Electrocardiogram-based mortality prediction in patients with COVID-19 using machine learning.

BACKGROUND AND PURPOSE: The electrocardiogram (ECG) is frequently obtained in the work-up of COVID-19 patients. So far, no study has evaluated whether ECG-based machine learning models have added value to predict in-hospital mortality specifically in COVID-19 patients. METHODS: Using data from the CAPACITY-COVID registry, we studied 882 patients admitted with COVID-19 across seven hospitals in the Netherlands. Raw format 12-lead ECGs recorded within 72 h of admission were studied. With data from five hospitals (n = 634), three models were developed: (a) a logistic regression baseline model using age and sex, (b) a least absolute shrinkage and selection operator (LASSO) model using age, sex and human annotated ECG features, and (c) a pre-trained deep neural network (DNN) using age, sex and the raw ECG waveforms. Data from two hospitals (n = 248) was used for external validation. RESULTS: Performances for models a, b and c were comparable with an area under the receiver operating curve of 0.73 (95% confidence interval [CI] 0.65-0.79), 0.76 (95% CI 0.68-0.82) and 0.77 (95% CI 0.70-0.83) respectively. Predictors of mortality in the LASSO model were age, low QRS voltage, ST depression, premature atrial complexes, sex, increased ventricular rate, and right bundle branch block. CONCLUSION: This study shows that the ECG-based prediction models could be helpful for the initial risk stratification of patients diagnosed with COVID-19, and that several ECG abnormalities are associated with in-hospital all-cause mortality of COVID-19 patients. Moreover, this proof-of-principle study shows that the use of pre-trained DNNs for ECG analysis does not underperform compared with time-consuming manual annotation of ECG features.

The Netherlands Sports Cardiology Map: a step towards sports cardiology network medicine for patient and athlete care.

Sports cardiology is a rapidly evolving subspecialty of cardiology, with a growing demand for expertise. To improve patient care, clinicians, patients, and athletes (recreational to elite) should be able to easily identify specialised care pathways, expertise centres and clinicians with sports cardiology expertise. To this purpose, several international societies and organisations recommend establishing a local and national sports cardiology infrastructure. We therefore aimed to establish The Netherlands Sports Cardiology Map. We conducted a web-based survey, which was published on the Netherlands Society of Cardiology home page (2019-2020) and in which each cardiology department or clinic was asked to provide information on sports cardiology expertise and the current infrastructure. Of the 46 respondent centres, 28 (61%) reported that they had expertise in sports cardiology, of which 22 (79%) had specific expertise in one or more specific types of sports. Integrated multidisciplinary meetings were reported by 43% of the centres (n = 12/28). Only two centres reported ongoing research projects that had been approved by an institutional review board. The Netherlands Sports Cardiology Map is an important step towards improving the existing infrastructure and developing network medicine for sports cardiology.

Exercise related ventricular arrhythmias are related to cardiac fibrosis in hypertrophic cardiomyopathy mutation carriers.

AIMS: Hypertrophic cardiomyopathy (HCM) is a frequent cause of sudden cardiac death (SCD) due to exercise-related ventricular arrhythmias (ERVA); however the pathological substrate is uncertain. The aim was to determine the prevalence of ERVA and their relation with fibrosis as determined by cardiac magnetic resonance imaging (CMR) in carriers of an HCM causing mutation. METHODS: We studied the prevalence and origin of ERVA and related these with fibrosis on CMR in a population of 31 HCM mutation carriers. RESULTS: ERVA occurred in seven patients (23%) who all showed evidence of fibrosis (100% ERVA(+) vs. 58% ERVA(-), p = 0.04). No ventricular tachycardia or ventricular fibrillation occurred. In patients with ERVA, the extent of fibrosis was significantly larger (8 ± 4% vs. 3 ± 4%, p = 0.02). ERVA originated from areas with a high extent of fibrosis or regions directly adjacent to these areas. CONCLUSIONS: ERVA in HCM mutation carriers arose from the area of fibrosis detected by CMR; ERVA seems closely related to cardiac fibrosis. Fibrosis as detected by CMR should be evaluated as an additional risk factor to further delineate risk of SCD in carriers of an HCM causing mutation.

Heterozygous loss of Rbm24 in the adult mouse heart increases sarcomere slack length but does not affect function.

