Design and evaluation of clinical trials of antimicrobial agents in surgery

Comparative trials of antibiotic therapy must be properly designed to produce meaningful results. Trials in the treatment of intra-abdominal infections are especially difficult to perform. Problems include finding adequate samples of comparable patients, establishing objective criteria by which to assess outcomes and controlling for the effects of nonexperimental factors. These difficulties are among the reasons many comparative trials of antibiotic therapy for intra-abdominal infection fail to find a difference in efficacy between the experimental regimens. Those trials that do demonstrate a therapeutic difference usually compare an adequate regimen against one that is clearly deficient; for example, one that lacks coverage of anaerobic organisms. For hospital acquired or surgical infections, the treatment regimen with the most extensive published support is the combination of an aminoglycoside and clindamycin. Single agent therapy for intra-abdominal infections may also be appropriate. Further clinical studies are needed, but two of the most promising agents are ticarcillin and clavulanic acid and imipenem and cilastatin.