Three hundred and twenty-two (322) clinical isolates were collected from patients admitted to intensive care units (ICUs) at eight Swedish hospitals between December 1996 and December 1998. Of the isolates, 244 (76%) were Enterococcus faecalis, 74 (23%) were Enterococcus faecium and four (1%) were other Enterococcus spp. MICs of ampicillin, imipenem, meropenem, piperacillin/tazobactam, ciprofloxacin, trovafloxacin, clinafloxacin, gentamicin, streptomycin, vancomycin, teicoplanin, quinupristin/dalfopristin, linezolid and evernimicin were determined by Etest. Susceptible and resistant isolates were defined according to the species-related MIC breakpoints of the British Society for Antimicrobial Chemotherapy (BSAC), the National Committee for Clinical Laboratory Standards (NCCLS) and the Swedish Reference Group for Antibiotics (SRGA). Tentative breakpoints were applied for new/experimental antibiotics. Multidrug resistance among enterococci in ICUs is not uncommon in Sweden, particularly among E. faecium, and includes ampicillin resistance and concomitant resistance to fluoroquinolones. Almost 20% of E. faecalis isolates showed high-level resistance to gentamicin and concomitant resistance to fluoroquinolones. Vancomycin-resistant enterococci were only found sporadically. Among the new antimicrobial agents, linezolid and evernimicin showed the best activity against all enterococcal isolates. There was good concordance between the BSAC, NCCLS and SRGA breakpoints in detecting resistance. When applying the SRGA breakpoints for susceptibility, isolates were more frequently interpreted as intermediate. This might indicate earlier detection of emerging resistance using the SRGA breakpoint when the native population is considered susceptible, but with the risk that isolates belonging to the native susceptible population will be incorrectly interpreted as intermediate.