Procalcitonin-guided antibiotic therapy in critically ill adults: a meta-analysis

• Main Authors: Tao Zhang, Yan Wang, Qianting Yang, Yalin Dong
• Format: Article
• Language: English
• Published: BMC 2017-07-01
• Series: BMC Infectious Diseases
• Subjects: Procalcitonin; ICU; Meta-analysis; Trial Sequential Analysis
• Online Access: http://link.springer.com/article/10.1186/s12879-017-2622-3

Abstract Background As a novel biomarker of inflammation, procalcitonin (PCT) has proven useful to guide antibiotic therapy in intensive care unit (ICU). However, there are controversial on mortality. The aim of this study was to evaluate the utility of PCT-guided antibiotic therapy in critically ill adults and determine whether studies are sufficient. Methods A systematic search in PubMed, Embase and Cochrane was performed. We included only randomized controlled trials which compared the safety and efficacy between PCT-guided or standard antibiotic therapy groups in ICU adults. Trial sequential analysis and GARDE approach were performed. Results Fifteen studies met our criteria for inclusion finally, with a cumulative number of 5486 ICU patients. There was no difference in 28-day mortality between two compared groups (P = 0.626), but significant decreases were observed in the duration of antibiotic therapy for the first episode of infection (P < 0.001) and length of hospitalization (P = 0.049). No significant deference was found in secondary endpoints except total duration of antibiotic therapy (P < 0.001). TSA revealed that the pooled sample sizes of 28-day mortality and the duration of antibiotic therapy for the first episode of infection exceeded the estimated required information size, but not the length of hospitalization. Conclusions PCT-guided therapy is a better and safer algorithm to be applied into ICU patients, which appears no effect on 28-day mortality while performing preferable utility in reducing the duration of antibiotic therapy for the first episode of infection. More studies on these endpoints were not recommended.