Impact of a hospital sepsis management protocol on the selection of empirical antibiotics in infectious disease consultations

Abstract

It is well-established that Infectious Diseases consultation (IDC) enhances the prognosis of bloodstream infections. However, it is unclear if adoption of an institutional sepsis protocol would lead to any further improvement in a setting where IDC and infectious diseases approval (IDA) - available throughout 7 days/24 hours -are mandatory for administering broad spectrum antibiotics. We aimed to evaluate the influence of the institutional sepsis protocol developed by Department of Infectious Diseases and Clinical Microbiology on the selection of appropriate empirical antibiotics by IDC through focusing on patients who had bloodstream infections caused by Extended-spectrum β-lactamase (ESBL) producing Escherichia coli and Klebsiella pneumoniae, which poses a therapeutic challenge. One hundred and fifty-three adult patients (58 patients in the pre-protocol period and 95 patients in the post-protocol period), who received empirical antibiotic treatment for ESBL-producing E. coli and K. pneumoniae, in whom at least one systemic antibiotic was started either on the day blood cultures were drawn or not later than 24 hours were included in the study, retrospectively. The primary outcome was whether the empirical treatment regimen included a carbapenem that was accepted as the appropriate treatment based on the results of the MERINO trial. Secondary outcomes included empirical treatment based on pre-defined risk factors suggesting multidrug resistance (MDR), 30-day inpatient mortality, and appropriate antibacterial treatment according to antimicrobial susceptibility test (AST) results. The median age (Interquartile range) was 61 (48-70.5) years and 76 (49.7%) out of 153 patients were male. The patients in the post-protocol period were older compared to the pre-protocol period (54 years vs 64 years, p = 0.045). The Charlson Comorbidity Index was higher during the post-protocol period compared to the pre-protocol period (4 vs 5, p=0.038). At least one risk factor for MDR bacteria infection was present in 147 (96.1%) of the 153 patients. While the rate of risk factors for MDR bacteria infections did not differ significantly between the pre-protocol and post-protocol periods, the post-protocol period showed a significantly higher level of appropriate antibiotic treatment according to the presence of MDR risk factors compared to the pre-protocol period (44.8% vs 64.2%, p=0.019). There was a significant increase in the use of carbapenems in the post-protocol period compared to the pre-protocol period (34.5% vs. 56.8%, p=0.007). When the subgroup of patients who were likely to have infection caused by ESBL-producing bacteria is taken into consideration, the carbapenem use was more frequent in the post-protocol period (37.8% vs 68.9%, p=0.002). The rate of appropriate empirical treatment according to AST was not statistically different between pre-protocol and post-protocol period. The 30-day mortality rates were similar in both periods (24.1% vs 31.5, p=0.33). However, the rate of susceptibility to piperacillin-tazobactam was statistically higher in the pre-protocol period (82.6% vs 46.2%, p=0.016) when 39.7% of the patients received piperacillin-tazobactam as the empirical treatment. This study highlights the significance of using a structured protocol to attain appropriate empirical treatment for patients suspected of sepsis, even in a setting where IDC is readily available.