We examined the pancreatic tissue concentrations of cefazolin in ten patients undergoing pancreatectomy, and determined the optimal intraoperative time to deliver a repeat dose of cefazolin. An intravenous bolus dose of 1 g cefazolin was administered at the time of skin incision. Peripheral blood, subcutaneous adipose tissue, and peritoneal samples were obtained intraoperatively every hour for 4h after the antibiotic was first administered, and pancreatic tissue samples were obtained at the time of pancreatectomy. To determine adquate tissue levels of cefazolin, minimum inhibitory concentrations (MIC) were measured for four bacterial species, namely 360 isolates of methicillin-sensitiveStaphylococcus aureus (MSSA), 204 isolates ofKlebsiella pneumoniae, 314 isolates ofEscherichia coli, and 30 isolates ofStreptococcus spp. The antibiotic concentrations in adipose tissue and peritoneum 3 h after the administration of cefazolin were lower than the MIC80 forK. pneumoniae, E. coli, andStreptococcus spp. Most pancreatic tissue samples showed antibiotic concentrations greater than the MIC80 for these bacterial species; however, those from four patients complicated by severe chronic pancreatitis, massive intraoperative bleeding, or obesity showed concentrations lower than the MIC80. Thus, we recommend that a second dose of cefazolin be given 3 h after thefirst administration to maintain adequate levels of antibiotic activity.