The definitive diagnosis of spontaneous bacterial peritonitis is made by a positive ascitic fluid culture. Causative organisms cannot be isolated in up to 65% of patients with well-defined spontaneous bacterial peritonitis, probably due to inadequate ascites culture techniques. We prospectively compared two ascites culture methods: conventional (on chocolate agar and thioglycolate broth) and modified (inoculation of 10 ml of ascites in a tryptic soy broth blood culture bottle at the patient's bedside). In a 10-month period, 31 cirrhotic patients met our diagnostic criteria for spontaneous bacterial peritonitis; both culture methods were performed on their ascitic fluid. The conventional method grew bacteria in only 16 of the 31 episodes (52%), whereas the modified method grew bacteria in 25 (81%), a significantly higher sensitivity (P less than 0.05). The modified method also shortened significantly the time for detection of bacterial growth. We conclude that ascites inoculation into a blood culture bottle at the patient's bedside should be the routine method for ascites culture.