Interpretation of EKG with Image Recognition and Convolutional Neural Networks.

Electrocardiograms (EKG) form the backbone of all cardiovascular diagnosis, treatment and follow up. Given the pivotal role it plays in modern medicine, there have been multiple efforts to computerize the EKG interpretation with algorithms to improve efficiency and accuracy. Unfortunately, many of these algorithms are machine specific and run-on proprietary signals generated by that machine, hence not generalizable. We propose the development of an image recognition model which can be used to read standard EKG strips. A convolutional neural network (CNN) was trained to classify 12-lead EKGs between 7 clinically important diagnostic classes. An austere variation of the MobileNetV3 model was trained from the ground up on publicly available labeled training set. The precision per class varies from 52% to 91%. This is a novel approach to EKG interpretation as an image recognition problem.

Safety of left atrial appendage closure in heart failure patients.

INTRODUCTION: Left atrial appendage closure (LAAC) is an intervention aimed at stroke prevention in nonvalvular atrial fibrillation (AF). There is a three-fold increased risk of stroke in patients with concomitant presence of AF and heart failure (HF). While anticoagulation is effective, only 60% receive it. We aimed at studying the safety of LAAC in HF patients using a national all-payer database. METHODS: We queried the National Inpatient Sample for the year 2016-2018 for WATCHMAN device insertion using ICD 10 procedure codes. We divided the study population into HF and non-HF groups. Outcomes were compared using appropriate statistical tests, p< .05 was considered significant. RESULTS: 34 385 LAAC procedures were identified of which 8530 (24.8%) were done in patients with HF. The mean (SD) age of the study population was 76 (7.9) years and 42% were female. There was no difference in mean age between HF and non-HF groups. Our findings indicate that there is no difference in inpatient mortality and cardiac complications between the HF and non-HF groups. However, noncardiac complications including acute kidney injury and respiratory failure were higher in the HF group. CONCLUSION: LAAC appears to be a safe procedure in patients with HF. The study is limited by a short follow up period and long-term follow-up is required before definitive conclusions can be made.

Reversible Mechanical Atrioventricular Block Caused By A Steerable Introducer Sheath During Transseptal Catheterization.

A 62-year-old woman presents for pulmonary vein isolation (PVI) for paroxysmal atrial fibrillation. During transseptal catheterization (TSC) the patient sustained mechanical injury to the atrioventricular node (AVN) with consequent complete heart block (CHB). Injury to the AVN and CHB recovered after approximately forty minutes. The patient subsequently underwent a successful PVI with the remainder of the hospital stay uneventful. We present a case of reversible injury to the AVN caused by a steerable introducer sheath during TSC and discuss the mechanisms of injury as well as potential measures to avoid such a complication in the future.

Cybersecurity: The need for data and patient safety with cardiac implantable electronic devices.

Remote monitoring of cardiac implantable electronic devices (CIEDs) has become routine practice as a result of the advances in biomedical engineering, the advent of interconnectivity between the devices through the Internet, and the demonstrated improvement in patient outcomes, survival, and hospitalizations. However, this increased dependency on the Internet of Things comes with risks in the form of cybersecurity lapses and possible attacks. Although no cyberattack leading to patient harm has been reported to date, the threat is real and has been demonstrated in research laboratory scenarios and echoed in patient concerns. The CIED universe comprises a complex interplay of devices, connectivity protocols, and sensitive information flow between the devices and the central cloud server. Various manufacturers use proprietary software and black-box connectivity protocols that are susceptible to hacking. Here we discuss the fundamentals of the CIED ecosystem, the potential security vulnerabilities, a historical overview of such vulnerabilities reported in the literature, and recommendations for improving the security of the CIED ecosystem and patient safety.

Shortness of breath and palpitations in an elderly man: Bad device behavior or malfunction?

70-year-old male with sinus node dysfunction (SND) and paroxysmal atrial fibrillation presents with shortness of breath and palpitations. Presenting EKG shows AF with rapid ventricular rates requiring direct current cardioversion (DCCV). Post-DCCV EKG shows sinus rhythm with competing ventricular pacing. Device interrogation demonstrates the patient's generator at the elective replacement indicator (ERI) and has been forced to VVI 65 bpm causing dyssynchronous ventricular pacing and inducing AF. This case highlights the importance of close device follow up with timely PPM generator change prior to ERI, especially in patients with Medtronic Adapta devices, to avoid unnecessary dyssynchronous ventricular pacing. In addition, device manufacturers should focus on maintaining AV synchrony in pacemakers when they reach ERI.

In memory of tachycardia: A wide complex tachycardia in a young male.

We present a case of a 24-year-old male with palpitations and a wide complex tachycardia. Baseline electrocardiogram (ECG) after termination of tachycardia demonstrates a normal rhythm but with inferior/anterolateral T-wave inversions (TWIs). Electrophysiologic study confirmed the diagnosis of posterior fascicular ventricular tachycardia successfully terminated by anatomic ablation of the left posterior fascicle. TWIs on the patient's baseline ECG were consistent with cardiac memory.

Brugada Phenocopy Induced by Recreational Drug Use.