RNA-binding proteins are key regulators of post-transcriptional processes such as alternative splicing and mRNA stabilization. Rbm24 acts as a regulator of alternative splicing in heart and skeletal muscle, and is essential for sarcomere assembly. Homozygous inactivation of Rbm24 in mice disrupts cardiac development and results in embryonic lethality around E12.5. In the present study, we generated somatic Rbm24 knockout (KO) mice and investigated the effects of reduced levels of Rbm24 in the adult heart. Due to the embryonic lethality of Rbm24 KO mice, we examined cardiac structure and function in adult Rbm24 heterozygotes (HETs). Rbm24 protein expression was 40% downregulated in HET hearts compared to WT hearts. Force measurements on isolated membrane-permeabilized myocytes showed increased sarcomere slack length and lower myofilament passive stiffness in adult Rbm24 HET compared to wildtype cardiomyocytes. As a result of the differences in sarcomere slack length, the relations between force development and sarcomere length differed between WT and Rbm24 HET hearts. No differences in sarcomere structure and titin isoform composition were observed. Likewise, in vivo cardiac function and myocardial structure was unaltered in Rbm24 HET mice compared to WT, at baseline and upon pressure overload after transverse aortic constriction. In conclusion, we generated a somatic Rbm24 KO model and recapitulated the previously reported embryonic phenotype. In adult Rbm24 HET cardiomyocytes we observed increased sarcomere slack length, but no difference in sarcomere structure and cardiac function.

Resistance to rice yellow mottle virus (RYMV) in cultivated African rice varieties containing RYMV transgenes.

The disease caused by rice yellow mottle virus (RYMV) is a serious problem for African rice growers in large-scale irrigated programs. As there are very few suitable natural sources of RYMV resistance, we have investigated a transgenic approach using widely grown, RYMV-susceptible cultivars of rice and a transgene encoding the RNA-dependent RNA polymerase of RYMV. Transformed lines were resistant to RYMV strains from different African locations. In the most extreme examples there was complete suppression of virus multiplication. Resistance was stable over at least three generations. Subject to satisfactory field testing, these transgenic lines may be suitable for introduction into RYMV-affected rice-growing areas. In the most resistant line, transcription analysis indicated that the resistance derives from an RNA-based mechanism associated with posttranscriptional gene silencing.

GENCOR: a national registry for patients and families suffering from a familial heart disease in the Netherlands.

INTRODUCTION: Developments in DNA-diagnostic techniques allow us to identify a significant proportion of patients with gene mutations causing familial heart diseass (arrhythmia syndromes, cardiomyopathies etc.) and to identify family members in early stages of the disease and/or even before symptoms occur. Early treatment can prevent sudden cardiac death and disease progression. However, data on long-term outcome in unselected genotyped patients are scarce due to a lack of large registries. In 2005, a national internet-based registry for familial heart diseases in the Netherlands, named GENCOR, was developed in collaboration with the Interuniversity Cardiology Institute of the Netherlands (ICIN). OBJECTIVES: GENCOR aims to assess the prevalence of familial heart diseases in patients and families in the Netherlands and to facilitate research to improve the quality of diagnostics and therapy in familial heart diseases. METHODS: Patients who visit the (cardio)genetic outpatient clinic are informed about GENCOR and asked to consent to the storage of information about cardiac examinations, family history and DNA diagnostics from all their visits. Patient data are entered into the internet-based GENCOR database by the cardiologist or clinical geneticist in attendance. Additional information can be stored for scientific research. RESULTS: Four university hospitals are actively obtaining informed consent from the patients, which resulted in the inclusion of more than 300 patients. In 2006, more university hospitals will start using GENCOR and the aim is that all university hospitals will participate. Three research projects have already started using GENCOR. CONCLUSION: GENCOR is already a success, regarding the number of included patients and the related research projects set up within a limited period of time. GENCOR provides easy internet-based access for authorised cardiologists, clinical geneticists and scientists throughout the country.

Cardiac endothelin system impairs left ventricular function in renin-dependent hypertension via decreased sarcoplasmic reticulum Ca(2+) uptake.

