Real-world clinical evaluation of the Inflammatix TriVerity test for sepsis at the emergency department: an Acutelines study.

Early sepsis recognition is challenging due to its heterogeneous presentations and limited rapid diagnostic and prognostic tools. The TriVerityTM Test, a high-multiplex host-response mRNA point-of-care test, supports clinical decision-making by estimating infection likelihood and illness severity. This study assesses TriVerity's accuracy in Dutch emergency department (ED) patients. Adult ED patients with suspected infection and Systemic Inflammatory Response Syndrome score (SIRS) ≥2 or quick Sequential Organ Failure Assessment (qSOFA) ≥2 were prospectively included by the Acutelines data/biobank. The TriVerity test provides three scores: likelihood of bacterial infection, viral infection, and illness severity. Infection status was determined post hoc by independent physicians through clinical consensus and forced adjudication. Illness severity was defined as ICU admission or death within 7 days after ED admission. Among the 467 patients included, 234 (50%) had a bacterial infection, 80 (17%) had a viral infection, 86 (18%) had a bacterial-viral co-infection, and 67 (14%) had no infection. For bacterial infections, the TriVerity assay demonstrated a rule-in specificity of 89-94% and a rule-out sensitivity of 87-92%, depending on the adjudication method (forced vs. consensus). For viral infections, the rule-in specificity was 92-95%, and the rule-out sensitivity ranged from 76 to 92%. The illness severity score showed a rule-in specificity of 83-86% and a rule-out sensitivity of 85-87%. The TriVerity test demonstrated high specificity and sensitivity for both bacterial and viral infections, with high sensitivity and modest positive predictive value in identifying patients at risk of developing severe illness.