Whole blood gene expression profiles distinguish clinical phenotypes of venous thromboembolism

Introduction Recurrent venous thromboembolism (VTE) occurs infrequently following a provoked event but occurs in up to 30% of individuals following an initial unprovoked event. There is limited understanding of the biological mechanisms that predispose patients to recurrent VTE. Objectives To identify whole blood gene expression profiles that distinguished patients with clinically distinct patterns of VTE. Patients/Methods We studied 107 patients with VTE separated into 3 groups: (1) 'low-risk' patients had one or more provoked VTE; (2) 'moderate-risk' patients had a single unprovoked VTE; (3) 'high-risk' patients had ≥2 unprovoked VTE. Each patient group was also compared to twenty-five individuals with no personal history of VTE. Total RNA from whole blood was isolated and hybridized to Illumina HT-12 V4 Beadchips to assay whole genome expression. Results Using class prediction analysis, we distinguished high-risk patients from low-risk patients and healthy controls with good receiver operating curve characteristics (AUC = 0.81 and 0.84, respectively). We also distinguished moderate-risk individuals and low-risk individuals from healthy controls with AUC's of 0.69 and 0.80, respectively. Using differential expression analysis, we identified several genes previously implicated in thrombotic disorders by genetic analyses, including SELP, KLKB1, ANXA5, and CD46. Protein levels for several of the identified genes were not significantly different between the different groups. Conclusion Gene expression profiles are capable of distinguishing patients with different clinical presentations of VTE, and genes relevant to VTE risk are frequently differentially expressed in these comparisons.