Frequency of empiric antibiotic de-escalation in an acute care hospital with an established Antimicrobial Stewardship Program

BACKGROUND: Expanding antimicrobial resistance patterns in the face of stagnant growth in novel antibiotic production underscores the importance of antibiotic stewardship in which de-escalation remains an integral component. We measured the frequency of antibiotic de-escalation in a tertiary care medical center with an established antimicrobial stewardship program to provide a plausible benchmark for de-escalation. METHODS: A retrospective, observational study was performed by review of randomly selected electronic medical records of 240 patients who received simultaneous piperacillin/tazobactam and vancomycin from January to December 2011 at an 885-bed tertiary care medical center. Patient characteristics including antibiotic regimen, duration and indication, culture results, length of stay, and hospital mortality were evaluated. Antibiotic de-escalation was defined as the use of narrower spectrum antibiotics or the discontinuation of antibiotics after initiation of piperacillin/tazobactam and vancomycin therapy. Subjects dying within 72 h of antibiotic initiation were considered not de-escalated for subsequent analysis and were subtracted from the study population in determining a modified mortality rate. RESULTS: The most commonly documented indications for piperacillin/tazobactam and vancomycin therapy were pneumonia and sepsis. Of the 240 patients studied, 151 (63%) had their antibiotic regimens de-escalated by 72 h. The proportion of patients de-escalated by 96 h with positive vs. negative cultures was similar, 71 and 72%, respectively. Median length of stay was 4 days shorter in de-escalated patients, and the difference in adjusted mortality was not significant (p = 0.82). CONCLUSIONS: The empiric antibiotic regimens of approximately two-thirds of patients were de-escalated by 72 h in an institution with a well-established antimicrobial stewardship program. While this study provides one plausible benchmark for antibiotic de-escalation, further studies, including evaluations of antibiotic appropriateness and patient outcomes, are needed to inform decisions on potential benchmarks for antibiotic de-escalation.