Study design and population

This study was conducted as part of the Molecular Diagnosis and Risk Stratification of Sepsis (MARS) project (ClinicalTrials.gov identifier {"type":"clinical-trial","attrs":{"text":"NCT01905033","term_id":"NCT01905033"}}NCT01905033), a prospective observational study conducted in two tertiary hospitals in the Netherlands between January 2011 and January 2014 (Academic Medical Center, Amsterdam, and University Medical Center Utrecht, Utrecht) [11]. For this investigation, we enrolled consecutive patients who received a sepsis diagnosis within 24 h of admission to the intensive care unit (ICU). The sepsis definition used was based on the Sepsis-3 criteria [1]; patients were retrospectively classified as meeting these criteria using prospectively collected data. In additional analyses, we compared the percentage of MS1 cells of patients with sepsis to those admitted to the ICU for non-infectious conditions. We excluded patients who were readmitted or transferred from another ICU, unless the transfer occurred on the initial day of ICU presentation. For definitions of comorbidities, organ dysfunctions and complications, see Additional file 1. Furthermore, we analyzed transcriptomic data from healthy controls (Gene Expression Omnibus access number: {"type":"entrez-geo","attrs":{"text":"GSE65682","term_id":"65682"}}GSE65682).