Journal Club

An Opt-Out Protocol Can Safely Reduce Antibiotic Use in Patients With Sepsis

By A. Krishna Rao, MD, MS

While early recognition and initiation of antibiotics for sepsis is associated with increased survival, sepsis is a challenging diagnosis, and noninfectious causes of hypotension, lactic acidosis, and/or organ dysfunction can occur. Antibiotic overuse can lead to adverse events such as Clostridioides difficile infection, increased antimicrobial resistance, and even subsequent sepsis. Thus, sepsis guidelines recommend daily review of antimicrobials for opportunities for de-escalation or cessation, and better tools to help accomplish this safely are needed. In this multicenter, randomized, controlled trial by Moehring et al., the investigators studied an opt-out intervention protocol that identifies patients where safe de-escalation can occur.

The De-escalating Empiric Treatment Opting-OUt of Rx for Selected Patients (DETOURS) trial included 10 acute care hospitals in the U.S. from 2018-2020. Adult patients hospitalized outside of intensive care units were eligible for screening if they had suspected sepsis (negative blood cultures at 48-96 hours, including suspected contaminants, and active orders for broad-spectrum antibiotics). These patients were further screened by a 23-item safety checklist that included components of ongoing signs of infection (e.g., fever), concerning/inadequate microbiological data (e.g., antibiotics started prior to culture), and high-risk comorbid disease/severe illness (e.g., immunocompromised status). Patients could not have any of the items on the checklist to be eligible for randomization. In the intervention arm, a study team member (usually a clinical pharmacist) contacted the primary treating clinician to notify them that antibiotics would be stopped unless they opted out.

After screening 9,440 patients with suspected sepsis, 767 (8%) passed the safety check and were randomized, with receipt of antibiotics prior to blood culture collection being the most common reason for exclusion (35%). Fewer patients had antibiotics continued after the intervention (301 [78.6%] vs. 324 [84.4%]; P = .04). Among those that did continue, there was no significant difference in days on therapy (mean, 10.4 vs. 9.9 days; ratio, 0.68; 95% confidence interval [CI], 0.47-0.98), but fewer were on extended-spectrum antibiotics (e.g., carbapenems) targeting multidrug-resistant organisms or Pseudomonas (ratio, 0.73; 95% CI, 0.55-0.98). The desirability-of-outcome ranking analysis calculated a 52% (95% CI, 48%-56%) probability of the intervention arm having a better outcome.

This first-of-its-kind study successfully implemented a safe, low-cost method to de-escalate antibiotics in patients with suspected sepsis. Since it used an opt-out protocol, the study accomplished its aims while preserving provider autonomy. This contrasts with traditional prospective audit and feedback stewardship strategies that involve engaging with treating clinicians and developing a mutual, trusting relationship over time. However, given that many patients were not eligible for the study intervention, the safety-checklist and opt-out strategies can be an important tool going forward but are unlikely to completely supplant traditional stewardship efforts. Future studies will be needed to understand how to best integrate these strategies. This study has demonstrated it is feasible to study stewardship interventions rigorously, using randomized, patient-level designs.

(Moehring et al. Clin Infect Dis. Published online: Sept. 28, 2022.)