BACKGROUND: We evaluated the role of the cardiac endothelin (ET) system in compensated hypertensive left ventricular (LV) hypertrophy (LVH) and after the transition toward LV dysfunction. METHODS AND RESULTS: Hypertensive transgenic rats overexpressing the Ren2 gene (Ren2 rats) were investigated between the ages of 10 and 30 weeks (Ren2-10 and Ren2-30 groups, respectively) and compared with age-matched normotensive Sprague-Dawley (SD) rats (SD-10 and SD-30 groups, respectively). Systolic blood pressure and LV weight were elevated in both Ren2 groups compared with their age-matched SD control groups (P:<0.0001). In Ren2-30 rats, LV end-diastolic pressure increased and -dP/dt(max) decreased compared with the values in SD-30 and Ren2-10 rats (P:<0.05). This was paralleled by an activation of LV mRNA expression of preproET-1 and ET-converting enzyme-1 and ET subtype A (ETA) receptor binding in Ren2-30 compared with Ren2-10 rats (P:<0.001). Cardiac fibrosis was increased and sarcoplasmic reticulum (SR) Ca(2+) reuptake was reduced in Ren2-30 compared with SD-30 and Ren2-10 rats (P:<0.05). Treatment of Ren2 rats with the selective ETA receptor antagonist Lu135252 between 10 and 30 weeks of age did not lower systolic blood pressure, heart weight, or cardiac fibrosis but completely prevented the deterioration of LV end-diastolic pressure and abolished alterations in -dP/dt(max) and SR Ca(2+) reuptake compared with no treatment in Ren2-30 and SD-30 rats (P:<0.05). CONCLUSIONS: Activation of the cardiac ET system accounts at least in part for the LV dysfunction that gradually develops in LVH. The protective effect of ETA antagonism can be attributed to the improvement of diastolic LV function that is due to normalization of impaired SR Ca(2+) uptake.

Deletion-type allele of the angiotensin-converting enzyme gene is associated with progressive ventricular dilation after anterior myocardial infarction. Captopril and Thrombolysis Study Investigators.

OBJECTIVES: This study sought to determine whether patients who are homozygous for the deletion (D)-type allele of the angiotensin-converting enzyme gene display augmented ventricular dilation after myocardial infarction. BACKGROUND: Recent evidence suggests that the deletion-type allele of the angiotensin-converting enzyme gene (DD genotype) is associated with an increased prevalence of myocardial infarction and myocardial hypertrophy. However, it is unknown whether the DD genotype is associated with adverse cardiac remodeling. To address this question we determined the genotype in patients enrolled in the Captopril and Thrombolysis Study (CATS), a prospective trial in which patients received either captopril or placebo during and after thrombolysis for a first anterior myocardial infarction. METHODS: Cardiac volume was determined by echocardiography immediately after thrombolysis and at 1-year follow-up. The genotype for the angiotensin-converting enzyme was determined in 96 patients. Norepinephrine levels were assessed during and immediately after thrombolysis. RESULTS: Immediately after thrombolysis, cardiac volume did not differ between genotype groups. However, at 1-year follow-up, both end-systolic and end-diastolic left ventricular volumes were significantly greater in the DD-genotype group. Norepinephrine increased to higher levels in the DD-genotype group that received placebo therapy. Captopril treatment effectively blunted both the norepinephrine increase and cardiac dilation in the DD-genotype group. CONCLUSIONS: This exploratory study suggests that homozygosity for the angiotensin-converting enzyme deletion-type allele is associated with augmented neurohumoral activation as well as augmented cardiac dilation after an acute anterior myocardial infarction, an effect that may be susceptible to angiotensin-converting enzyme inhibition.

High- versus low-dose ACE inhibition in chronic heart failure: a double-blind, placebo-controlled study of imidapril.

OBJECTIVES: To determine dose-related clinical and neurohumoral effects of angiotensin-converting enzyme (ACE) inhibitors in patients with chronic heart failure (CHF), we conducted a double-blind, placebo-controlled, randomized study of three doses (2.5 mg, 5 mg and 10 mg) of the long-acting ACE inhibitor imidapril. BACKGROUND: The ACE inhibitors have become a cornerstone in the treatment of CHF, but whether high doses are more effective than low doses has not been fully elucidated, nor have the mechanisms involved in such a dose-related effect. METHODS: In a parallel group comparison, the effects of three doses of imidapril were examined. We studied 244 patients with mild to moderate CHF (New York Heart Association class II-III: +/-80%/20%), who were stable on digoxin and diuretics. Patients were treated for 12 weeks, and the main end points were exercise capacity and plasma neurohormones. RESULTS: At baseline, the four treatment groups were well-matched for demographic variables. Of the 244 patients, 25 dropped out: 3 patients died, and 9 developed progressive CHF (3/182 patients on imidapril vs. 6/62 patients on placebo, p < 0.05). Exercise time increased 45 s in the 10-mg group (p = 0.02 vs. placebo), but it did not significantly change in the 5-mg (+16 s), and 2.5-mg (+11 s) imidapril group, compared to placebo (+3 s). Physical working capacity also increased in a dose-related manner. Plasma brain and atrial natriuretic peptide decreased (p < 0.05 for linear trend), while (nor)epinephrine, aldosterone and endothelin were not significantly affected. Renin increased in a dose-related manner, but plasma ACE activity was suppressed similarly (+/-60%) on all three doses. CONCLUSIONS: Already within 3 months after treatment initiation, high-dose ACE inhibition (with imidapril) is superior to low-dose. This is reflected by a more pronounced effect on exercise capacity and some of the neurohormones, but it does not appear to be related to the extent of suppression of plasma ACE.

