Antimicrobial prophylaxis: a critique of recent trials.

Most evaluations of antimicrobial prophylaxis have serious defects in design or fail to assess the clinical importance of observed differences. Reports that were published in the last decade and that meet stringent criteria indicate that antimicrobial prophylaxis is justified in few circumstances and nearly always only in very short courses, often just a single dose. These situations include vaginal hysterectomies (cephalosporin or penicillin), total abdominal hysterectomies (cephalosporin), high-risk cesarean sections (cephalosporin), elective colorectal surgery (oral erythromycin-neomycin, kanamycin-metronidazole, or doxycycline), vascular grafts of the abdominal aorta or lower extremity vasculature (cephalosporin), total hip replacement (cephalosporin or penicillinase-resistant penicillin), head and neck cancer surgery (cephalosporin), travelers' diarrhea (doxycycline), prevention of pneumonia due to Pneumocystis carinii in susceptible cancer patients (trimethoprim-sulfamethoxazole), and recurrent urinary tract infections in females (trimethoprim-sulfamethoxazole). Elective high-risk gastric and biliary tract surgery and prosthetic cardiac valve replacement may also merit prophylaxis, but the information is less conclusive.