Author + information
- Received August 24, 1998
- Revision received November 6, 1998
- Accepted December 23, 1998
- Published online April 1, 1999.
- ↵*Reprint requests and correspondence: Dr. Nabil El-Sherif, SUNY Health Science Center, Cardiology Division, Box 1199, 450 Clarkson Ave., Brooklyn, New York 11203.
The purpose of this study was to investigate the electrophysiologic mechanism(s) that underlie the transition of one or more short–long (S-L) cardiac sequences to ventricular tachyarrhythmias (VTs) in the long QT syndrome.
One or more S-L cardiac cycles, usually the result of a ventricular bigeminal rhythm, frequently precedes the onset of VT in patients with either normal or prolonged QT interval. Electrophysiologic mechanisms that underlie this relationship have not been fully explained.
We investigated electrophysiologic changes associated with the transition of a S-L cardiac sequence to VT in the canine anthopleurin-A model, a surrogate of LQT3. Experiments were performed on 12 mongrel puppies after administration of anthopleurin-A. Correlation of tridimensional activation and repolarization patterns was obtained from up to 384 electrograms. Activation–recovery intervals were measured from unipolar electrograms and were considered to represent local repolarization.
We analyzed 24 different episodes of a S-L sequence that preceded VT obtained from 12 experiments. The VT followed one S-L sequence (five episodes), two to five S-L sequences (12 episodes) and more than five S-L sequences (seven episodes). The single premature ventricular beats coupled to the basic beats were consistently due to a subendocardial focal activity (SFA). There were two basic mechanisms for the development of VT after one or more S-L sequences: 1) in 10 examples of a S-L sequence due to a stable unifocal bigeminal rhythm, the occurrence of a second SFA, which arose consistently from a different site, infringed on the pattern of dispersion of repolarization (DR) of the first SFA to initiate reentrant excitation; 2) in the remaining 14 episodes of a S-L sequence, a slight lengthening (50 to 150 ms) in one or more preceding cycle lengths (CLs) resulted in alterations of the spatial pattern of DR at key sites to promote reentry. The lengthening of the preceding CL produced differentially a greater degree of prolongation of repolarization at midmyocardial and endocardial sites compared with epicardial sites with consequent increase of DR. The increased DR at key adjacent sites resulted in the development of de novo zones of functional conduction block and/or slowed conduction to create the necessary prerequisites for successful reentry.
The occurrence of VT after one or more S-L cardiac sequences was due to well defined electrophysiologic changes with predictable consequences that promoted reentrant excitation.
One or more short–long (S-L) cardiac cycles, usually the result of a ventricular bigeminal rhythm, frequently precede the onset of malignant ventricular tachyarrhythmias (VTs). This is seen in patients with organic heart disease and apparently normal QT interval (1)as well as in patients with either the congenital (2)or acquired (3,4)long QT syndrome (LQTS). Electrophysiologic mechanism(s) that underlie this relationship have not been fully explored. The present study investigates the transition from S-L cardiac cycles to VT by correlating the tridimensional activation and repolarization patterns in a well documented experimental surrogate model of the congenital LQT3 syndrome (5,6).
The study was approved by the Animal Studies Subcommittee of the local institutional review board and conformed to the guiding principles of the Declaration of Helsinki. Experiments were performed on 12 purpose-bred mongrel puppies, 12 to 14 weeks old. Puppies were preanesthetized with sodium thiopental (17.5 mg/kg intravenously) via cannulation of the cephalic vein. Puppies were then intubated and anesthetized with 1.0% to 2.0% isoflurane (vaporized in 100% O2) via a positive ventilation anesthesia machine (F500, The Forreger Co., Hauppage, New York). Catheters were inserted in the femoral vein for administration of fluids and drugs, and the femoral artery to monitor blood pressure. Electrocardiographic leads I, aVF and V1 and blood pressure (Statham, Gould, Oxnard, California) were continuously monitored on a physiologic recorder (VR12, PPG Industries, Pleasantville, New York). The heart was exposed through a midsternotomy, and a pericardial cradle was constructed. Wild-type anthopleurin-A (7)was administered as an intravenous bolus of 25 μg/kg followed by a maintenance dosage of 1.0 μg/kg/min. To slow the heart rate, vagal stimulation was taccomplished by insertion of polyimide-coated silver wires (75 μm diameter), exposed 2 to 3 mm at the tip, into the right and left cervical vagosympathetic trunks. Square pulses of 0.3 ms were delivered at 0.1 to 3 V and a frequency of 20 Hz. On completion of the experiment, the animal was euthanized by electrical induction of ventricular fibrillation and extirpation of the heart performed under general anesthesia.
