Imipenem is approved by the U.S. Food and Drug Administration (FDA) for treatment of infections caused by Enterococcus faecalis. However, there are no NCCLS guidelines for testing susceptibility of enterococci against imipenem. To assess whether or not ampicillin or penicillin could be used as a surrogate for broth microdilution (BMD) testing of imipenem versus Enterococcus species, 633 strains of E. faecalis, E. faecium, and other enterococci isolated from blood cultures of patients at three geographically distinct university hospitals were tested by the NCCLS BMD and disk diffusion (DD) methods. Using FDA susceptibility breakpoints for imipenem and NCCLS breakpoints for penicillin and ampicillin, categorical agreement (CA) for penicillin-imipenem and ampicillin-imipenem tested with E. faecalis and E. faecium by BMD was >/=94% but was </=90% for other enterococci. Using the DD method, CA for ampicillin-imipenem tested with E. faecalis and E. faecium was >/=98% and was 92% for other enterococci; CA for penicillin-imipenem was 91% for E. faecalis, 98% for E. faecium, and 87% for other enterococci. Further analysis showed that testing E. faecalis with ampicillin resulted in no false-susceptible (FS) or false-resistant (FR) results by BMD, no FS results by DD, and a single FR result by DD (0.2%), whereas testing with penicillin resulted in no FS results by BMD or DD and two FR results by BMD (0.4%). For E. faecium and other enterococci, the combination of FS and FR results was such that surrogate testing with penicillin or ampicillin appears not to be sufficiently reliable to be used clinically. We conclude that ampicillin is an accurate predictor of the in vitro activity of imipenem against E. faecalis.