Computerized Analysis of the Ventricular Fibrillation Waveform Allows Identification of Myocardial Infarction: A Proof-of-Concept Study for Smart Defibrillator Applications in Cardiac Arrest

Role of artificial intelligence in defibrillators: a narrative review

Automated external defibrillators (AEDs) and implantable cardioverter defibrillators (ICDs) are used to treat life-threatening arrhythmias. AEDs and ICDs use shock advice algorithms to classify ECG tracings as shockable or non-shockable rhythms in clinical practice. Machine learning algorithms have recently been assessed for shock decision classification with increasing accuracy. Outside of rhythm classification alone, they have been evaluated in diagnosis of causes of cardiac arrest, prediction of success of defibrillation and rhythm classification without the need to interrupt cardiopulmonary resuscitation. This review explores the many applications of machine learning in AEDs and ICDs. While these technologies are exciting areas of research, there remain limitations to their widespread use including high processing power, cost and the ‘black-box’ phenomenon.

The effect of the localisation of an underlying ST-elevation myocardial infarction on the VF-waveform: A multi-centre cardiac arrest study

INTRODUCTION

In cardiac arrest, ventricular fibrillation (VF) waveform characteristics such as amplitude spectrum area (AMSA) are studied to identify an underlying myocardial infarction (MI). Observational studies report lower AMSA-values in patients with than without underlying MI. Moreover, experimental studies with 12-lead ECG-recordings show lowest VF-characteristics when the MI-localisation matches the ECG-recording direction. However, out-of-hospital cardiac arrest (OHCA)-studies with defibrillator-derived VF-recordings are lacking.

METHODS

Multi-centre (Amsterdam/Nijmegen, the Netherlands) cohort-study on the association between AMSA, ST-elevation MI (STEMI) and its localisation. AMSA was calculated from defibrillator pad-ECG recordings (proxy for lead II, inferior vantage point); STEMI-localisation was determined using ECG/angiography/autopsy findings.

RESULTS

We studied AMSA-values in 754 OHCA-patients. There were statistically significant differences between no STEMI, anterior STEMI and inferior STEMI (Nijmegen: no STEMI 13.0mVHz [7.9-18.6], anterior STEMI 7.5mVHz [5.6-13.8], inferior STEMI 7.5mVHz [5.4-11.8], p = 0.006. Amsterdam: 11.7mVHz [5.0-21.9], 9.6mVHz [4.6-17.2], and 6.9mVHz [3.2-16.0], respectively, p = 0.001). Univariate analyses showed significantly lower AMSA-values in inferior STEMI vs. no STEMI; there was no significant difference between anterior and no STEMI. After correction for confounders, adjusted absolute AMSA-values were numerically lowest for inferior STEMI in both cohorts, and the relative differences in AMSA between inferior and no STEMI was 1.4-1.7 times larger than between anterior and no STEMI.

CONCLUSION

This multi-centre VF-waveform OHCA-study showed significantly lower AMSA in case of underlying STEMI, with a more pronounced difference for inferior than for anterior STEMI. Confirmative studies on the impact of STEMI-localisation on the VF-waveform are warranted, and might contribute to earlier diagnosis of STEMI during VF.

Towards individualised treatment of out-of-hospital cardiac arrest patients: an update on technical innovations in the prehospital chain of survival

Out-of-hospital cardiac arrest (OHCA) is a major healthcare problem, with approximately 200 weekly cases in the Netherlands. Its critical, time-dependent nature makes it a unique medical situation, of which outcomes strongly rely on infrastructural factors and on-scene care by emergency medical services (EMS). Survival to hospital discharge is poor, although it has substantially improved, to roughly 25% over the last years. Recognised key factors, such as bystander resuscitation and automated external defibrillator use at the scene, have been markedly optimised with the introduction of technological innovations. In an era with ubiquitous smartphone use, the Dutch digital text message alert platform HartslagNu (www.hartslagnu.nl) increasingly contributes to timely care for OHCA victims. Guidelines emphasise the role of cardiac arrest recognition and early high-quality bystander resuscitation, which calls for education and improved registration at HartslagNu. As for EMS care, new technological developments with future potential are the selective use of mechanical chest compression devices and extracorporeal life support. As a future innovation, ‘smart’ defibrillators are under investigation, guiding resuscitative interventions based on ventricular fibrillation waveform characteristics. Taken together, optimisation of available prehospital technologies is crucial to further improve OHCA outcomes, with particular focus on more available trained volunteers in the first phase and additional research on advanced EMS care in the second phase.

Coronary angiography findings in patients with shock-resistant ventricular fibrillation cardiac arrest

INTRODUCTION

Shock-resistant ventricular fibrillation (VF) poses a therapeutic challenge during out-of-hospital cardiac arrest (OHCA). For these patients, new treatment strategies are under active investigation, yet underlying trigger(s) and substrate(s) have been poorly characterised, and evidence on coronary angiography (CAG) data is often limited to studies without a control group.

METHODS

In our OHCA-registry, we studied CAG-findings in OHCA-patients with VF who underwent CAG after hospital arrival. We compared baseline demographics, arrest characteristics, CAG-findings and outcomes between patients with VF that was shock-resistant (defined as >3 shocks) or not shock-resistant (≤3 shocks).

