Complement activation characterizes a high mortality subgroup in pulmonary embolism: An observational proteomic study

Background Complement and coagulation pathways are mechanistically linked in thromboinflammation, yet the contribution of complement activation to pulmonary embolism (PE) pathophysiology and risk stratification remains largely unexplored. We used plasma proteomics and functional assays to identify thromboinflammatory endotypes in patients with suspected PE. Methods We enrolled 196 patients referred for CT pulmonary angiography at Uppsala University Hospital (2019–2022). Plasma was analyzed using the Olink Explore 384 Inflammation II panel. Differentially expressed proteins (PE vs no-PE) were used for supervised feature selection, followed by unsupervised clustering to define diagnostic clusters in all patients and prognostic endotypes in confirmed PE cases. Complement and coagulation activation were validated using functional assays. Results Eighty-nine patients (45%) had confirmed PE. Twenty-one proteins differed between PE and non-PE (FDR < 0.05), including complement C9 (log2FC 0.53, FDR 2.7 × 10 −5 ), associated with increased soluble C5b-9 (sC5b-9) supporting terminal complement activation. Three diagnostic clusters were identified; the PE-enriched cluster showed higher mortality and evidence of complement and coagulation activation (increased sC5b-9 and thrombin-antithrombin complexes), with intermediate native complement protein levels consistent with consumption. Among confirmed PE cases, three prognostic endotypes emerged, including a high-mortality cluster (50% one-year mortality) with reduced complement protein levels, without parallel increase in functional activation markers. Conclusions Targeted proteomic profiling identified elevated terminal complement C9 in PE, and evidence of complement and coagulation activation, alongside reduced native complement levels. Complement protein levels were also reduced in the high-mortality prognostic endotype. These findings support proteomic endotyping as a biologically informative approach to PE risk stratification.