Use of serial changes in biomarkers vs. baseline levels to predict left ventricular remodelling after STEMI.

AIMS: Cellular communication network factor 1 (CCN1) is an independent predictor of MACE after ACS and elevated levels correlated with infarct size after STEMI. We compared the prognostic accuracy of baseline levels of CCN1, NT-proBNP, hsTnT, and ST2 and changes in levels over time to predict the development of structural and functional alterations typical of LV remodelling. METHODS: Serial 3-T cMRI scans were performed to determine LVEF, LVEDV, LVESV, infarct size, and relative infarct size, which were correlated with serial measurements of the four biomarkers. The prognostic significance of these biomarkers was assessed by multiple logistic regression analysis by examining their performance in predicting dichotomized cardiac MRI values 12 months after STEMI based on their median. For each biomarker three models were created using baseline (BL), the Δ value (BL to 6 months), and the two values together as predictors. All models were adjusted for age and renal function. Receiver operator curves were plotted with area under the curve (AUC) to discriminate the prognostic accuracy of individual biomarkers for MRI-based structural or functional changes. RESULTS: A total of 44 predominantly male patients (88.6%) from the ETiCS (Etiology, Titre-Course, and Survival) study were identified at a mean age of 55.5 ± 11.5 (SD) years treated by successful percutaneous coronary intervention (97.7%) at a rate of 95.5% stent implantation within a median pain-to-balloon time of 260 min (IQR 124-591). Biomarkers hsTnT and ST2 were identified as strong predictors (AUC > 0.7) of LVEDV and LVEF. BL measurement to predict LVEF [hsTnT: AUC 0.870 (95% CI: 0.756-0.983), ST2: AUC 0.763 (95% CI: 0.615-0.911)] and the Δ value BL-6M [hsTnT: AUC 0.870 (95% CI: 0.756-0.983), ST2: AUC 0.809 (95% CI: 0.679-0.939)] showed a high prognostic value without a significant difference for the comparison of the BL model vs. the Δ-value model (BL-6M) for hsTnT (P = 1) and ST2 (P = 0.304). The combined model that included baseline and Δ value as predictors was not able to improve the ability to predict LVEF [hsTnT: AUC 0.891 (0.791-0.992), P = 0.444; ST2: AUC 0.778 (0.638-0.918), P = 0.799]. Baseline levels of CCN1 were closely associated with LVEDV at 12 months [AUC 0.708 (95% CI: 0.551-0.865)] and infarct size [AUC 0.703 (95% CI: 0.534-0.872)]. CONCLUSIONS: Baseline biomarker levels of hsTnT and ST2 were the strongest predictors of LVEF and LVEDV at 12 months after STEMI. The association of CCN1 with LVEDV and infarct size warrants further study into the underlying pathophysiology of this novel biomarker.

Hemodynamic markers of pulmonary vasculopathy for prediction of early right heart failure and mortality after heart transplantation.

BACKGROUND: Elevated pulmonary vascular resistance (PVR) is broadly accepted as an imminent risk factor for mortality after heart transplantation (HTx). However, no current HTx recipient risk score includes PVR or other hemodynamic parameters. This study examined the utility of various hemodynamic parameters for risk stratification in a contemporary HTx population. METHODS: Patients from seven German HTx centers undergoing HTx between 2011 and 2015 were included retrospectively. Established risk factors and complete hemodynamic datasets before HTx were analyzed. Outcome measures were overall all-cause mortality, 12-month mortality, and right heart failure (RHF) after HTx. RESULTS: The final analysis included 333 patients (28% female) with a median age of 54 (IQR 46-60) years. The median mean pulmonary artery pressure was 30 (IQR 23-38) mm Hg, transpulmonary gradient 8 (IQR 5-10) mm Hg, and PVR 2.1 (IQR 1.5-2.9) Wood units. Overall mortality was 35.7%, 12-month mortality was 23.7%, and the incidence of early RHF was 22.8%, which was significantly associated with overall mortality (log-rank HR 4.11, 95% CI 2.47-6.84; log-rank p < .0001). Pulmonary arterial elastance (Ea) was associated with overall mortality (HR 1.74, 95% CI 1.25-2.30; p < .001) independent of other non-hemodynamic risk factors. Ea values below a calculated cutoff represented a significantly reduced mortality risk (HR 0.38, 95% CI 0.19-0.76; p < .0001). PVR with the established cutoff of 3.0 WU was not significant. Ea was also significantly associated with 12-month mortality and RHF. CONCLUSIONS: Ea showed a strong impact on post-transplant mortality and RHF and should become part of the routine hemodynamic evaluation in HTx candidates.

