Paced ventricular electrogram fractionation and sudden death in hypertrophic cardiomyopathy and other non-coronary heart diseases.

1. Introduction minimum required predictive accuracy of the test will vary amongst physicians. As there is no consensus on the The prediction of sudden arrhythmic death (SCD) in criteria for ICD implantation, we hope to achieve a patients with non-coronary heart disease has become positive predictive accuracy (PPA) of 0.3 so that three out increasingly important because of the need to select high of every ten patients implanted with an ICD would develop risk patients who require prophylactic implantable car-VF. dioverter-defibrillators (ICD). Hypertrophic car-The relatively low SCD rate in HCM causes problems in diomyopathy (HCM) is an important cause of SCD, and designing a trial to establish the SCD rate or the PPA of a while it is clear that ICDs are effective in aborting SCD method of predicting SCD and in the interpretation of due to ventricular fibrillation [1–3], there is no generally published studies. Assume that there are two populations accepted method of risk stratification to allow large-scale, of patients, one with a 4% and the other with 1.5% annual selective ICD implantation in such patients [4]. This article mortality, which are observed for 3 years. Fig. 1 shows the reviews the various trials which have been performed to 95% confidence limits of the observed SCD rate as the predict SCD in patients with HCM. Then an emerging number of patients are increased and shows that there are technique which was initially developed for HCM, paced large potential errors in measuring the SCD rate with small electrogram fractionation analysis (PEFA), is reviewed. numbers of patients. Therefore, the SCD rate in HCM is This technique is based on detection of the substrate for VF difficult to establish, purely on the basis of sample sizes and has promise in identifying high risk patients with and this is complicated by selection bias, differing age HCM and other non-coronary heart diseases which cause structures, different phenotypic expression [5–7] and dif-SCD. Finally, preliminary results obtained in man and ferent genotypes which may carry different risks of SCD animal models are compared suggesting that PEFA may be [8]. The differences in published SCD rates which vary used to interpret electrophysiological data from humans at from 7.5 to 0% per annum in small study groups can be risk of SCD in the context of animal experiments. explained on the basis of sample size [9–14] while the variability of the estimated SCD is lower in studies …