Abstract Introduction Various clinical characteristics have emerged as prognostic factors in patients undergoing transcatheter edge-to-edge mitral valve repair (M-TEER). The use of biomarkers for outcome prediction has enormous potential. The goal of this study was to identify novel biomarkers linked with adverse events after M-TEER. Methods We prospectively included 242 patients undergoing M-TEER at our center and obtained pre-procedural blood samples for biomarker analysis. Additionally, 94 follow-up samples were taken three months after M-TEER. A commercial multiplex protein assay (Cardiovascular III, Olink, Uppsala, Sweden) was used to quantify expression levels of 92 defined biomarkers. We analyzed differences in biomarker expression levels between patients that had died within one year after M-TEER and patients that had survived. Results 39 patients (16.3%) had died within one year after M-TEER. In these patients the most significantly upregulated biomarkers were NT-proBNP (expression 3.0-fold compared to survivors, p < 0.001), Cystatin B (CSTB; 2.0-fold, p < 0.001), fatty acid binding protein 4 (FABP4; 2.0-fold, p < 0.001), insulin-like growth factor-binding protein 1 (IBP-1; 1.9-fold, p < 0.001) and suppression of tumourigenicity 2 (ST2; 1.7-fold, p < 0.001). In non-survivors, expression levels of NT-proBNP (6.1-fold, p = 0.036), CSTB (3.2-fold, p = 0.050), FABP4 (3.7-fold, p = 0.037) and ST2 (2.1-fold, p = 0.017) were persistently elevated at the time of follow-up. Since its prognostic impact has already been described elsewhere, NT-proBNP was excluded from further analysis. Kaplan Meier analysis for 1-year mortality was carried out using tertiles of the respective biomarker expression levels. Mortality risk increased significantly with rising expression levels (CSTB: 3.4% in tertile 1, 12.2% in tertile 2, 37.9% in tertile 3; FABP4: 5.1%, 14.2% and 31.7%; IBP-1: 3.3%, 13.2% and 36.2%; ST2: 4.9%, 11.9% and 36.7%; p by log-rank-test < 0.001 for all biomarkers). Patients in which all 4 biomarkers were elevated to the highest tertile had an excessively high mortality risk of 52.9 vs. 13.5% (p by log-rank-test < 0.001). Receiver operating characteristic (ROC) analysis suggested higher predictive performances for mortality of all 4 biomarkers as compared to the EuroSCORE II (CSTB: area under the curve (AUC) 78.1%, 95% CI 70.7 – 85.5%; FABP4: AUC 71.6%, 95% CI 63.1 – 80.0%; IBP-1: AUC 73.9%, 95% CI 65.7 – 82.1%; ST2: AUC 76.2%, 95% CI 68.5 – 84.0%; EuroSCORE II: AUC 68.4%, 95% CI 59.3 – 77.5%). Conclusion This analysis has revealed novel biomarkers (CSTB, FABP4, IBP-1, ST2), which so far have not been characterized in M-TEER and suggests significant prognostic implications. Predictive performance of each of the respective biomarkers was superior to the most commonly used pre-operative risk stratification tool: the surgical EuroSCORE II. This hypothesis-generating study warrants further examination of these promising biomarkers.
Read full abstract