Objectives: Proteomic analysis of patients with advanced cardiovascular disease was conducted to identify possible biomarkers for atrial fibrillation (AF).
Methods: A total of 123 patients undergoing cardiac surgery (22 with AF and 101 without AF) and 20 healthy subjects were recruited. Demographic data, patient history and blood samples were collected. Growth/differentiation factor 15, resistin, intracellular adhesion molecule-1, galectin-3, trefoil factor 3, tissue inhibitor of metalloproteinases 1, matrix metallopeptidase 9, high-sensitive troponin T, interleukins 6, 1α, 3, 4, 8, 20 and 22, tumour necrosis factor alpha, C-X-C motif chemokines 10 and 11, S100A6 and Type III procollagen were measured in blood serum. Differential expression between any 2 groups for any of the measured proteins was identified by fitting linear models, whereas Matthews Correlation Coefficient was used to evaluate their predictive capacity. Combined markers using more than 1 protein were attained via weighted Support Vector Machines.
Results: Although serum levels of the markers were higher in patients with cardiovascular disease than in healthy subjects, only growth/differentiation factor 15 and resistin were significantly higher in patients with AF among the subpopulation who underwent heart surgery (P = 0.029 and P = 0.007, respectively). Specific pairs of several biomarkers had mediocre predictive capacity for AF.
Conclusions: Growth/differentiation factor 15 and resistin are 2 markers that could be helpful in stratifying risk for AF in patients with cardiovascular disease. Yet, more research in terms of proteomics and investigation of possible molecular pathways implicated is required.