Background This study aimed to explore predictive biomarkers of coronary collateralization in patients with chronic total occlusion. Methods and Results By using a microarray expression profiling program downloaded from the Gene Expression Omnibus database, weighted gene coexpression network analysis was constructed to analyze the relationship between potential modules and coronary collateralization and screen out the hub genes. Then, the hub gene was identified and validated in an independent cohort of patients (including 299 patients with good arteriogenic responders and 223 patients with poor arteriogenic responders). Weighted gene coexpression network analysis showed that SERPING1 in the light-cyan module was the only gene that was highly correlated with both the gene module and the clinical traits. Serum levels of serpinG1 were significantly higher in patients with bad arteriogenic responders than in patients with good arteriogenic responders (472.53±197.16 versus 314.80±208.92 μg/mL; P<0.001) and were negatively associated with the Rentrop score (Spearman r=-0.50; P<0.001). Receiver operating characteristic curve analysis indicated that the area under the curve was 0.77 (95% CI, 0.72-0.81; P<0.001) for serum serpinG1 in prediction of bad arteriogenic responders. After adjusting for traditional cardiovascular risk factors, serum serpinG1 levels (per SD) remained an independent risk factor for bad arteriogenic responders (odds ratio, 2.20 [95% CI, 1.76-2.74]; P<0.001). Conclusions Our findings illustrate that SERPING1 screened by weighted gene coexpression network analysis was associated with poor collateralization in patients with chronic total occlusion.
Keywords: chronic total occlusion; coronary artery disease; coronary collateralization; weighted gene coexpression network analysis.