Search

Abstract Aims In recent years, there has been remarkable development in machine learning (ML) models, showing a trend towards high prediction performance. ML models with high prediction performance often become structurally complex and are frequently perceived as black boxes, hindering intuitive interpretation of the prediction results. We aimed to develop ML models with high prediction performance, interpretability, and superior risk stratification to predict in‐hospital mortality and worsening heart failure (WHF) in patients with acute heart failure (AHF). Methods and results Based on the Kyoto Congestive Heart Failure registry, which enrolled 4056 patients with AHF, we developed prediction models for in‐hospital mortality and WHF using information obtained on the first day of admission (demographics, physical examination, blood test results, etc.). After excluding 16 patients who died on the first or second day of admission, the original dataset (n = 4040) was split 4:1 into training (n = 3232) and test datasets (n = 808). Based on the training dataset, we developed three types of prediction models: (i) the classification and regression trees (CART) model; (ii) the random forest (RF) model; and (iii) the extreme gradient boosting (XGBoost) model. The performance of each model was evaluated using the test dataset, based on metrics including sensitivity, specificity, area under the receiver operating characteristic curve (AUC), Brier score, and calibration slope. For the complex structure of the XGBoost model, we performed SHapley Additive exPlanations (SHAP) analysis, classifying patients into interpretable clusters. In the original dataset, the proportion of females was 44.8% (1809/4040), and the average age was 77.9 ± 12.0. The in‐hospital mortality rate was 6.3% (255/4040) and the WHF rate was 22.3% (900/4040) in the total study population. In the in‐hospital mortality prediction, the AUC for the XGBoost model was 0.816 [95% confidence interval (CI): 0.815–0.818], surpassing the AUC values for the CART model (0.683, 95% CI: 0.680–0.685) and the RF model (0.755, 95% CI: 0.753–0.757). Similarly, in the WHF prediction, the AUC for the XGBoost model was 0.766 (95% CI: 0.765–0.768), outperforming the AUC values for the CART model (0.688, 95% CI: 0.686–0.689) and the RF model (0.713, 95% CI: 0.711–0.714). In the XGBoost model, interpretable clusters were formed, and the rates of in‐hospital mortality and WHF were similar among each cluster in both the training and test datasets. Conclusions The XGBoost models with SHAP analysis provide high prediction performance, interpretability, and reproducible risk stratification for in‐hospital mortality and WHF for patients with AHF.

Abstract Aims Hospitalizations are common in patients with heart failure and are associated with high mortality, readmission and economic burden. Detecting early signs of worsening heart failure may enable earlier intervention and reduce hospitalizations. The HeartLogic algorithm is designed to predict worsening heart failure using diagnostic data from multiple device sensors. The main objective of this analysis was to evaluate the sensitivity of the HeartLogic alert calculation in predicting worsening heart failure events (HFEs). We also evaluated the false positive alert rate (FPR) and compared the incidence of HFEs occurring in a HeartLogic alert state to those occurring out of an alert state. Methods The HINODE study enrolled 144 patients (81 ICD and 63 CRT‐D) with device sensor data transmitted via a remote monitoring system. HeartLogic alerts were then retrospectively simulated using relevant sensor data. Clinicians and patients were blinded to calculated alerts. Reported adverse events with HF symptoms were adjudicated and classified by an independent HFE committee. Sensitivity was defined as the ratio of the number of detected usable HFEs (true positives) to the total number of usable HFEs. A false positive alert was defined as an alert with no usable HFE between the alert onset date and the alert recovery date plus 30 days. The patient follow‐up period was categorized as in alert state or out of alert state. The event rate ratio was the HFE rate calculated in alert to out of alert. Results The patient cohort was 79% male and had an average age of 68 ± 12 years. This analysis yielded 244 years of follow‐up data with 73 HFEs from 37 patients. A total of 311 HeartLogic alerts at the nominal threshold (16) occurred across 106 patients providing an alert rate of 1.27 alerts per patient‐year. The HFE rate was 8.4 times greater while in alert compared with out of alert (1.09 vs. 0.13 events per patient‐year; P

