OBJECTIVE: The goal of antibiotic therapy for surgical sepsis is to kill bacteria that intermittently or continuously reach the bloodstream from the residue of an operatively treated focus. While sepsis and conditions leading to sepsis compromise the immune system, antibiotics may become a fundamental determinant of the host's defense. No data from sound prospective randomized clinical antibiotic trials dealing with sepsis are available. Therefore we tested the hypothesis that treatment recommendations can be based on pharmacodynamics comparing in vitro activity of commonly used antimicrobials with concentrations sustained in vivo to provide for full coverage for bacteria of concern. RESULTS: The application of strict criteria for antibiotic choice to avoid selection of primary resistant strains reveals that most commonly used antibiotics render insufficient activity to eliminate pathogens that commonly cause surgical sepsis. Antibiotics that sustain in vivo concentration exceeding fourfold the MIC100 (highest minimal inhibitory concentration for all (100%) species tested) of Escherichia coli, for example, are 400 mg ciprofloxacin IV (MIC100 of 1224 strains=0.06 mg/dl, in vivo concentration=1 mg/dl for 12 h), and 1000 mg imipenem/cilastatin (MIC100 of 3142 strains=0.14 mg/dl, in vivo concentration=2 mg/dl for 6 h). The third choice is one of the fourth- or, less convincingly, third-generation cephalosporins. Similar data for most pathogens causing sepsis are provided. First- and second-generation cephalosporins and penicillin beta-lactamase inhibitor combinations generally do not achieve sufficient concentrations to cover the most important pathogens of sepsis. CONCLUSION: Sepsis is defined as a whole body's inflammatory response that is characterized by systemic signs and symptoms secondary to a focal infection. While many antibiotic trials have dealt with a focal infection, no prospective randomized antibiotic trial has dealt with sepsis per se. Antibiotic trials on focal infections generally exclude patients when their focal infection has progressed to sepsis. To circumvent the lack of controlled clinical trials we show that pharmacodynamics may provide sound foundation for antibiotic choice for sepsis.