Prognostic Power of Pulmonary Arterial Compliance Is Boosted by a Hemodynamic Unloading Test With Glyceryl Trinitrate in Heart Failure Patients With Post-capillary Pulmonary Hypertension.

Background: Pulmonary hypertension (PH) is an established risk factor in patients with heart failure (HF). However, right heart catheterisation (RHC) and vasoreactivity testing (VRT) are not routinely recommended in these patients. Methods: The primary objective of the present study was to explore the impact of VRT using sublingual glyceryl trinitrate (GTN) on transplant/ventricular assist device-free survival in HF patients with post-capillary PH. RHC parameters were correlated retrospectively with the primary outcome. Results: The cohort comprised 154 HF patients with post-capillary PH undergoing RHC with GTN-VRT at a tertiary heart failure centre. Multiple parameters were associated with survival. After adjustment for established prognosis-relevant clinical variables from the MAGGIC Score, variables with the most relevant odds ratios (OR) obtained after GTN-VRT were: calculated effective pulmonary arterial (PA) elastance (adjusted OR 2.26, 95%CI 1.30-3.92; p = 0.004), PA compliance (PAC-GTN; adjusted OR 0.45, 95%CI 0.25-0.80; p = 0.006), and total pulmonary resistance (adjusted OR 2.29, 95%CI 1.34-3.93; p = 0.003). Forest plot analysis including these three variables as well as PAC at baseline, delta PAC, and the presence of combined post- and pre-capillary PH revealed prognostic superiority of PAC-GTN, which was confirmed by Kaplan-Meier analysis. Conclusions: In our cohort of symptomatic HF patients with post-capillary PH, improved PAC after administration of GTN was associated with survival independent of established hemodynamic and clinical risk factors. VRT using GTN may be better described as unloading test due to GTN's complex effects on the circulation. This could be used for advanced prognostication and should be investigated in further studies.

Circulating Monocyte Subsets Are Associated With Extent of Myocardial Injury but Not With Type of Myocardial Infarction.

Inflammation is a hallmark of the period after a myocardial infarction (MI) that is either promoted or resolved by distinct subtypes of circulating inflammatory cells. The three main monocyte subpopulations play different roles inflammation. This study examined whether the type of MI (type 1 or type 2) or the extent of myocardial injury is associated with differences in monocyte subpopulations. For this purpose, peripheral whole blood from patients with a suspected MI was used for flow cytometric measurements of the monocyte subpopulations, and myocardial injury was classified by cardiac troponin levels in serum. In patients with acute coronary syndrome (n = 82, 62.2% male) similar proportions of the monocyte subsets were associated with the two types of MI, whereas total monocyte counts were increased in patients with substantial myocardial injury vs. those with minor injury (p = 0.045). This was accompanied by a higher proportion of intermediate (p = 0.045) and classical monocytes (p = 0.059); no difference was found for non-classical monocytes (p = 0.772). In patients with chronic coronary syndrome (n = 144, 66.5% male), an independent association with myocardial injury was also observed for classical monocytes (p = 0.01) and intermediate monocytes (p = 0.08). In conclusion, changes in monocyte subpopulation counts, particularly for classical and intermediate monocytes, were related to the extent of myocardial injury in acute and stable coronary artery disease but not to the type of MI.

Detecting myocardial scar using electrocardiogram data and deep neural networks.