Abstract Aims A considerable proportion of candidates for transcatheter aortic valve implantation (TAVI) have underlying heart failure (HF) with preserved ejection fraction (HFpEF), which can be challenging for diagnosis because significant valvular heart disease should be excluded before diagnosing HFpEF. This study investigated the long‐term prognostic value of the pre‐procedural H2FPEF score in patients with preserved ejection fraction (EF) undergoing TAVI. Methods and results Patients who underwent TAVI between October 2013 and May 2017 were enrolled from the Optimized CathEter vAlvular iNtervention–Transcatheter Aortic Valve Implantation Japanese multicentre registry. After excluding 914 patients, 1674 patients with preserved EF ≥ 50% (median age: 85 years, 72% female) were selected for calculation of the H2FPEF score and were dichotomized into two groups: the low H2FPEF score [0–5 points; n = 1399 (83.6%)] group and the high H2FPEF score [6–9 points; n = 275 (16.4%)] group. Patients with high H2FPEF scores were associated with a higher prevalence of New York Heart Association Functional Class III/IV (59.3% vs. 43.7%, P

Background The impact of high‐risk coronary artery plaques identified using optical coherence tomography on late luminal narrowing and clinical events remains poorly understood. Methods and Results This multicenter prospective study included 176 patients who underwent percutaneous coronary intervention and serial optical coherence tomography at baseline and 1‐year follow‐up to investigate nontarget regions with angiographically intermediate stenosis. At 1 year after percutaneous coronary intervention, the coronary artery lumen area decreased significantly from 6.06 (95% CI, 5.60–6.53) mm2 to 5.88 (95% CI, 5.41–6.35) mm2 (difference, −0.18; 95% CI, −0.22 to −0.14 mm2; P

Abstract Background His bundle pacing (HBP) and left bundle branch area pacing (LBBAP) emerge as better alternatives to right ventricular apical pacing (RVAP) in patients with bradycardia requiring permanent cardiac pacing. We aimed to compare the clinical outcomes of LBBAP, HBP, and RVAP in Japanese patients with bradycardia. Methods A total of 424 patients who underwent successful pacemaker implantation (HBP, n = 53; LBBAP, n = 75; and RVAP, n = 296) were retrospectively enrolled in this study. The primary study endpoint was the cumulative incidence of heart failure hospitalization (HFH) during the follow‐up. Results The success rate for implantation was higher in the LBBAP group than in the HBP group (94.9% and 81.5%, respectively). Capture threshold increase >1V during the follow‐up occurred in the HBP and RVAP groups (9.4% and 5.1%, respectively), while it did not in the LBBAP group. The cumulative incidence of HFH was significantly lower in the LBBAP group than the RVAP (adjusted hazard ratio, 0.12 [95% confidence interval: 0.02–0.86]; p = .034); it did not differ between the HBP and RVAP groups (adjusted hazard ratio, 0.48 [95% confidence interval: 0.17–1.34]; p = .16). Advanced age, mean percent right ventricular pacing (per 10% increase), left ventricular ejection fraction

Background The effects of aspirin‐free strategy on bleeding and cardiovascular events in patients undergoing percutaneous coronary intervention with oral anticoagulation (OAC) have not been fully elucidated. Methods and Results We conducted the prespecified subgroup analysis based on the use of OAC, including vitamin K antagonist and direct oral anticoagulants, within 7 days before percutaneous coronary intervention in the STOPDAPT‐3 (Short and Optimal Duration of Dual Antiplatelet Therapy‐3) trial, which randomly compared prasugrel monotherapy (2984 patients) to dual antiplatelet therapy (DAPT) with prasugrel and aspirin (2982 patients) in patients with acute coronary syndrome or high bleeding risk. The coprimary end points were major bleeding events (Bleeding Academic Research Consortium types 3 or 5) and cardiovascular events (a composite of cardiovascular death, myocardial infarction, definite stent thrombosis, or ischemic stroke) at 1 month. Among 5966 study patients, there were 530 patients (8.9%) with OAC (no aspirin: N=248, and DAPT: N=282) and 5436 patients (91.1%) without OAC (no aspirin: N=2736, and DAPT: N=2700). Regardless of the use of OAC, the effects of no aspirin compared with DAPT were not significant for the bleeding end point (OAC: 4.45% and 4.27%, hazard ratio [HR], 1.04 [95% CI, 0.46–2.35]; no‐OAC: 4.47% and 4.75%, HR, 0.94 [95% CI, 0.73–1.20]; P for interaction=0.82), and for the cardiovascular end point (OAC: 4.84% and 3.20%, HR, 1.53 [95% CI, 0.64–3.62]; no‐OAC: 4.06% and 3.74%, HR, 1.09 [95% CI 0.83–1.42]; P for interaction =0.46). Conclusions The no‐aspirin strategy compared with the DAPT strategy failed to reduce major bleeding events irrespective of the use of OAC. There was a numerical excess risk of the no‐aspirin strategy relative to the DAPT strategy for cardiovascular events in patients with OAC.

