Biomedical Engineering Reference
In-Depth Information
vation that alternation of the T wave on the electrocardiogram is correlated with
a higher risk of sudden cardiac death (41,42). Since the T wave is associated
with action potential duration, and alternation of action potential duration can be
associated with a period-doubling bifurcation (43), one possible origin of T-
wave alternans is a period-doubling bifurcation induced by rapid cardiac rates.
T-wave alternans could also arise from a 2:1 block in some regions of the heart
and a 1:1 conduction in other regions. Independent of the cause of the alternans,
the attention to this abnormality provides a new medical approach to the assess-
ment of risk of sudden death. Others have noted that reduced heart rate variabil-
ity is correlated with a higher risk for sudden death (44,45). Since reduced heart
rate variability would be expected in patients with impaired ventricular function
and elevated sinus rate associated with higher circulating catecholamines, it
might not be surprising that reduced heart rate variability is associated with a
higher risk of sudden death. It seems clear that there is an enormous amount of
information hidden in the fluctuations of the sinus rate and in the dynamics of
arrhythmias in patients who have intermittent arrhythmias. For example, ar-
rhythmias with frequent premature ventricular contractions that appear to be the
same if looked at superficially are in fact quite different when the dynamics are
dissected in detail, and only rarely has it been possible to find a good theoretical
understanding of the mechanisms consistent with the observed dynamics over
long periods. In order to find quantitative agreement between computed ar-
rhythmias based on mathematical models and observed arrhythmias, it may be
necessary to use stochastic models (17,46). Since the opening and closing of
single ion channels assume stochastic mechanisms, a physiological basis for
stochastic macroscopic behavior may exist. Indeed, recent theoretical models of
cardiac cells have employed stochastic models for ion channels dynamics to
help interpret experimental data (47) (see also Part III, chapter 3.2, by Winslow).
Algorithms for assessment of arrhythmia may be helpful in other ways. Tat-
eno and I have developed a method to assess the presence of atrial fibrillation
based on the timing of ventricular complexes. The basic idea is to compare the
histogram of the changes in RR intervals (for a given mean RR rate) with stan-
dard histograms collected from patients experiencing atrial fibrillation (48). This
type of algorithm might be implemented in a portable device, and so might be
helpful in assessing the incidence of atrial fibrillation in a patient at risk or as-
sessing the efficacy of drug therapy in a patient who has atrial fibrillation.
Work is also underway to develop new methods to stimulate the heart to
control or terminate arrhythmia (49). Simulations of defibrillation combined
with experimental studies might be useful in helping to improve the shapes and
placement of electrodes or the optimal wave form for defibrillation. In addition,
new methods might be found to regularize or terminate arrhythmias. Current
medical devices now use pacing protocols to terminate tachycardias, but it is
possible that theoretical analyses and simulation of arrhythmia might yield im-
proved algorithms (25,26). In addition, alternating reentrant cardiac arrhythmias