Cardiovascular end-organ damage in Ren-2 transgenic rats compared to spontaneously hypertensive rats.

To compare hypertensive end-organ damage in two genetic forms of hypertension we assessed cardiovascular function in two rat strains of genetic hypertension: transgenic rats overexpressing the mouse Ren-2 gene [(TGR(mREN2)27]) and blood pressure matched spontaneously hypertensive rats (SHR). Despite similarly elevated blood pressure, systolic dp/dt (mmHg/s) was more impaired in transgenic rats (3099 +/- 446) than in SHR (3571 +/- 272) and normals (4342 +/- 119; P < 0.05). Left ventricular weight (mg/g body weight) increased more in the transgenic rats (40 +/- 3) than in SHR (31 +/- 2) and normals (26 +/- 2). Endothelium-dependent relaxation was significantly decreased only in the transgenic rats. This study shows significantly more cardiac and endothelial dysfunction in transgenic, hypertensive TGR (mREN2)27 than in age and blood pressure matched SHR. This supports the hypothesis that chronic activation of the renin-angiotensin system significantly contributes to hypertensive end-organ damage.

Overexpression of the human angiotensin II type 1 receptor in the rat heart augments load induced cardiac hypertrophy.

Angiotensin II is known to stimulate cardiac hypertrophy and contractility. Most angiotensin II effects are mediated via membrane bound AT1 receptors. However, the role of myocardial AT1 receptors in cardiac hypertrophy and contractility is still rarely defined. To address the hypothesis that increased myocardial AT1 receptor density causes cardiac hypertrophy apart from high blood pressure we developed a transgenic rat model which expresses the human AT1 receptor under the control of the alpha-myosin heavy-chain promoter specifically in the myocardium. Expression was identified and quantified by northern blot analysis and radioligand binding assays, demonstrating overexpression of angiotensin II receptors in the transgenic rats up to 46 times the amount seen in nontransgenic rats. Coupling of the human AT1 receptor to rat G proteins and signal transduction cascade was verified by sensitivity to GTP-gamma-S and increased sensitivity of intracellular Ca2+ [Ca2+]i to angiotensin II in fluo-3 loaded transgenic cardiomyocytes. Transgenic rats exhibited normal cardiac growth and function under baseline conditions. Pronounced hypertrophic growth and contractile responses to angiotensin II, however, were noted in transgenic rats challenged by volume and pressure overload. In summary, we generated a new transgenic rat model that exhibits an upregulated myocardial AT1 receptor density and demonstrates augmented cardiac hypertrophy and contractile response to angiotensin II after volume and pressure overload, but not under baseline conditions.

Lessons from rat models of hypertension: from Goldblatt to genetic engineering.

Over the past 50 years various animal models of hypertension have been developed, predominantly in the rat. In this review we discuss the use of the rat as a model of hypertension, and evaluate what these models have taught us. Interestingly, the spontaneously hypertensive rat (SHR) is by far the most widely used rat model, although it reflects only a rare subtype of human hypertension, i.e. primary hypertension that is inherited in a Mendelian fashion. Many other aspects of the etiology of hypertension are found in other rat models, but these models are less frequently employed. The widespread use of the SHR suggests that this rat model is often chosen without considering alternative (and possibly better suited) models. To illustrate the importance of the choice for a particular model, we compared the natural history and response to antihypertensive drugs in different rat models of hypertension (SHR, Dahl, deoxycorticosterone acetate (DOCA)-salt, two-kidney one-clip, transgenic TGR(mRen2)27. This revealed that the outcome of hypertension can be similar in some respects, as all models exhibit cardiac hypertrophy, and all demonstrate impaired endothelium-dependent relaxations. However, the more severe forms of end-organ damage such as heart failure, stroke and kidney failure, occur only in some models and then only in a subset of the hypertensive rats. The effects of antihypertensives varies even more in the different models: antihypertensive treatment only attenuates end-organ damage if it decreases blood pressure. Moreover, if a given antihypertensive is effective, it sometimes even attenuates end-organ damage in nonhypotensive doses. On the other hand, some agents do decrease blood pressure but do not prevent end-organ damage (e.g. hydralazine in SHR). Furthermore, not all classes of antihypertensives are equally effective in all rat models of hypertension: endothelin-receptor antagonists are not effective in SHR, but have beneficial effects in the DOCA-salt model. The comparison of models, and the comparison of treatment effects suggests that end-organ damage critically depends upon not only on the stress imposed by high blood pressure and its underlying biochemical disturbance, but also upon the ability of the organism to recruit adequate 'coping' mechanisms. These coping mechanisms deserve greater attention, as failure to recruit such mechanisms may indicate an increased risk. The current development of transgenic techniques will provide new opportunities, to develop specific models to address this balance between stress and coping.