Background There is a scarcity of data on the prevalence and clinical impact of cerebrovascular disease detected on preprocedural computed tomography (CT) before aortic valve replacement (AVR) in patients with severe aortic stenosis. Methods and Results Among patients with severe aortic stenosis undergoing AVR, the authors compared clinical outcomes between patients with and without cerebrovascular disease detected on preprocedural CT, which was defined as chronic brain infarction or hemorrhage. The primary outcome measure in this study was a composite of all‐cause death or stroke. Among 567 study patients, 200 patients (35.3%) had cerebrovascular disease on preprocedural CT. Among 200 patients with cerebrovascular disease on preprocedural CT, only 28.5% of patients had a clinical history of symptomatic stroke. The cumulative 3‐year incidence of death or stroke was higher in patients with cerebrovascular disease on preprocedural CT than in those without cerebrovascular disease on preprocedural CT (40.7% versus 24.1%, log‐rank P

Patent foramen ovale (PFO) complicated with carcinoid heart disease (CHD) can cause severe hypoxia and worsening clinical conditions. We report the case of a patient with CHD in poor general condition with multiple severe valve regurgitations and PFO, who underwent successful percutaneous closure of the PFO.

Transcatheter aortic valve replacement in extra-large annulus is challenging because of the limited sizes and data of the current commercial prosthesis sizes available. We present a case of successful transcatheter aortic valve replacement with a new balloon-expandable SAPIEN 3 Ultra RESILIA valve for an aortic annulus area >1,000 mm2 on computed tomography.

BACKGROUND: A high permanent pacemaker implantation (PPI) risk remains a concern of self-expandable transcatheter aortic valve implantation, despite the continued improvements in implantation methodology. We aimed to assess the impact of real-time direct visualization of the membranous septum using transjugular intracardiac echocardiography (ICE) during transcatheter aortic valve implantation on reducing the rates of conduction disturbances including the need for PPI. METHODS: Consecutive patients treated with Evolut R and Evolut PRO/PRO+ from February 2017 to September 2022 were included in this study. We compared outcomes between the conventional implantation method using the 3-cusps view (3 cusps without ICE group), the recent method using cusp-overlap view (cusp overlap without ICE group), and our novel method using ICE (cusp overlap with ICE group). RESULTS: Of the 446 patients eligible for analysis, 211 (47.3%) were categorized as the 3 cusps without ICE group, 129 (28.9%) were in the cusp overlap without ICE group, and 106 (23.8%) comprised the cusp overlap with ICE group. Compared with the 3 cusps without ICE group, the cusp overlap without ICE group had a smaller implantation depth (2.2 [interquartile range, 1.0-3.5] mm versus 4.3 [interquartile range, 3.3-5.4] mm; P<0.001) and lower 30-day PPI rates (7.0% versus 14.2%; P=0.035). Compared with the cusp overlap without ICE group, the cusp overlap with ICE group had lower 30-day PPI rates (0.9%; P=0.014), albeit with comparable implantation depths (1.9 [interquartile range, 0.9-2.9] mm; P=0.150). Multivariable analysis showed that our novel method using ICE with the cusp-overlap view was independently associated with a 30-day PPI rate reduction. There were no group differences in 30-day all-cause mortality (1.4% versus 1.6% versus 0%; P=0.608). CONCLUSIONS: Our novel implantation method using transjugular ICE, which enable real-time direct visualization of the membranous septum, achieved a predictably high position of prostheses, resulting in a substantial reduction of conduction disturbances requiring PPI after transcatheter aortic valve implantation. [ABSTRACT FROM AUTHOR]