Data acquisition and isochronal mapping
Details of the mapping technique have been previously reported (5,6). Unipolar electrograms were acquired using three variable gain 128-channel multiplexed data acquisition systems (DSC 2000, INET Corporation, Newark, California), allowing simultaneous recording of up to 384 signals. Each electrogram was amplified and filtered with a fixed high pass setting of 0.05 Hz and an adjustable low pass setting of 500 to 1,000 Hz. The analog data were digitized with 12-bit resolution at a sampling rate of 1,000 to 2,000 samples per second per channel. The digitized signals were then stored on hard disk on an IBM-compatible computer system (486PC, Touche Co., Chicago, Illinois). For activation–recovery interval (ARI) calculation, electrograms were further filtered off-line. Up to 384 unipolar electrograms could be recorded, from which 192 bipolar electrograms could be synthesized off-line, by taking the arithmetic difference between two neighboring unipolar recording sites (5). The timing of selected landmarks in each activation and recovery complex was automatically computed and stored for later analysis. Activation times were determined using previously published criteria (8,9). Computer-generated isochrones of activation were derived from the activation time data and delineated by closed contours at 20-ms intervals beginning with the earliest detected time of activation, and labeled in the figures as 1, 2, 3 and so on to make it easier to follow the activation patterns of successive VT beats. For the whole ventricle activation maps, zones of functional unidirectional conduction block were identified using previously defined criteria (8,9). A continuous line, or surface, was drawn through these regions and was defined as a zone of functional conduction block.
Unipolar signals were low pass filtered utilizing a digital eight-pole Butterworth filter (fc= 50 Hz) before computation of temporal derivatives. Activation–recovery interval was defined as the interval between the time of minimum first derivative (Vmin) of the QRS and the maximum first derivative (Vmax) of the T wave of unipolar electrograms (5).
Where applicable, data were analyzed using repeated measures analysis of variance and unpaired ttest (SYSTAT for Windows, version 5.0, SPSS Inc., Chicago, Illinois). A p value ≤0.05 was considered statistically significant.
A S-L cardiac sequence in the form of a premature ventricular beat coupled to the long QT interval of the basic (sinus) beat commonly preceded the occurrence of VT (defined as three or more consecutive beats). We analyzed 24 different episodes of a S-L sequence that preceded VT obtained from 12 experiments. The VT followed one S-L sequence (five episodes), two to five S-L sequences (12 episodes) and more than five S-L sequences (seven episodes). The single premature ventricular beats coupled to the basic beats were consistently due to a subendocardial focal activity (SFA). There were two basic mechanisms for the development of VT after one or more S-L sequences.
A second SFA initiates reentrant excitation
In 10 separate examples of a S-L sequence due to a stable unifocal bigeminal rhythm, the occurrence of a second SFA, which arose consistently from a different site, infringed on the pattern of dispersion of repolarization (DR) of the first SFA to initiate reentrant excitation. This is illustrated in Figures 1 and 2. ⇓Figure 1Ashows a regular bigeminal rhythm due to a SFA from the same site (marked by stars) coupled to sinus beats having a long QT interval. In panel B, a second SFA from a different site (with different QRS configuration) developed during the QT segment of the first SFA and initiated three consecutive reentrant beats (marked R). The next sinus beat was followed by a short-coupled SFA from a different site that initiated sustained reentrant VT. The second beat of the VT is labeled as possibly focal because a complete reentrant pathway could not be mapped between the first and second beats.
Figure 2illustrates the tridimensional activation maps of the first and second beats of the nonsustained VT shown in Figure 1B(labeled V1 and V2, respectively) as well as selected electrograms along the reentrant pathway of the V2 beat. The V1 beat had a subendocardial focal origin (marked by a star in section 2). Ventricular activation was completed within 100 ms, and there was no evidence of slowed conduction or conduction block. The V2 beat had a different subendocardial focal origin (marked by a star in section 1). Even though the V2 beat had a relatively long coupling interval to the V1 beat (560 ms), nevertheless, it infringed on the DR after the V1 beat, resulting in multiple zones of functional conduction block and a circulating wave front that initiated reentrant excitation. The selected electrograms in the right panel of Figure 2illustrates the diastolic bridging of activation between the V2 and V3 beats. The recording also shows that the SFA in electrogram A preceded the onset of QRS of the V2 beat in the surface electrocardiogram (ECG) by 80 to 100 ms.
