Thrombolytic therapy is a widely used therapeutic procedure for acute myocardial infarction (MI). Currently, the decision whether this procedure was successful, which bears important clinical implications, is based on clinical information. Therefore, the main goal of this research project was to develop a set of quantitative tools for diagnosing reperfusion, based on the analysis of the heart rate signal. This research project has three parts: A) Development of reliable and robust tools for time-dependent spectral analysis of HRV, B) Data collection and C) Analysis of the data and the development of a criterion for reperfusion and reocclusion.
In the last year, we focused in the first part of the project. This report briefly describes our achievements.
Vagal activity is typically assessed from the power in the high frequency (HF) region of the heart rate power spectrum. When analyzing the heart rate signal in the time-frequency plane, one may find difficulties in determining the frequency limits of the HF peak. This problem is even more pronounced in patients whose physiologic condition changes over time, as in our case. To solve this problem, we developed an algorithm for automatic and objective detection of the changing frequency limits of the HF peak. This algorithm is based on the analyzing the morphology of the HF peak as found by the wavelet analysis of the heart rate signal. This algorithm was verified in simulations and in data from normal subjects
. Patients during thrombolytic therapy exhibit frequent arrhythmias. These arrhythmias are a major limitation in heart rate variability analysis, and especially in time-dependent analysis. However, isolated arrhythmias cause a brief reduction in cardiac output and may serve as a “stimulus” to the cardiac rate control system. Lately, analysis of the heart rate response to isolated PVCs revealed a relation between this response and sudden death. This method is termed heart rate turbulence. We hypothesized that such an analysis will provide important information regarding cardiac control in MI patients. We developed a modified version of the heart turbulence scheme and applied it the our data set of MI patients undergoing thrombolytic therapy.
Another confounding factor in the analysis of heart rate variability is the occurrence of saturation in the vagal system. Such saturation was found in our data set. This phenomenon hinders the interpretation of heart rate variability patterns and reduces the accuracy of detection of reperfusion from patterns of heart rate oscillations. We collaborate with the research group at the Section of Cardiology in Northwestern University who performed careful experiments to study this phenomenon. We developed a set of algorithms, based on analyzing the temporal relations between heart rate and respiration, to detect saturation of the vagal system. Our algorithms are currently being refined and verified using the data set provided by the Northwestern group.
We plan to increase the number of recordings of MI patients undergoing thrombolytic therapy, as the next step of our project. The above mentioned algorithms will be incorporated into our already developed scheme of detecting reperfusion in order to improve the accuracy of detection.