Recreational drugs are commonly abused in all age groups. Intoxication with these substances can induce silent but significant electrocardiographic signs which may lead to sudden death. In this case study, we present a 49-year-old male with no medical comorbidities who came to the emergency department requesting opioid detoxification. Toxicology screen was positive for cocaine, heroin, and cannabis. Initial electrocardiogram (EKG) showed features of a Brugada pattern in the right precordial leads, which resolved within one day into admission. This presentation is consistent with the recently recognized clinical entity known as Brugada phenocopy.

Inappropriate shock in a subcutaneous cardiac defibrillator due to residual air.

Inappropriate shock due to residual air in subcutaneous implantable cardiac defibrillators is not a well-known complication. Obtaining overpenetrated X-rays, recognizing electrocardiogram findings, limiting blunt finger dissection, and switching to sense at another vector are techniques which might lead to avoidance of unnecessary wound exploration or device removal.

Feasibility of early discharge after implantable cardioverter-defibrillator procedures.

INTRODUCTION: Registry data demonstrate considerably low complication rates after implantable cardioverter-defibrillator (ICD) procedures for primary prevention of sudden death. Yet standard of care includes postimplant overnight in-hospital observation that may levy substantial unnecessary financial burden on health care systems. In appropriate patients, discharge soon after implant could translate into significant cost savings, if such practice does not result in complications. We applied a simple clinical algorithm to assess feasibility of discharge on the same day of ICD implantation in patients at low risk for procedural complications. METHODS: We prospectively randomized primary prevention ICD candidates at low risk for complications (not pacing-dependent or requiring bridging heparin anticoagulation) to next-day discharge with overnight in-hospital observation, or same-day discharge with remote monitoring for 24 hours after ICD implant. Implants were performed via cephalic vein access, and randomization occurred after 4-hours clinical observation and device interrogation. All patients were followed for a minimum of 6 weeks to assess acute procedural complications. RESULTS: 71 patients comprised the study cohort (mean age 62, 79% male) after 3 were excluded. The most common indication for ICD implant was ischemic cardiomyopathy with ejection fraction ≤35%. Device data obtained through 24-hour remote monitoring was comparable to 4-hour postimplant parameters in same-day discharge patients. No acute complications occurred in same-day discharge patients; 1 next-day discharge patient developed pneumothorax. CONCLUSION: ICD implantation with same-day discharge is reasonable in patients at low risk for complications. Remote monitoring can be useful in indicating lead-parameter stability during the immediate postoperative period.

The use of a modified sphygmomanometer to detect atrial fibrillation in outpatients.

This study was designed to assess the accuracy of a modified sphygmomanometer, that measures pulse irregularity, to detect atrial fibrillation (AF). An irregularity index, defined as the standard deviation of the time intervals between beats divided by the mean of the time intervals, was used to analyze standard 12-lead ECGs from hospitalized patients. A threshold irregularity index was selected such that all ECGs with AF exceeded this irregularity index value. A modified automatic blood pressure monitor was designed to detect AF by calculating the irregularity index of the pulse. The device was used to calculate the irregularity index in an unselected group of outpatients during scheduled office visits in which a standard 12-lead ECG was performed. A total of 125 ECGs, 53 with AF, were analyzed. Using a threshold irregularity index of 0.066, the sensitivity for detecting AF was 100%, the specificity was 92% and diagnostic accuracy 95%. A modified sphygmomanometer was used to analyze the pulse of 450 outpatients, 54 of whom were documented by ECG to be in AF. When paired readings were analyzed, the rhythm was considered to be irregular if both readings were greater than the threshold index. Using a threshold index of 0.06, all the AF patients were correctly identified while 37 non-AF patients also exceeded the threshold irregularity index. In this analysis, the sensitivity was 100%, the specificity 91%, and the diagnostic accuracy 92% for detecting AF. The irregularity index determined using a modified sphygmomanometer can accurately identify AF.

Determinants of left atrial size in patients with newly diagnosed untreated hypertension.

The present study determined the relationships between ambulatory blood pressure, left ventricular mass, body mass index, and other clinical and demographic variables to left atrial size in previously untreated hypertensive and normotensive subjects. Left atrial size was measured uni-dimensionally using M-mode echocardiography in 58 newly diagnosed never-treated hypertensive patients (office blood pressure 149/96 +/- 15/7 mmHg) and 28 normotensive control subjects (office blood pressure, 122/78 +/- 8/8 mmHg). Left ventricular mass, septal and posterior wall thickness were significantly increased in hypertensive compared to normotensive subjects (230 +/- 63 g versus 181 +/- 45 g, 1.1 +/- 0.2 cm versus 0.94 +/- 0.2 cm, and 1.04 +/- 0.2 cm versus 0.92 +/- 0.2 cm respectively; all p < 0.001). Left ventricular internal diameter (4.9 +/- 0.6 versus 4.8 +/- 0.4 cm, = 0.54) and left atrial size (3.74 +/- 0.48 versus 3.70 +/- 0.34 cm, p = 0.86) were not different between the two groups respectively. Body mass index, weight, left ventricular mass, wall thickness, and 24-h pulse pressure were significant correlates of left atrial size in the entire group and in the hypertensive subgroup. In the normotensive subgroup, body weight, body mass index, 24-h systolic and pulse pressure, and left ventricular mass were significant correlates. Multiple regression analyses in the entire group and the hypertensive subgroup alone showed that body mass index and left ventricular mass were the two best predictors of left atrial dimension. These data demonstrate that body mass index and left ventricular mass were the main correlates of left atrial size in patients with previously untreated stage I-II hypertension.