Alterations of the pattern of DR that promote reentrant excitation
In the remaining 14 episodes, slight lengthening of one or a few preceding cycle lengths (CLs) resulted in critical increase of the local pattern of DR that promoted reentrant excitation. In some of these episodes, an additional factor was a change in the site of origin and local coupling of the premature beat that initiated reentry. Figures 3 to 5⇓⇓illustrate a typical example of VT that followed a single S-L sequence. Figure 3shows the surface ECG. After a series of relatively regular sinus rhythm at CLs of 600 to 620 ms, there was a sudden increase of the sinus CL to 700 ms with obvious lengthening of the QT interval of the sinus beat and the occurrence of a single ventricular premature beat (V1). The premature beat was followed after a compensatory pause of 830 ms by a sinus beat, thus creating a S-L sequence. The QT interval of the sinus beat that followed the S-L sequence showed further prolongation of the QT interval and the occurrence of a ventricular premature beat with a different QRS configuration (V2) that initiated VT.
Figure 4illustrates the tridimensional activation pattern of the two ventricular premature beats, V1 and V2, shown in Figure 3, as well as selected electrograms along the reentrant pathway initiated by the V2 beat. The two V beats arose from two different sites (marked by stars in section 3 of the maps and by electrograms C and A for V1 and V2 beats, respectively). The V2 beat had a shorter “local” coupling interval compared to V1. The V1 beat resulted in multiple zones of functional conduction block, but there was no significant area of slow conduction, and the total ventricular activation time was 100 ms. By contrast, the V2 beat resulted in more extensive zones of functional conduction block and a slow circulating wave front in section 2 to initiate the first reentrant cycle.
Figure 5shows unipolar electrograms recorded from two plunge needle electrodes (in sections 1 and 3, respectively). Recordings from electrode sites #1 and #4 in needle A are not shown. The recordings illustrate the alterations in the repolarization pattern and DR that followed the lengthening of preceding CL and that created the substrate for reentrant excitation. The numbers in the figure represent the local ARIs, and the numbers in parentheses are the cardiac CLs. Needle A shows that the increase of the sinus CL preceding V1 resulted in lengthening of ARI of all epicardial (Epi), midmyocardial (Mid) and endocardial (End) sites compared with preceding sinus beats with shorter and relatively constant CLs. The longer compensatory CL after V1 resulted in further lengthening of ARIs of the next sinus beat. Critical analysis revealed that the degree of lengthening of ARI at Epi sites was less compared with subEpi, Mid and End sites, resulting in greater dispersion between these sites. For needle A, the dispersion of ARIs between Epi site #8 and “adjacent” subEpi site #7, separated by 1 mm, was 10 ms during the stable sinus rhythm at a CL of 600 ms, and increased to 19 ms after the lengthening of the last sinus cycle before V1 to 700 ms. The dispersion of ARI then increased to 37 ms after the longer CL of 833 ms of the S-L sequence. Needle B showed similar directional increases of local ARIs after the lengthening of the preceding CL, but the degree of lengthening was more pronounced. Still, the lengthening of ARI at Epi sites was less marked compared with Mid and End sites. The lengthening of the sinus CL from 600 ms to 700 ms resulted in 19-ms and 38-ms increase of the ARI at the two most Epi sites #8 and #7, respectively, compared with Mid/End sites (ranging from 65 ms at site #6 to 195 ms at site #2). The most illustrative consequence of differential changes in ARI in response to lengthening of preceding CL is seen in the sinus beat after the S-L sequence in needle B. Conduction block occurred between Mid sites #5 and #4. The ARIs could only be estimated at sites #6 to #8 and showed further lengthening compared with the sinus beat before V1. The ARI could not be accurately estimated at sites #1 to #5 because of superimposition of the local activation potential (site #5) or electrotonic potentials (sites #1 to #4) on the repolarization wave. However, it is clear that the dispersion of local ARI between sites #5 and #4 was the substrate for the resulting functional conduction block.
