Clin Res Cardiol (2022). https://doi.org/10.1007/s00392-022-02002-5

Does a 6-month change in circulating biomarkers improve the prognostic power of baseline values for predicting cardiac MRI pathologies in patients with STEMI from the ETiCS cohort?
F. Holtkamp1, D. Grün1, A. Frey2, V. Jahns3, R. Jahns4, T. Gassenmaier5, C. W. Hamm6, S. Frantz7, T. Keller1, R. Klingenberg8, für die Studiengruppe: ETICS
1Franz-Groedel-Institut (FGI), Justus-Liebig-Universität Giessen, Bad Nauheim; 2Medizinische Klinik und Poliklinik I, ZIM Kardiologie, Universitätsklinikum Würzburg, Würzburg; 3Institut für Pharmakologie und Toxikologie, Universitätsklinikum Würzburg, Würzburg; 4Interdisziplinäre Biomaterial- und Datenbank Würzburg (IBDW), Universitätsklinikum Würzburg, Würzburg; 5Radiologie, Universitätsklinikum Würzburg, Würzburg; 6Medizinische Klinik I - Kardiologie und Angiologie, Universitätsklinikum Gießen und Marburg GmbH, Gießen; 7Medizinische Klinik und Poliklinik I, Universitätsklinikum Würzburg, Würzburg; 8Abteilung für Kardiologie, Kerckhoff Klinik GmbH, Bad Nauheim;

Objective:

We have recently shown in STEMI patients that hsTnT is the only one of four circulating biomarkers (hsTnT, NT-proBNP, ST2 and CCN1 [cellular communication network factor 1]) that strongly correlates with structural changes typical of left ventricular remodeling and heart failure after 12 months assessed by cardiac MRI. It remained unclear whether the timing of biomarker measurement has an impact on predicting these structural changes.

Methods:

Cardiac MRI data and serum were available from 49 patients with acute manifestation of STEMI from the ETiCS cohort recruited at a single center. Biomarkers were determined in serum at baseline (day 1-9), at 6 and 12 months, including hsTnT, NT-proBNP, ST2, and CCN1. Cardiac MRI 3 Tesla scans were performed to determine left ventricular ejection fraction (LVEF), left ventricular end-systolic and end-diastolic volumes (LVESV and LVEDV), infarct mass, and relative infarct mass at baseline (day 1-4), short-term (day 7-9), and long-term (12 months). The prognostic power of biomarkers was assessed using multiple logistic regression analysis by investigating their performance in predicting dichotomized cardiac MRI values based on the median (above or below the median). Three models for each biomarker were calculated by using baseline (BL), Δ value (BL - 6 months), and both values together as predictors. All models were adjusted for age and MDRD-based eGFR at initial hospitalization. Using these regression models, ROC (receiver operating characteristics) analysis was performed, and the AUC (area under the curve) with CI (confidence interval) was calculated.

Results:

We focused the current analysis on LVEF and LVESV assessed by cardiac MRI at 12 months. The logistic regression analysis identified hsTnT and ST2 as significant and strong predictors of dichotomized cardiac MRI parameters after 12 months (AUC >0.7). Both hsTnT and ST2 had a strong prognostic value in predicting a decrease in LVEF at 12 months. BL measurement (hsTnT: AUC 0.849 [CI 0.733-0.961], ST2: AUC 0.689 [CI 0.537-0.840]) as well as the Δ value BL-6M (hsTnT: AUC 0.849 [CI 0.731-0.967], ST2 0.758 [CI 0.621-0.895]) showed a high prognostic value without a statistically significant difference for the comparison of BL model vs. Δ value BL-6M model for hsTnT (p=1) and ST2 (p=0.177). The combined model including baseline and Δ value as predictors was not able to improve the ability to predict LVEF (hsTnT: AUC 0.847 [CI 0.733-0.961], p=0.918 (Figure 1); ST2: 0.743 [CI 0.601-0.884], p=0.448). Similarly, with respect to the ability of hsTnT to predict LVESV, the BL measurement of AUC 0.807 [CI 0.680-0.935] was not different from the Δ value BL-6M of AUC 0.807 [CI 0.680-0.935] (p=1), nor was the ability of the combined model AUC 0.845 [CI 0.732-0.959] different from that of the BL model (p=0.218) (Figure 2). The predictive power of ST2 BL was not improved (BL: AUC 0.714 [0.560-0.867], Δ value BL-6M: AUC 0.759 [CI 0.617-0.900]), when comparing the BL model vs. the Δ value BL-6M model for ST2 (=0.414). The combined model was not able to improve the ability of ST2 to predict LVESV (AUC 0.759 [0.617-0.900]), compared with the BL model (p=0.414).

Conclusion:

HsTnT and ST2 were identified as biomarkers that are able to predict LVEF 12 months after STEMI. Combining baseline biomarkers with their Δ changes BL-6M yielded no advantage in prognostic value; hsTnT at baseline showed good predictive power for both LVEF and LVESV.

Figure 1  Figure 2




https://dgk.org/kongress_programme/jt2022/aP2011.html