Decision-Making Analysis for Selection of Antibiotic Treatment in Intra-Abdominal Infection Using Preference Measurements

Abstract

Background: Antimicrobial therapy of abdominal infections is important to the prognosis of affected patients. The choice of antimicrobial therapy must consider effectiveness, safety, cost, and antibiotic resistance, among numerous factors. However, in reality, decisions are made assuming bioequivalence between regimens, without considering the specific attributes of any particular regimen. The objective was to determine the best antibiotic regimen for patients with community-acquired abdominal infection on the basis of a decision analysis that included effectiveness as well as safety, measured as adverse effects. Methods: A decision tree was built using information from a systematic review of the literature on the effectiveness of antimicrobial regimens tested in randomized clinical trials (RCTs) and the frequency and severity of adverse effects. The quality of the articles was assessed with the Oxford criteria for RCTs. The main outcome was preferences reported by surgeons, measured on a numeric scale. Preferences were obtained using a standard survey that reported each adverse effect with its respective intensity, reversibility, sequelae, duration of symptoms, and necessity for change of antibiotic. Each of the surgeons had to assign a value blindly from 0 to 10, where 10 was the most severe. A sensitivity analysis was conducted varying the frequency of adverse effects. Results: The regimens analyzed were amikacin-metronidazole, amikacin-clindamycin, ciprofloxacin-metronidazole, ampicillin-sulbactam, ceftriaxone-metronidazole, piperacillintazobactam, and ertapenem. The perceived severity of adverse effects reported were: Acute neuromuscular blockade (8.0), severe allergic reaction (7.5), ototoxicity (7.4), nephrotoxicity (7.1), antibiotic-associated colitis (7.0), peripheral neuropathy (5.3), general neurological symptoms (4.9), gastrointestinal symptoms (3.1), and other general symptoms (2.6). Favored regimens were ceftriaxone-metronidazole (1.15), ampicillin-sulbactam (1.24), piperacillin-tazobactam (1.27) and ertapenem (1.28). These strategies dominated the other therapeutic schemes. Sensitivity analysis showed no changes in the dominance reported when the frequency of adverse effects was maintained in the known clinical range. Conclusions: Antibiotic regimens that contain aminoglycosides are not bioequivalent to those without aminoglycosides when effectiveness and adverse effects are considered simultaneously. Antibiotic regimens that do not use aminoglycosides must be the first line of treatment for abdominal sepsis acquired in the community.