Figure 6illustrates one example of VT initiation after a series of unifocal ventricular premature beats in association with lengthening of the preceding CL(s). The recording was obtained during vagally induced bradycardia. The coupling interval of the bigeminal beats was constant at 520 ms, and the returning sinus cycle was 1,220 ms. However, the two returning sinus cycles preceding the onset of VT were prolonged by 30 and 100 ms, respectively. Figure 7, right panel, illustrates the activation map of section 3 of the penultimate bigeminal beat (A) and the last begimeninal beat that initiated VT (B). Figure 7, left panel illustrates selected electrograms of both beats. The numbers in parentheses represent the calculated ARI in ms of the sinus beat to which the bigeminal beat was coupled. The electrograms labeled A∗ and B∗ were recorded at the site of origin of bigeminal beats A and B, respectively. Both beats arose from the same right ventricular subendocardial site in section 2 (not shown in the figure) and had similar coupling interval to the preceding local sinus activation. Electrograms A1 to A4 and B1 to B4 represent the same four sites on the activation maps of bigeminal beats A and B, respectively. The ARIs at sites 1 to 3 were 19 to 21 ms longer during B when compared with A; the ARI at site 4 was 37 ms longer. The dispersion of ARI between sites A3 and A4 was 20 ms and increased to 37 ms between sites B3 and B4. The longer ARI dispersion between adjacent sites B3 and B4 could explain the development of functional conduction block between these two sites. The development of zones of functional conduction block in section 3 was critical for the occurrence of reentrant excitation after bigeminal beat B compared with preceding bigeminal beats.
We analyzed 64 plunge needle recordings from nine different experiments in which an episode of VT onset was associated with lengthening of the preceding CL. In each episode, we analyzed ARIs from an average of 14 sites (range 13 to 15) before and after CL prolongation in each dog (unpaired Student ttest) as well as in the entire group (repeated measures analysis of variance). We compared the changes in ARI at Epi/subEpi sites (plunge needle electrodes #8 and #7) and Mid sites (plunge needle electrodes #6 and #5). The average control ARI (before CL lengthening) at Epi sites was 405 ± 89 ms (n = 128) and increased to 425 ± 96 ms after a 50- to 150-ms lengthening of the preceding CL. On the other hand, the average control ARI at Mid sites was 438 ± 101 ms (n = 128) and increased to 482 ± 105 ms after the same range of lengthening of the preceding CL. The degree of ARI increase in response to lengthening of the preceding CL was significantly more pronounced at Mid compared with Epi sites (p < 0.01). We also analyzed the increase in ARI dispersion between adjacent plunge needle electrode sites #7 (subEpi) and #6 (Mid) in the 64 plunge needle recordings. The ARI dispersion significantly increased from 19 ± 8 ms to 38 ± 12 ms (p < 0.002).
The present study illustrates two different electrophysiologic mechanisms that underlie the relationship between the S-L sequence and the onset of VT: 1) a second SFA (always from a different site) could infringe on the DR of the first SFA to initiate reentrant excitation; and 2) a slight lengthening (50 to 150 ms) of one or more preceding CLs could result in alterations of the spatial pattern of DR at key sites to promote reentry. The lengthening of the preceding CL produced differentially a greater degree of prolongation of repolarization at Mid/End sites compared with Epi sites with consequent increase of DR. The increased DR at key adjacent sites could result in the development of de novo zones of functional conduction block and/or slowed conduction to create the necessary prerequisites for successful reentry. Although it may be difficult to correlate an increased DR at adjacent sites with slowed conduction, it is reasonable to suggest that if a critical degree of DR between adjacent sites can result in failure of propagation of the activation wave front, a lesser degree of dispersion may result in slowed propagation (10). The SFA that initiates VT could arise from the same site of previous bigeminal SFA(s) and have the same local coupling interval (Figs. 6 and 7), or from a different site with a different coupling interval (Figs. 3 to 5). In the latter example, a combination of a shorter local coupling interval, a greater degree of DR and possibly a more appropriate site of origin of the SFA in relation to the underlying pattern of DR may all combine to create the substrate of reentrant excitation.
