RCT: Assessing Procalcitonin-Based Antibiotic Management in Critically Ill Patients With Sepsis

Background: Optimal antibiotic duration for sepsis remains uncertain. Procalcitonin (PCT) and C-reactive protein (CRP) are thought to support shorter courses, but prior research was small-scale or at risk of bias. This multicenter, randomized trial (ADAPT-Sepsis) evaluated whether daily PCT- or CRP-guided protocols could reduce antibiotic use without increasing 28-day all-cause mortality in critically ill adults with suspected sepsis.

Objective: To determine if daily biomarker-guided (PCT or CRP) strategies decrease total antibiotic days among critically ill adults while maintaining acceptable 28-day mortality, compared with standard care.

Methods: From 2018 to 2024 (with enrollment paused March–August 2020 due to COVID-19), 2760 adults (≥18 years) on intravenous antibiotics for suspected sepsis (acute organ dysfunction and presumed infection) and likely to continue antibiotics for at least 72 hours were randomized across 41 UK NHS ICUs within 24 hours of antibiotic initiation. They were assigned in a 1:1:1 ratio to (1) daily PCT-guided advice (n=918), (2) daily CRP-guided advice (n=924), or (3) standard care (n=918). Biomarker results were concealed; clinicians received automated daily prompts recommending continuation or discontinuation. The co-primary outcomes were (1) total antibiotic duration (randomization to day 28) and (2) 28-day all-cause mortality. Secondary measures included antibiotic duration for the initial sepsis episode, 90-day mortality, readmissions, and length of stay.

Results: Among 2760 participants (mean age, 60.2 years; 60.3% men; ~50% with septic shock), over 96% provided 28-day data. Patients in the PCT-guided arm had a statistically significant mean reduction in total antibiotic duration vs standard care (9.8 vs 10.7 days; difference, 0.88 days; 95% CI, 0.19–1.58; p=0.01). The PCT strategy met the prespecified 5.4% noninferiority margin for 28-day mortality (20.9% vs 19.4%; absolute difference, 1.57; 95% CI, –2.18 to 5.32; p=0.02), implying noninferiority but not fully excluding a small risk of excess mortality. CRP-guided protocols did not shorten total antibiotic use (10.6 vs 10.7 days; p=0.79) and were inconclusive for noninferiority regarding mortality (21.1% vs 19.4%; difference, 1.69; 95% CI, –2.07 to 5.45; p=0.03). Notably, 90-day mortality also showed no significant differences. A post-trial commentary (PulmCCM) emphasized that some uncertainty remains with the 5.4% margin and warned that patient-level randomization could subtly discourage earlier antibiotic discontinuation in standard care, which received no explicit “stop” prompts.

Conclusions: In critically ill patients with suspected sepsis, a PCT-guided antibiotic discontinuation protocol shortened overall antibiotic use by nearly one day without exceeding the predefined noninferiority threshold for 28-day mortality. However, the chosen 5.4% margin allows for the possibility of clinically relevant harm. A CRP-guided protocol did not reduce total antibiotic use and showed inconclusive mortality findings.

Implications for Practice: Adopting PCT-based stewardship may modestly decrease antibiotic exposure without a clear short-term mortality penalty, potentially limiting antibiotic resistance. Clinicians should remain vigilant, recognizing the risk tolerance implied by the 5.4% margin. PCT results should complement, not replace, comprehensive clinical judgment.

Study Strengths and Limitations: Strengths include the large sample size, multi-center design, blinded biomarker allocation, and distinct emphasis on both effectiveness and safety outcomes. Limitations include the acceptance of a 5.4% potential excess mortality as the noninferiority threshold, uncertainty about rare but significant harms, and the possibility of bias introduced by patient-level randomization. Generalizability to lower-resource settings may also be limited.

Future Research: Further randomized trials with lower noninferiority margins or cluster-level allocation are needed to better define the safety and efficacy of PCT-guided strategies for reducing antibiotic duration in sepsis. Additional investigations are needed for long-term patient-centered outcomes, cost-effectiveness, and the role of alternative biomarkers or combined strategies in sepsis care.

Reference:

Dark P, Hossain A, McAuley DF, et al. Biomarker-Guided Antibiotic Duration for Hospitalized Patients With Suspected Sepsis: The ADAPT-Sepsis Randomized Clinical Trial. JAMA. 2024; published online December 9. DOI: http://doi.org/10.1001/jama.2024.26458

PulmCCM Commentary: “Is procalcitonin ‘safe’ to guide antibiotic use in patients with sepsis? ADAPT-Sepsis tests the strategy in the U.K., with global ambitions.” Jan 02, 2025. https://www.pulmccm.org/p/is-procalcitonin-safe-to-guide-antibiotic