Antibiotic appropriateness for Gram-negative bloodstream infections: impact of infectious disease consultation

Abstract

Background

We investigated the role of infectious disease consultation (IDC) on therapeutic appropriateness in Gram-negative bloodstream infections (GNBSIs) in a setting with a high proportion of antibiotic resistance. Secondary outcomes were in-hospital mortality and the impact of rapid diagnostic tests (RDTs).

Methods

Retrospective study on hospitalised patients with GNBSIs. Therapy was deemed appropriate if it had the narrowest spectrum considering infection and patients’ characteristics. Interventional-IDC (I-IDC) group included patients with IDC-advised first appropriate or last non-appropriate therapy. Time to first appropriate therapy and survival were evaluated by Kaplan-Meier curves. Factors associated with therapy appropriateness were assessed by multivariate Cox proportional-hazard models.

Results

471 patients were included. High antibiotic resistance rates were detected: quinolones 45.5%, third-generation cephalosporins 37.4%, carbapenems 7.9%. I-IDC was performed in 31.6% of patients (149/471), RDTs in 70.7% (333/471). The 7-day probability of appropriate treatment was 91.9% (95% confidence interval [95%CI]: 86.4–95.8%) vs. 75.8% (95%CI: 70.9–80.4%) with and without I-IDC, respectively (p-value = 0.0495); 85.5% (95%CI: 81.3–89.1%) vs. 69.4% (95%CI: 61.3–77.2%) with and without RDTs, respectively (p-value = 0.0023). Compared to RDTs alone, the combination with I-IDC was associated with a higher proportion of appropriate therapies at day 7: 81.9% (95%CI: 76.4–86.7%) vs. 92.6% (95%CI: 86.3–96.7%). At multivariate analysis, I-IDC and RDTs were associated with time to first appropriate therapy [adjusted hazard-ratio 1.292 (95%CI: 1.014–1.647) and 1.383 (95%CI: 1.080–1.771), respectively], with no impact on mortality.

Conclusions

In a setting with a high proportion of antibiotic resistance, IDC and RDTs were associated with earlier prescription of appropriate therapy in GNBSIs, without impact on mortality.

Keywords:

• Gram-negative bacterial infections
• bloodstream infection
• inappropriate treatment
• molecular diagnostic techniques

Disclosure statement

A. Castagna has received consultancy payments and speaking fees from Bristol-Myers Squibb, Gilead Sciences, ViiV Healthcare, Merck Sharp & Dohme and Janssen-Cilag. All other authors have no relevant financial or non-financial interests to disclose.