Peng-Wie Hsia

Genesis of Sigmoidal Dose-Response Curve During Defibrillation by Random Shock: A Theoretical Model Based on Experimental Evidence for a Vulnerable Window During Ventricular Fibrillation

The sigmoidal dose‐response curve (percent success vs shock energy) suggests a probabilistic nature of defibrillation. The mechanism is still largely unknown, however, random variation in the excitable state during ventricular fibrillation (VFJ is suspected. A canine defibrillation study was designed to determine whether random variation in absolute VF voltage (AVFVJ of crude marker of number of excitable cells) was related to success of defibrillufion, using a DC shock successful at the 50% level. The results were: (a) fransmyocardial resistance (73.4 ± 1.4 vs 73.6 ± 1.5 ohms) and delivered energy (6.1 ±1.2 vs 6.2 ± 1.2 joules) were similar; however, (b) AVFV 2 msec prior to DC shock was greater for successful as compared to unsuccessful attempts (0.5 ± 0.1 vs 0.3 ± 0.0 mV, P < 0.01). A mathematical model was subsequently developed based on fluctuation in the number of excitable cells. Variation in the state of excitability resulted in a cyclic window potentially vulnerable to defibrillation. The vulnerable window occurred at a point when the number of excitable cells was low, i.e., a higher state of total depolarization, which was in agreement with the experimental finding. For a given VF pattern, duration of the vulnerable window was regulated by the shock energy. A larger shock energy generated a wider vulnerable window and, in turn, a higher success rate. Finally, the sigmoidal dose‐response curve of defibrillafion was theoretically constructed by calculating the variable chances of a random DC shock occurring either in a vulnerable window or elsewhere during VF. It is concluded that a vulnerable window susceptible to defibrillation can be demonstrated in the early stages (10 sec) of VF. The mathematical model provides a theoretical basis for the vulnerable window and helps elucidate the probabilistic nature of defibrillation.

Changes in transmyocardial impedance during prolonged ventricular fibrillation. Implications for current flow and delivered energy during DC countershock

Transthoracic resistance (TTR) and transmyocardial resistance (TMR) were measured during 10 minutes of uninterrupted ventricular fibrillation (VF) in a canine model. TMR was measured at 10- to 50-second intervals with two wire-mesh patch electrodes in 16 dogs. TTR was measured through two identical low-impedance electrodes. A monophasic exponentially truncated pulse with a duration of 5 msec was used for measurement of TMR as well as TTR. Low-energy pulses of 100 V were used for TMR measurements and pulses of 300 V for TTR measurements. TMR showed an increase of 22.8 +/- 5.14 omega (from 96.2 +/- 12.3 omega) after 600 seconds of uninterrupted VF (p less than 0.0006). TTR showed a change of 2.4 +/- 1.94 omega, which was not statistically significant in comparison with the initial value of 69.0 +/- 11.4 omega. A mathematical model was developed to predict energy delivered to the heart after a transthoracic shock. Observed values of TMR and TTR were used in this model. With the use of this model, the predicted fall in transmyocardial current after 600 seconds of uninterrupted VF and 19.3%, and the fall in energy delivered to the heart was 14%. Our study suggests that increase in TMR may contribute to the observed lack of successful defibrillation during prolonged VF.

Variation of cycle length in epicardial electrograms: a quantification of chaos of ventricular fibrillation using a sock electrode array

In order to quantify the randomness of ventricular fibrillation (VF) the variance in electrogram in the heart was measured. Cycle length (CL) of activation was determined, and 10936 activations were used to calculate the variance of CL globally and at each electrode on the surface of the heart. A fractal dimension and chaotic trajectory analysis using a nonlinear dynamic model indicated that VF signal is a non-specific random event with chaotic measurement. It is concluded that CL and its variance become progressively greater with time. The change in the variance is linear, suggesting an orderly increase in conduction time during VF. The VF is not completely random. These findings suggest that complete VF can be described by a linear part (constant increase of variance of CL) and nonlinear part (mean CL).<<ETX>>

Association of short term heart rate spectral power with onset of spontaneous ventricular tachycardia or ventricular fibrillation

Heart rate (HR) variability immediately prior to the onset of a spontaneous ventricular tachycardia or ventricular fibrillation (VT/VF) may be higher. Power spectral estimation using periodogram technique was performed on HR obtained from 135 sets of 1000 RR intervals that preceded spontaneous onset of VT/VF and 125 control sets. Ten frequency bands, each of 0.1 Hz bandwidth were analyzed. Time course analysis of spectral power was also performed at 7 non-overlapping segments of 100 samples (50 seconds) each prior to time of onset of VT/VF (ONSET) and on control data sets. The RR interval rhythms associated with ONSET had higher power in all spectral bands compared to the control group (p<0.01). It was also found that spectral power increased suddenly in the 50 seconds before ONSET. A HR spectral analysis may be useful in predicting the ONSET.

Synchronized defibrillation shock delivery based on endocardial ventricular fibrillation waveform analysis is lead dependent

Previous studies have demonstrated that the defibrillation efficacy can be improved if the shock is synchronized at a high absolute VF voltage (AVFV) peak recorded from lead II of the ECG. In this study, two different endocardial lead configurations SY1:RV-coil(+)/spl rlhar2/(SVC-coil+Can)(-) and SY2:RV-coil(+)/spl rlhar2/ Can(-), were tested utilizing the peak shock method in a close-chest canine model (SY1:n=2, SY2:n=7). The leads were used both for sensing the V/F signal and for the shock delivery. The performance of this new peak shock method was compared to the conventional method of shocking at fixed time in 225 paired trials. For SY1, the peak shock method did not improve the success comparing with the conventional shock method. For SY2, peak shock method performed better in four (4) of the seven (7) subjects (8%-40% increase) and no difference in one. In the remaining two cases, additional 4-5 shock trial pairs were tested using lead II of the ECG for sensing and the performance with the peak shock was improved. The synchronized shock based on VF waveform analysis using endocardial electrograms recorded from the same electrodes for shocking may be potentially useful for implantable devices if further studies can identify, an optimal endocardial lead.