Methods: Patients with Gram-negative BSI who were hospitalized for < 14 days at Palmetto Health Richland and Baptist Hospitals in Columbia, SC, USA from January 1, 2010 to December 31, 2013 and discharged on oral antibiotics were included in the study. The cohort was stratified into three groups based on bioavailability of oral antibiotics prescribed (high: ≥ 95%, moderate: 75-94% and low: < 75%). Kaplan-Meier analysis and multivariate Cox proportional hazards regression were used to examine clinical failure, which was defined as either mortality or recurrent infection within 90 days of initial BSI.
Results: Among a total of 322 patients, high, moderate, and low bioavailability oral antibiotics were prescribed to 89, 162, and 71 patients, respectively, for definitive therapy of Gram-negative BSI. Overall, the median age was 64 years, 193 (60%) were women, 221 (69%) had Escherichia coli BSI and 227 (70%) had urinary source of infection. There were no major differences in baseline demographic and clinical characteristics across the three groups. The median treatment duration with appropriate intravenous antibiotics before switching to oral agents was five days in the three groups. Clinical failure rates were 1.4% (95% confidence intervals [CI] 0-4%) in the high bioavailability group, 13% (95% CI 7-19%) in the moderate bioavailability group and 13% (95% CI 5-21%) in the low bioavailability group (p=0.02). The risk of clinical failure in multivariate Cox model was higher in patients who received antibiotics with moderate (hazard ratio [HR] 13.0, 95% CI 2.6-236.5, p<0.001) and low bioavailability (HR 14.9, 95% CI 2.6-281.8, p=0.001) as compared to those who received oral antibiotics with high bioavailability.
Conclusion: These data demonstrate effectiveness of oral antibiotics with high bioavailability for definitive therapy of Gram-negative BSI. The risk of clinical failure increases as bioavailability of the oral regimen declines.
P. B. Bookstaver, Forest Labs: Grant Investigator and Scientific Advisor , Consulting fee
J. Kohn, None
H. Albrecht, None
M. Al-Hasan, None