Treatment of peritonitis requires prompt surgery and antibiotic therapy. It usually takes two or three days to obtain definitive results of peritoneal cultures and to adapt empirical antibiotic therapy. We assessed the potential time gain associated with direct culture of peritoneal samples on antibiotic agar (AA).Peritoneal samples from 31 consecutive patients undergoing surgery for suspected community-acquired peritonitis were cultured according to the standard method and on AA containing one of the following five regimens: amoxicillin/clavulanic acid + gentamicin, ticarcillin/clavulanic acid + gentamicin, cefotaxime +metronidazole, piperacillin/tazobactam, or ertapenem. We compared the treatment modifications made by physicians aware only of the results of the standard method with the modifications the AA method would have indicated.Fewer isolates were identified by direct culture on AA than by the standard method (17 vs. 45; p = 0.0001), but definitive results were obtained much more rapidly (median 1 [range 1-3] days vs. 3 [range 2-7] days; p0.0001). Antibiotic regimens were changed for 14 patients on the basis of the results of the standard method (broader antibiotic spectrum and narrower spectrum in seven patients each). With the AA method, these changes could have been indicated after a median of 1 (range 1-2) days instead of 4 (range 1-11) days (p = 0.0006). The AA method missed only one resistant bacterial strain and isolated nine strains not detected by the standard method, including an extended-spectrum beta-lactamase-producing Escherichia coli. A complicated outcome was more frequent in patients having isolates found with the AA but not the standard method (86% vs. 21%; p = 0.003).Use of the AA method for culture of peritoneal samples from patients with community-acquired peritonitis speeds appropriate adaptation of antibiotic therapy and warrants further investigation.