De-escalation practice pattern in an Indian intensive care unit

Introduction

Antibiotic de-escalation is thought to be beneficial by reducing the selection pressure for resistance. This study was carried out to identify variables that influence de-escalation practices.

Methods

Prospective observational study during a 1-year period (July 2008 to June 2009) in a 50-bed ICU in a tertiary care hospital. Any patient admitted to the ICU during this period in whom at least one specimen was sent for microbiological culture at or before starting antibiotics was included in the study and subsequently followed up for antibiotic change according to the culture sensitivity report. Antibiotics covering Gram-negative organisms were ranked as per Figure 1.

Figure 1

Antimicrobial therapy ranking.

Results

Seven hundred and seventy-eight patients were included, of whom cultures were positive in 551 (70.8%) and negative in 227 (29.2%) cases. In 350 (44.9%) patients, neither escalation nor de-escalation of therapy was done. Overall escalation of therapy occurred in 192 (24.7%) patients and de-escalation in 236 (30.3%). The mortality rate was lowest among patients in whom therapy was de-escalated (8.9%) compared with categories of no change (14.2%) or escalation (23.4%). De-escalation occurred more frequently among patients in whom no pathogen was isolated (45.8%) compared with culture-positive cases (23.9%) and in 18.3% of patients with growth of drug-resistant pathogens, compared with 38.9% of patients with susceptible pathogens. De-escalation occurred most frequently where an antibiotic having only Gram-positive coverage was included in the initial empiric therapy (60.3%), and in cases where third-generation cephalosporin was the initial empiric therapy (54.4%).

Conclusions

This study highlights no change of therapy as the most prevalent practice pattern of antibiotic use. Mortality remained low in patients in whom therapy was de-escalated. Variables favoring de-escalation practices were non-isolation of pathogen, growth of susceptible organisms, and stopping of Gram-positive coverage.