Likelihood of Inadequate Treatment A Novel Approach to Evaluating Drug-Resistance Patterns

Abstract

Objective.To provide a novel way to predict the likelihood that antibiotic therapy will result in prompt, adequate therapy on the basis of local microbiological data.Design and Setting.Prospective study conducted at 3 medical intensive care units at the Viennese General Hospital, a tertiary care medical university teaching hospital in Vienna, Austria.Patients.One hundred one patients who received mechanical ventilation and who met the criteria for having ventilator-associated pneumonia.Design.Fiberoptic bronchoscopic examination was performed, and bronchoalveolar samples were collected. Samples were analyzed immediately by a single technician. Minimum inhibitory concentrations were determined for imipenem, cephalosporins (cefepime and cefpirome), ciprofloxacin, and piperacillin-tazobactam, and drug resistance rates were calculated. These drug resistance rates were translated into the likelihood of inadequate therapy (LIT; the frequency of inadequately treated patients per antibiotic and drug-resistant strain), cumulative LIT (the cumulative frequency of inadequately treated patients), and syndrome-specific LIT.Results.Amongthe 101 bronchoalveolar samples, culture yielded significant (at least 1 × 104 colony-forming units per raL) polymicrobial findings for 34 and significant monomicrobial findings for 31; 36 culture results were negative. Of the isolates from patients with ventilator-associated pneumonia who had monomicrobial culture findings, 33% were gram-positive bacteria and 20% were gram-negative bacteria. LIT suggested that 1 of 2 patients was treated inadequately for Pseudomonas aeruginosa infection. The LIT for patients with ventilator-associated pneumonia revealed that the rank order of antibiotics for appropriate therapy was (1) imipenem, (2) cephalosporins, (3) ciprofloxacin, and (4) piperacillin-tazobactam. These calculations were based solely on microbiological data.Conclusions.The novel ratio LIT may help clinicians use microbiological data on drug resistance to predict which antimicrobial agents will provide adequate therapy. In daily practice, this new approach may be helpful for choosing adequate antimicrobial therapy.