Converting enzyme inhibition after experimental myocardial infarction in rats: comparative study between spirapril and zofenopril.

STUDY OBJECTIVE: The aim was to compare the effects of two novel angiotensin converting enzyme (ACE) inhibitors, spirapril and zofenopril, on cardiac remodelling in rats with congestive heart failure after myocardial infarction. Spirapril contains no sulphydryl group, whereas zofenopril is a sulphydryl containing ACE inhibitor. DESIGN: Experimental myocardial infarction was induced by ligation of the left coronary artery. Sham operated animals served as controls. Treatment with spirapril (2-2.5 mg.kg-1.d-1) or zofenopril (12-15 mg.kg-1.d-1) added to the drinking water was started immediately after myocardial infarction or sham operation and continued for six weeks. After the treatment period, all rats were killed. The heart was rapidly removed and perfused as described by Langendorff. Heart rate and left ventricular pressure were measured both at baseline and during stimulation with isoprenaline (6 nM). Heart and lung weights were determined. SUBJECTS: Normotensive male Wistar rats (220-240 g) were used. MEASUREMENTS AND MAIN RESULTS: Experimental myocardial infarction considerably increased left ventricular cavity volume. Chronic treatment with either spirapril or zofenopril significantly attenuated this increase in volume. In infarcted rats, the increase in total heart and lung weight was also significantly reduced by chronic treatment with spirapril and zofenopril, indicating that these compounds reduce cardiac mass and pulmonary congestion in congestive heart failure due to myocardial infarction. There were no significant differences between treatment with spirapril and zofenopril. In the isolated and perfused rat heart, myocardial infarction significantly decreased both heart rate and left ventricular pressure. Converting enzyme inhibition only affected heart rate. Heart rate was significantly higher in infarcted animals treated with spirapril and zofenopril than in untreated infarcted animals. CONCLUSIONS: Both spirapril and zofenopril attenuated ventricular enlargement and cardiac hypertrophy in rats with congestive heart failure after myocardial infarction when treatment was started in the acute phase of myocardial infarction. No additional role could be attributed to the sulphydryl moiety of zofenopril. It is also suggested that these two ACE inhibitors modify cardiac sympathetic activity in rats with congestive heart failure, but more studies are needed to confirm these findings.

Selective and time related activation of the cardiac renin-angiotensin system after experimental heart failure: relation to ventricular function and morphology.

OBJECTIVE: The cardiac renin-angiotensin system is activated in experimental heart failure, but it is unknown at what stage of heart failure it becomes activated, and whether activation is related to ventricular dysfunction and dilatation. Changes in activity of cardiac, renal, and plasma angiotensin converting enzyme (ACE) were therefore examined at different stages of experimental heart failure, with simultaneous measurements of left ventricular pressure, systolic dP/dt, and inner ventricular radius. METHODS: Heart failure was induced by experimental infarction in 17 normotensive male Wistar rats; 14 rats were sham operated. Rats were killed 3, 5, or 80 d after infarction. In an isolated heart perfusion, left ventricular pressure and systolic dP/dT were measured. ACE activity was determined in samples of the left and right cardiac ventricle, kidney, and plasma. Radius of the ventricular cavity was planimetrically determined in transverse sections of the left ventricle. RESULTS: At the different stages both left ventricular pressure and systolic dP/dT progressively decreased and inner radius of the left ventricle increased in all heart failure groups. ACE activity in the left ventricle increased significantly in all heart failure groups and correlated inversely with left ventricular pressure (R = -0.81; p < 0.001) and dP/dt (R = -0.85; p < 0.001). ACE activity in the kidney was only increased 80 d after the induction of heart failure [17(SEM 1) v 11.2(0.5) nM His-Leu generated per min.mg-1, p < 0.01], while plasma ACE activity was not increased in any heart failure group. CONCLUSIONS: Cardiac ACE is activated in the early stage after induction of heart failure and is related to the amount of dysfunction. ACE in the kidney is activated only in the chronic stage. The cardiac renin-angiotensin system therefore already appears to be an important neurohumoral adjustment in the early stage of heart failure and is thereby a suitable target for early intervention by ACE inhibitors.