Scope and Method of Study: ECG analysis and other available clinical techniques are unable to predict potentially fatal myocardial ischemic events---they are diagnostic tools, not prognostic tools. Methods derived from the field of nonlinear dynamics have shown tremendous promise as quantitative measures of cardiovascular (CV) system stability during disease. However, they suffer from severe limitations (i.e. nonstationarity and skipped beats). In this study, heart rate variability (HRV) signals were extracted from healthy subjects and patients with myocardial ischemia. Subsequently, these signals were analyzed using recurrence plot (RP) and recurrence quantification analysis (RQA) techniques.

Findings and Conclusions: In this study, new methods capable of removing the limitation imposed by time series nonstationarity were designed. This allowed for the characterization of complex CV system behavior specific to changes that accompany orthostasis, senescence, and development of clinical myocardial ischemia. More specifically, RQA heart rate (HR) complexity measures change in a predictable way following orthostasis. Additionally, the complexity of nonstationary HRV signals, as measured by RQA complexity statistics, decreases with age. Finally, criteria have been developed that are most specific to impending clinical myocardial ischemia. Software technology has been developed that is capable of predicting ~54% of all ischemic events using a HR time series from a simple 5 to 10 minute single-lead ECG signal. On average ischemic events can be predicted ~24 minutes before the onset of clinically diagnosed myocardial ischemia. For major ischemic events lasting longer than 20 minutes, prediction sensitivity and advanced prediction time is improved to ~83% and ~38 minutes, respectively. Overall, this study revealed that HR complexity could be accurately quantified using RQA statistics regardless of the level of nonstationarity. Furthermore, these measures could be used as indexes of subject-specific cardiovascular health.