RESULTS

Baseline demographics, arrest location, bystander resuscitation and AED-use did not differ between 105 patients with and 196 patients without shock-resistant VF. Shock-resistant VF-patients required more shocks, with higher proportions endotracheal intubation, mechanical CPR, amiodaron and epinephrine. In both groups, significant coronary artery disease (≥1 stenosis >70%) was highly prevalent (78% vs. 77%, p = 0.76). Acute coronary occlusions (ACOs) were more prevalent in shock-resistant VF-patients (41% vs. 26%, p = 0.006). Chronic total occlusions did not differ between groups (29% vs. 33%, p = 0.47). There was an association between increasing numbers of shocks and a higher likelihood of ACO. Shock-resistant VF-patients had lower proportions 24-h survival (75% vs. 93%, p < 0.001) and survival to discharge (61% vs. 78%, p = 0.002).

CONCLUSION

In this cohort of OHCA-patients with VF and CAG after transport, acute coronary occlusions were more prevalent in patients with shock-resistant VF compared to VF that was not shock-resistant, and their clinical outcome was worse. Confirmative studies are warranted for this potentially reversible therapeutic target.

Electrocardiographic recording direction impacts ventricular fibrillation waveform measurements: A potential pitfall for VF-waveform guided defibrillation protocols

Aim

In cardiac arrest, ventricular fibrillation (VF) waveform analysis has identified the amplitude spectrum area (AMSA) as a key predictor of defibrillation success and favorable neurologic survival. New resuscitation protocols are under investigation, where prompt defibrillation is restricted to cases with a high AMSA. Appreciating the variability of in-field pad placement, we aimed to assess the impact of recording direction on AMSA-values, and the inherent defibrillation advice.

Methods

Prospective VF-waveform study on 12-lead surface electrocardiograms (ECGs) obtained during defibrillation testing in ICD-recipients (2010–2017). AMSA-values (mVHz) of simultaneous VF-recordings were calculated and compared between all limb leads, with lead II as reference (proxy for in-field pad position). AMSA-differences between leads I and II were quantified using Bland-Altman analysis. Moreover, we investigated differences between these adjacent leads regarding classification into high (≥15.5), intermediate (6.5–15.5) or low (≤6.5) AMSA-values.

Results

In this cohort (n = 243), AMSA-values in lead II (10.2 ± 4.8) differed significantly from the other limb leads (I: 8.0 ± 3.4; III: 12.9 ± 5.6, both p < 0.001). The AMSA-value in lead I was, on average, 2.24 ± 4.3 lower than in lead II. Of the subjects with high AMSA-values in lead II, only 15% were classified as high if based on assessments of lead I. For intermediate and low AMSA-values, concordances were 66% and 72% respectively.

Conclusions

ECG-recording direction markedly affects the result of VF-waveform analysis, with 20–30% lower AMSA-values in lead I than in lead II. Our data suggest that electrode positioning may significantly impact shock guidance by ‘smart defibrillators’, especially affecting the advice for prompt defibrillation.

Computerized Analysis of the Ventricular Fibrillation Waveform Allows Identification of Myocardial Infarction: A Proof-of-Concept Study for Smart Defibrillator Applications in Cardiac Arrest

Background

In cardiac arrest, computerized analysis of the ventricular fibrillation (VF) waveform provides prognostic information, while its diagnostic potential is subject of study. Animal studies suggest that VF morphology is affected by prior myocardial infarction (MI), and even more by acute MI. This experimental in‐human study reports on the discriminative value of VF waveform analysis to identify a prior MI. Outcomes may provide support for in‐field studies on acute MI.

Methods and Results

We conducted a prospective registry of implantable cardioverter defibrillator recipients with defibrillation testing (2010–2014). From 12‐lead surface ECG VF recordings, we calculated 10 VF waveform characteristics. First, we studied detection of prior MI with lead II, using one key VF characteristic (amplitude spectrum area [AMSA]). Subsequently, we constructed diagnostic machine learning models: model A, lead II, all VF characteristics; model B, 12‐lead, AMSA only; and model C, 12‐lead, all VF characteristics. Prior MI was present in 58% (119/206) of patients. The approach using the AMSA of lead II demonstrated a C‐statistic of 0.61 (95% CI, 0.54–0.68). Model A performance was not significantly better: 0.66 (95% CI, 0.59–0.73), P=0.09 versus AMSA lead II. Model B yielded a higher C‐statistic: 0.75 (95% CI, 0.68–0.81), P<0.001 versus AMSA lead II. Model C did not improve this further: 0.74 (95% CI, 0.67–0.80), P=0.66 versus model B.

Conclusions

This proof‐of‐concept study provides the first in‐human evidence that MI detection seems feasible using VF waveform analysis. Information from multiple ECG leads rather than from multiple VF characteristics may improve diagnostic accuracy. These results require additional experimental studies and may serve as pilot data for in‐field smart defibrillator studies, to try and identify acute MI in the earliest stages of cardiac arrest.