Ischaemic heart disease is among the most frequent causes of death. Early detection of myocardial pathologies can increase the benefit of therapy and reduce the number of lethal cases. Presence of myocardial scar is an indicator for developing ischaemic heart disease and can be detected with high diagnostic precision by magnetic resonance imaging. However, magnetic resonance imaging scanners are expensive and of limited availability. It is known that presence of myocardial scar has an impact on the well-established, reasonably low cost, and almost ubiquitously available electrocardiogram. However, this impact is non-specific and often hard to detect by a physician. We present an artificial intelligence based approach - namely a deep learning model - for the prediction of myocardial scar based on an electrocardiogram and additional clinical parameters. The model was trained and evaluated by applying 6-fold cross-validation to a dataset of 12-lead electrocardiogram time series together with clinical parameters. The proposed model for predicting the presence of scar tissue achieved an area under the curve score, sensitivity, specificity, and accuracy of 0.89, 70.0, 84.3, and 78.0%, respectively. This promisingly high diagnostic precision of our electrocardiogram-based deep learning models for myocardial scar detection may support a novel, comprehensible screening method.

Myeloid-related protein 8/14 and high-sensitivity cardiac troponin I to differentiate type 2 myocardial infarction.

Myocardial infarction (MI) not only defines acute MI with obstructed coronary arteries (T1MI) but also myocardial necrosis caused by myocardial oxygen supply/demand mismatch as type 2 MI (T2MI); only T1MI patients benefit from an early invasive management. Myeloid-related protein(MRP)-8/14 is a biomarker described in various inflammatory diseases and in MI patients. Here we evaluate the potential of MRP-8/14 and high-sensitivity troponin I (hs-cTnI) to differentiate T2MI from T1MI. Patients with final diagnosis NSTEMI (n = 254; 33.1% female) enrolled in a prospective biomarker registry between 08/2011 and 10/2016 were analysed. Median baseline MRP-8/14 levels were higher in T2MI (n = 55; 3.37(1.88-6.48)µg/mL) than in T1MI (n = 199; 2.4 [1.4-3.79]µg/mL) (p = .013) patients, in contrast to hs-cTnI (T2MI:52[11.65-321.4]ng/L vs. T1MI:436.5 [61.25-1973.8]ng/L; p < .001). To detect the strength of this association odds ratios(OR) were calculated with MRP-8/14 yielding 2.13(1.16-3.92; p = .015) to predict T2MI and 0.47(0.26-0.87; p = .015) for T1MI. As expected, hs-cTnI yielded an OR of to predict T2MI 0.34(0.17-0.65; p = .001) and 2.98(1.53-5.81; p = .001) for T1MI. Both markers show comparable and independent results if adjust to hs-cTnI/MRP-8/14, TIMI risk score and CRP. T2MI is associated with higher MRP-8/14 and lower hs-cTnI concentrations than T1MI. Our data suggest that MRP-8/14 as a marker of inflammation might provide usable discriminatory information complementing hs-cTnI in a diagnostic procedure evaluating the type of MI directly upon hospital admission.

Identifying Heart Failure in ECG Data With Artificial Intelligence-A Meta-Analysis.

Introduction: Electrocardiography (ECG) is a quick and easily accessible method for diagnosis and screening of cardiovascular diseases including heart failure (HF). Artificial intelligence (AI) can be used for semi-automated ECG analysis. The aim of this evaluation was to provide an overview of AI use in HF detection from ECG signals and to perform a meta-analysis of available studies. Methods and Results: An independent comprehensive search of the PubMed and Google Scholar database was conducted for articles dealing with the ability of AI to predict HF based on ECG signals. Only original articles published in peer-reviewed journals were considered. A total of five reports including 57,027 patients and 579,134 ECG datasets were identified including two sets of patient-level data and three with ECG-based datasets. The AI-processed ECG data yielded areas under the receiver operator characteristics curves between 0.92 and 0.99 to identify HF with higher values in ECG-based datasets. Applying a random-effects model, an sROC of 0.987 was calculated. Using the contingency tables led to diagnostic odds ratios ranging from 3.44 [95% confidence interval (CI) = 3.12-3.76] to 13.61 (95% CI = 13.14-14.08) also with lower values in patient-level datasets. The meta-analysis diagnostic odds ratio was 7.59 (95% CI = 5.85-9.34). Conclusions: The present meta-analysis confirms the ability of AI to predict HF from standard 12-lead ECG signals underlining the potential of such an approach. The observed overestimation of the diagnostic ability in artificial ECG databases compared to patient-level data stipulate the need for robust prospective studies.