Sepsis triggers and tools to support early identification in healthcare settings: An integrative review

• Published in: Australian critical care Vol. 36; no. 6; pp. 1117 - 1128
• Main Authors: Kumar, Ashwani, Abbenbroek, Brett, Delaney, Anthony, Hammond, Naomi, Grattan, Sarah, Finfer, Simon
• Format: Journal Article
• Language: English
• Published: Australia Elsevier Ltd 01.11.2023
• Subjects: Lactate; qSOFA; Sepsis; SIRS; Triggers; Tools; SIRS; Sepsis; Triggers; Lactate; qSOFA
• DOI: 10.1016/j.aucc.2023.01.001
• ISSN: 1036-7314

There is no universal trigger or tool to aid sepsis diagnosis. The objective of this study was to identify triggers and tools to assist the early detection of sepsis that can be readily implemented across various health care settings. A systematic integrative review was conducted using MEDLINE, CINAHL, EMBASE, Scopus, and the Cochrane Database of Systematic Reviews. Relevant grey literature and subject-matter expert consultation also informed the review. Study types included systematic reviews, randomised controlled trials, and cohort studies. All patient populations across prehospital, emergency department, and acute hospital inpatient settings, excluding the intensive care unit, were included. Sepsis triggers and tools were evaluated for efficacy in detecting sepsis and association with process measures and patient outcomes. Methodological quality was appraised using Joanna Briggs Institute tools. Of the 124 included studies, most were retrospective cohort (49.2%) in adults (83.9%) within the emergency department (44.4%). The most commonly evaluated sepsis tools were qSOFA (12 studies) and SIRS (11 studies) with a median sensitivity of 28.0% versus 51.0% and a specificity of 98.0% versus 82.0%, respectively, for sepsis diagnosis. Lactate plus qSOFA (two studies) had a sensitivity between 57.0 and 65.5%, whereas the National Early Warning Score (four studies) demonstrated median sensitivity and specificity >80%, but the latter was considered difficult to implement. Amongst triggers, lactate (18 studies) at the threshold of ≥2.0 mmol/L showed higher sensitivity for predicting sepsis-related clinical deterioration than <2.0 mmol/L. Automated sepsis alerts and algorithms (35 studies) showed median sensitivity between 58.0 and 80.0% and specificity between 60.0 and 93.1%. There were limited data for other sepsis tools and maternal, paediatric, and neonatal populations. Overall methodological quality was high. No single sepsis tool or trigger is applicable across various settings and populations, but considering efficacy and ease of implementation, there is evidence to use lactate plus qSOFA for adult patients. More research is needed in maternal, paediatric, and neonatal populations.