Assessment of 352 plasma biomarkers for predicting ischaemic stroke risk in atrial fibrillation patients without anticoagulation - A proteomic profiling study
European Heart Journal

Abstract
Atrial fibrillation (AF) is a major risk factor for ischaemic stroke. Recent advancements in proteomics enable high-throughput screening of proteins using small plasma samples. However, studies on biomarkers in AF patients without oral anticoagulation are lacking.
This proteomics screening study aims to identify plasma biomarkers associated with ischaemic stroke risk in AF patients without oral anticoagulation, thereby improving the mechanistic understanding of this complication.
This case-cohort design study included 166 cases with ischaemic stroke and a random sample of 1,437 controls from patients in the ACTIVE-A and AVERROES trials randomized to aspirin, with baseline plasma samples available. All ischaemic stroke outcomes were independently adjudicated. A total of 352 unique biomarkers were measured with OLINK Target 96 panels using proximity extension assay (CVDII, CVDIII, inflammation, and oncology), along with conventional and prototype immunoassays (Roche Diagnostics). The association between biomarkers and outcomes was evaluated by Random Survival Forest analysis, including Boruta and Cox models adjusted for clinical characteristics, comorbidities, and kidney function.
Out of the 352 biomarkers, the top 15 risk predictors for ischaemic stroke in AF according to Random Survival Forest are shown in the Figure. The corresponding Boruta and Cox regression model p-values and hazard ratios (95% CI) for an interquartile difference validated 6 of these biomarkers (Table).
In AF patients not receiving oral anticoagulation, out of 352 plasma biomarkers, the six biomarkers most strongly associated with subsequent ischaemic stroke represent cardiac dysfunction/injury (NT-proBNP and hs-troponin T), atrial stress (BMP10), angiogenesis (CYR61), and myocardial remodelling/fibrosis (PRELP and WISP-1). The latter three proteins emerged as novel markers associated with ischaemic stroke in AF.





