New Research on T-Wave Alternans: Interview with Charles Swerdlow, MD, FHRS

Explain T-wave alternans testing. What were your goals in doing this T-wave alternans research? T-wave alternans (TWA) is a measure of cardiac electrical instability that has been used to predict long-term risk of arrhythmia by measuring from the surface electrocardiogram. The purpose of electrocardiographic testing is to determine if patients need implantable cardioverter defibrillators (ICDs). The goal of this particular study was different. ICDs are very effective at stopping arrhythmias after they start; however, the shocks they deliver are painful. Therefore, the long-term goal of our approach is to find a way to predict and prevent arrhythmias in patients with ICDs. We wanted to predict short-term exactly when these patients were going to have ventricular tachycardia (VT) or ventricular fibrillation (VF), and then potentially intervene with a unique pacing algorithm or some other method to alter the electrical properties of the heart for the short term and prevent ventricular arrhythmias as opposed to stopping them after they start. So the goal of this study was to determine whether TWA was different immediately before spontaneous episodes of VT/VF in ICD patients from the amount of TWA that is present at other times of the day. Tell us about your study methods. Describe the simple averaging (AVE) method that was developed for this study. We faced two challenges in this study. One was that the vast majority of TWA data have been reported from surface ECGs, but we wanted to record it from intracardiac electric electrograms. The second is that the TWA data have typically been recorded over a period of minutes — either during atrial pacing or exercise testing — and then sophisticated signal processing algorithms are used to identify very low amplitude TWA from noise. In contrast, the memory in ICDs before VT/VF is only 10–20 beats. So we needed a method that would reliably record TWA from 10–20 beats. We knew from a previous study we did that the magnitude or amplitude of the TWA was a lot greater — about 10 to 100 times greater — when recorded from the inside of the heart than from the surface of the body, and this worked in our favor. It permitted us to measure a beat-to-beat average of TWA, that is, we subtracted the amplitude of the second T wave from that of the first to get a difference value, repeated this process from all pairs of beats, and then averaged these values. In contrast, the most commonly used method from the surface ECG — the fast Fourier transform (FFT) method — requires many more beats than are stored in memory for us to use it. We found that the other surface ECG method, the modified moving average method, had some technical problems when it was applied to signals that came out of ICD electrograms. For that reason we had to develop a new method that would be able to record TWA in a small number of beats. We spent a considerable amount of time validating that method with simulation. What were the findings from this study? The major finding is that the magnitude of TWA immediately before life-threatening ventricular arrhythmias is much greater than it is when recorded at other times in patients with ICDs. Will more research be done? Yes. We are now conducting prospective multicenter studies of a substantially larger number of patients in which we use the same technique. We expect to finish collecting data sometime this summer, and with any luck we will have the data analyzed and at least have an abstract ready for the Heart Rhythm Society meeting next year as well as a paper. The next step will be to actually write code that can be downloaded into existing ICDs, run that software in real time 24/7, and assess the accuracy which with we can predict a life-threatening arrhythmia. The final step is a set of intervention studies in which we would intervene to reduce the frequency of VT/VF by adaptive pacing algorithms implemented when abnormal TWA was detected. The goal would be to identify the most effective way to reduce the frequency of VT/VF in ICD patients. How soon could we see this detection technology incorporated into ICDs? I would estimate the timeframe to market release product would be about five years. For more information, please visit: