Dynamic behavior of heart rate in ischemic stroke

Traditional spectral and nonspectral methods have shown that heart rate (HR) variability is reduced after stroke. Some patients with poor outcome, however, show randomlike, complex patterns of HR behavior that traditional analysis techniques are unable to quantify. Therefore, we designed the present study to evaluate the complexity and correlation properties of HR dynamics after stroke by using new analysis methods based on nonlinear dynamics and fractals ("chaos theory").

In addition to the traditional spectral components of HR variability, we measured instantaneous beat-to-beat variability and long-term continuous variability analyzed from Poincaré plots, fractal correlation properties, and approximate entropy of R-R interval dynamics from 24-hour ambulatory ECG recordings in 30 healthy control subjects, 31 hemispheric stroke patients, and 15 brain stem stroke patients (8 medullary, 7 pontine) in the acute phase of stroke and 6 months after stroke.

In the acute phase, the traditional spectral components of HR variability and the long-term continuous variability from Poincaré plots were impaired (P<0.01) in patients with hemispheric and medullary brain stem stroke, but not in patients with pontine brain stem stroke, in comparison with control subjects. At 6 months after stroke, measures of HR variability in hemispheric stroke patients were still lower (P<0.05) than those of the control subjects. Various complexity and fractal measures of HR variability were similar in patients and control subjects. The conventional frequency domain measures of HR variability as well as the Poincaré measures showed strong correlations (Pearson correlation coefficient, r=0.68 to r=0.90) with each other but only weak correlations (r=0.09 to r=0.56) with the complexity and fractal measures of HR variability.

Hemispheric and medullary brain stem infarctions seem to damage the cardiovascular autonomic regulatory system and appear as abnormalities in the magnitude of HR variability. These abnormalities can be more easily detected with the use of analysis methods of HR variability, which are based on moment statistics, than by methods based on nonlinear dynamics. Abnormal HR variability may be involved in prognostically unfavorable cardiac complications and other known manifestations of autonomic failure associated with stroke.