Several studies have shown that in ventricular myocardium, as in other cardiac tissue, the refractory period is dependent on the immediately preceding CL as well as on a number of preceding CLs, the so-called cardiac memory (11–14). If the change in the preceding CL is maintained, several cycles will be required before a steady state of refractoriness is reached. The effect of the immediately preceding CL is always the greatest, but the degree of change varies from one cardiac tissue to the other. For example, in the canine postinfarction heart, we previously reported that the interposition of a single long cycle in a pacing protocol could facilitate the induction of reentrant excitation because of differential lengthening of refractoriness in the epicardial border of the ischemic zone (10). The lengthening of refractoriness associated with a single long cycle ranged from 44% to 79% of the total increase in refractoriness that would have occurred if the same long cycle was to be maintained until a steady state is reached (10). The epicardial sites closer to the center of the ischemic zone with longer refractoriness showed a great dependence on the immediately preceding cycle compared with normal sites with shorter refractoriness. It was suggested that ischemic myocardium has less memory of the cumulative effects of preceding CLs compared with normal myocardium (10). In the present study, the degree of prolongation of ARI after lengthening of the preceding CL was greater at Mid/End zones compared with Epi zones. This is consistent with the original observation of Antzelevitch et al. showing that M cells had the steepest action potential duration (APD)–CL relationship compared with subepicardial cells (15). This observation was later confirmed in vivo in the present experimental model of LQT3 (5). However, the same may not apply to other models of LQTS.
The trigger and the substrate
This and previous reports (5,6)have clearly demonstrated that the trigger for VT in this model of LQTS is a SFA that interacts with an underlying substrate of DR to initiate reentrant excitation. The onset of VT after one or a series of bigeminal beats represents the culmination of interaction between two independent processes. The mere presence of a bigeminal rhythm is a demonstration that premature focal discharge is possible (most probably an early afterdepolarization [EAD]-triggered action potential). In the absence of an underlying suitable substrate of DR, the bigeminal rhythm will not initiate reentrant excitation. On the other hand, in the absence of an appropriate trigger, a substrate of DR may not be sufficient per se to initiate reentry. It is correctly argued that a bradycardic pause can facilitate both the initiation of an EAD-trigger beat and an increase of the underlying DR. However, the dynamics of the interaction between a bradycardic pause and these two processes may be independent.
Other mechanisms by which the premature beat can perpetuate an arrhythmia independent of the compensatory (or long) pause that follows have recently been reported (16). A single premature beat can induce or facilitate transient EAD activity, APD prolongation in canine ventricular M cells and increased DR in preparations pretreated with an IKrblocking agent.
It was difficult to accurately calculate the ARI when a local depolarization potential was closely associated with the terminal part of the T wave or when an electrotonic potential was superimposed on the T wave. However, the changes in the local ARI in the study were always correlated with appropriate changes in propagation of the activation wave front. The positive correlation of ARI and changes in propagation also confirms the validity of measurement of ARI as a surrogate for local refractoriness. Care should be exercised when extrapolating the result of the present study to the clinical setting at large. The clinical LQT3, of which the present experimental model is a valid surrogate, seems to be the least common variant of the congenital LQTS. On the other hand, the common observation of the S-L cardiac sequence preceding the onset of VT in both congenital and acquired LQTS (4,5), as well as in the absence of long QT (1), argues that the present mechanisms may also apply to these settings.
We wish to thank Dr. Dmitry O. Kozhevnikov and Victoria Stoyanovsky for their excellent contribution to surgery and manuscript preparation. We also thank Dr. Cathy Chunlian Luo for technical support. The authors also wish to acknowledge Joyce Ince for exemplary care of the animals.
☆ Supported in part by Veterans Administration Medical Research Funds to N.E.S. and M.R. and a Medtronic Japan Fellowship Award to M.C.
- action potential duration
- activation–recovery interval
- cycle length
- dispersion of repolarization
- early afterdepolarization
- long QT syndrome
- subendocardial focal activity
- ventricular tachyarrhythmia
- Received August 24, 1998.
- Revision received November 6, 1998.
- Accepted December 23, 1998.
- American College of Cardiology
- Leclerq J.F.,
- Maison-Blanche P.,
- Cauchemez B.,
- Coumel P.
- Viskin S.,
- Alla S.R.,
- Barron H.V.,
- et al.
- Kay G.N.,
- Plumb V.J.,
- Arciniegas J.G.,
- et al.
- El-Sherif N.,
- Caref E.B.,
- Yin H.,
- Restivo M.
- El-Sherif N.,
- Chinushi M.,
- Caref E.B.,
- Restivo M.
- Gallagher M.J.,
- Blumenthal K.M.
- Restivo M.,
- Gough W.B.,
- El-Sherif N.,
- Gough W.B.,
- Restivo H.
- Han J.,
- Moe G.K.
- Janse M.J.,
- van der Stern A.B.M.,
- van Dam R.H.T.,
- Durrer D.
- Wiener I.,
- Kunkes S.,
- Rubin D.,
- et al.
- Elharrar V.,
- Surawicz B.
- Antzelevitch C.,
- Sicouri S.,
- Litovsky S.,
- et al.