Document resource
Introduction Sepsis is a life-threatening condition in intensive care units (ICUs), where any delay in diagnosis and treatment can lead to organ dysfunction, prolonged hospital stay and increased mortality. Early identification of sepsis prior to its clinical manifestation may enable timely intervention and improve outcomes. This review aims to identify, synthesise and categorise predictor variables assessed in adult ICU patients prior to sepsis diagnosis.Methods/design This protocol, in accordance with the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P) guideline, will involve a comprehensive search of PubMed/Medline, Scopus, Embase, Web of Science, IEEE Xplore and Cochrane Library for English-language studies from database inception. The search is planned to be conducted between March 2026 and June 2026. We will include prospective, retrospective, cohort, case–control and cross-sectional studies, and both randomised and non-randomised designs. Additionally, we will consider secondary analysis of electronic health records, clinical registries and routinely acquired clinical data that determine risk factors and predictive variables that will help in the early identification of sepsis using machine learning, artificial intelligence, computational or statistical methods. Two independent reviewers will perform title and abstract screening followed by full-text review and data extraction. The risk of bias will be evaluated using appropriate tools depending on study design: the Newcastle-Ottawa Scale for observational studies, Cochrane Risk of Bias tool for randomised trials and PROBAST for prediction model studies. We will perform a narrative synthesis structured by variable-type, timing and performance metrics (area under the curve, sensitivity, ORs). Subgroup analyses will be conducted based on study design, sepsis definition, type of ICU (medical, surgical, mixed) and geographic area (high-income and low- and middle-income countries).Ethics and dissemination Findings from this systematic review will be disseminated through publication in a peer-reviewed journal and/or presentation at scientific conferences. The data used will not include individual patient data, so ethical approval is not required.PROSPERO registration number CRD420251113781.