susceptibility testing of penicillin, ceftriaxone, meropenem, and amoxicillin. Essential agreement was ≥90% for the majority of the β-lactams and methods tested, particularly for penicillin and ceftriaxone. Categorical agreements (CAs) for penicillin using meningeal and nonmeningeal breakpoints were ≥90%; CAs using penicillin oral breakpoints were 84-89%. Ceftriaxone CAs using nonmeningeal and meningeal breakpoints were 68-88% for Etest, M.I.C.E., and Vitek2 with 6-12% very major errors (VMEs) using meningeal breakpoints. Sensititre CAs for ceftriaxone, amoxicillin, and meropenem were ≥90% with no VMEs. In the context of the current guidelines, there exists considerable method-dependent variability in the susceptibility of S. pneumoniae to β-lactams.