We investigated the in vitro susceptibilities of 23 Propionibacterium acnes ophthalmic isolates to ertapenem, meropenem, and cefepime by utilizing the Etest. The MICs ranged from 0.094 g/ml to 0.75 g/ml, 0.094 g/ml to 1.5 g/ml, and 1 g/ml to 12 g/ml for ertapenem, meropenem, and cefepime, respectively. Based on our excellent in vitro carbapenem susceptibility results, in vivo studies using ertapenem and meropenem in a rabbit model of P. acnes endophthalmitis are warranted.Propionibacterium acnes, an anaerobic, gram-positive bacillus, is known to be one of the most frequent causes of postsurgical endophthalmitis. Direct intravitreal antibiotic instillation has been shown to be safe and effective for the treatment of many aerobic organisms (8,9,11). Optimal treatment of P. acnes, which has been shown to cause a chronic, indolent infection after cataract surgery and lens implantation despite intravitreal vancomycin instillation, vitrectomy, and removal and exchange of the intraocular lens, remains unknown (2, 4). Despite recent research indicating the susceptibility of P. acnes to older beta-lactams, fluoroquinolones, and macrolides, in vivo success with antibiotics that are currently available for intravitreal instillation has been inconsistent, and additional antibiotic alternatives are needed (6, 10).In the present study, the in vitro susceptibilities of P. acnes ophthalmic isolates to two of the newer carbapenems, ertapenem and meropenem, and the "fourth-generation" cephalosporin cefepime were investigated. Cefepime was of interest because of previous reports demonstrating very good in vitro activity against P. acnes using the broad-spectrum cephalosporin ceftriaxone (6, 10). In addition, cefepime provides a broad spectrum of activity and may be useful in the treatment of P. acnes infections. Further studies using a rabbit model of P. acnes endophthalmitis would be justified if this study illustrates in vitro susceptibility to any of the three study drugs.The P. acnes isolates were identified using the Vitek Rapid ANI system (bioMérieux). The MIC (in g/ml) of each of the test antibiotics was determined using the Etest (AB Biodisk of North America, Piscataway, NJ) (3, 10). The Division of Microbiology at the Long Island Jewish Medical Center, New Hyde Park, N.Y., supplied 23 clinical ophthalmic isolates of P. acnes obtained from eyes of infected patients in addition to control strains (P. acnes ATCC 11827, P. acnes ATCC 11828, and Bacteroides fragilis ATCC 25285) used in this study. Twelve of the 23 P. acnes strains were the same isolates used in our previous study (6). Ceftriaxone, clindamycin, and metronidazole were also added to the antibiotic test battery. These antibiotics had been evaluated against some of the P. acnes strains that were used in our previous study and served as additional experimental controls. Susceptibility testing was performed using the Etest. In our previous study, equivalent MIC results were obtained using the Etest and the Clinical and Laboratory Standards Institute (CLSI) (formerly ...