Three-dimensional Mapping of the Initiation of Nonsustained Ventricular Tachycardia in the Human Heart
Background Elucidation of the electrophysiological mechanisms of nonsustained ventricular tachycardia (VT) in humans is required to define the relationship between nonsustained VT and sustained VT. This goal requires, at least in part, analysis of transmural ventricular activation in patients with both sustained and nonsustained VTs.
Methods and Results We analyzed three-dimensional intraoperative cardiac maps of extrastimuli and beats during 44 nonsustained VTs and the initiating beats of 6 sustained VTs from six patients with healed myocardial infarcts who were undergoing arrhythmia surgery. The coupling interval, total activation time, and diastolic interval of each extrastimulus and beat of nonsustained VT were compared with counterparts during sustained VT. Sites activated last during extrastimuli initiating nonsustained or sustained VTs occurred in the same region, and activation times were comparable. However, the site of earliest activation during the initial or subsequent beats of nonsustained VT was discordant from the site activated earliest during the first and subsequent beats of sustained VT in 74% of cases. The mean variance in coupling interval, but not total activation time or diastolic interval, was significantly greater for VT that terminated before the 10th cycle than for VT that sustained. When analyzed from the last extrastimulus up to the fifth VT cycle, the standard deviation of the coupling interval, but not of the total activation time, was greater for nonsustained than for sustained VTs. Electrode density was sufficient to define an arrhythmia mechanism for 36 beats of nonsustained VT. Twenty-one (58%) initiated in the subendocardium, midmyocardium, or epicardium by a macroreentrant mechanism, and 15 (42%) initiated in the subendocardium by a focal mechanism.
Conclusions Compared with sustained VT, nonsustained VT initiates at discordant sites, is characterized by oscillations in coupling interval but not in total activation time, and initiates by either a macroreentrant or a focal mechanism.
The electrophysiological mechanisms of nonsustained VT and how they relate to those responsible for sustained VT in patients with healed myocardial infarction remain to be determined. A more thorough understanding of the determinants of whether VT terminates or sustains is desirable and should contribute to the evolution of management strategies in patients with nonsustained VT.1
Nonsustained and sustained VTs may differ in the extent of conduction delay induced by programmed extrastimuli, the spatial location of myocardial sites critical to VT initiation, the degree of oscillations in activation or recovery during the initial beats of VT, and/or the underlying electrophysiological mechanisms. The first step in addressing these issues is elucidation of the complex sequence of transmural ventricular activation during nonsustained and sustained VTs; this requires, as shown previously,2 simultaneous recording from multiple intramural as well as endocardial and epicardial sites. Accordingly, we performed intraoperative cardiac mapping and analyzed the three-dimensional activation sequences of nonsustained and sustained VTs induced by programmed stimulation in six patients undergoing arrhythmia surgery to determine (1) the extent to which conduction delay elicited by programmed ventricular extrastimuli is a determinant of whether VT sustains or terminates, (2) the concordance of myocardial sites that initiate sustained and nonsustained VTs, (3) the influence of dynamic changes in coupling interval and conduction on whether VT is sustained or nonsustained, and (4) the electrophysiological mechanism(s) of nonsustained VT. The electrophysiological mechanisms responsible for termination of VT were also analyzed in these patients and are reported in a companion study.3
VT was considered sustained if it was >30 seconds in duration or it required termination in <30 seconds because of loss of consciousness. Nonsustained VT was <30 seconds in duration.
Subjects included six patients with ischemic heart disease and sustained monomorphic VT who underwent intraoperative three-dimensional mapping and arrhythmia surgery at Barnes Hospital and in whom nonsustained VT was induced by programmed stimulation. The mean age was 57±7 years. Each patient had a healed (2 months to 12 years) myocardial infarct (inferior, n=3; anterior, n=3). The mean left ventricular ejection fraction was 37±13%. Programmed ventricular stimulation and endocardial catheter mapping during sustained VT were performed preoperatively in all patients. Sustained VT of at least one morphology was induced in all six patients. Sustained monomorphic VT was induced in five of these patients and nonsustained VT was induced in all six patients at the time of intraoperative mapping. Antiarrhythmic drugs, with the exception of amiodarone, were discontinued for at least five half-lives. In patient 4, the preoperative and intraoperative studies were performed during treatment with amiodarone. Amiodarone was discontinued 1 week before surgery in patient 5. Both patients remained inducible into sustained and nonsustained VTs despite the use of amiodarone.
Three-dimensional Intraoperative Mapping
Transmural and transseptal ventricular mapping was performed under normothermic bypass with a computer-assisted three-dimensional cardiac mapping system capable of simultaneously recording from 160 transmural sites.2 3 4 5 6 As many as 34 plunge-needle electrodes (136 total recording sites), each containing four bipolar electrode pairs separated by 4 mm (with an interbipole distance of 0.5 mm), were inserted in the left and right ventricles. The interelectrode distance ranged from 1.0 to 3.0 cm. Electrode placement was determined by the size and location of the infarct and the results of the preoperative electrophysiological study. Needle density was concentrated in regions surrounding the infarct as described previously.2 Surface ECG leads I, aVF, and V5R were recorded. Programmed stimulation with bipolar electrodes contained on plaques sutured to the epicardial surfaces of the right and left ventricles was performed at twice diastolic threshold with a 2-ms pulse width. The choice of paced cycle lengths used to induce arrhythmias was guided by data acquired during the preoperative electrophysiology study. The methods of stimulation were similar during initiation of sustained or nonsustained VTs except for the coupling interval of the extrastimuli.
Bipolar electrograms were sampled at 2 kHz, filtered from 40 to 500 Hz, amplified, and digitized with 12-bit precision. Digital data were stored continuously in 12 parallel bits on a Sangamo-Weston Sabre IV high-density recorder and analyzed off-line with the use of a Micro VAX II computer (Digital Equipment Corp) equipped with high-resolution color graphics. Electrode localization was determined with the use of a 53-site epicardial grid and a high-resolution video camera (Panasonic) mounted over the operating table that continuously recorded the placement of all electrodes on videotape. Multiple detailed still-frame photographs from the video recordings were used to draw maps to scale on a coordinate system.2
Construction of Activation Maps
Electrograms were analyzed with an automated system.2 3 4 5 6 Computer-assigned activation times were based on a peak criterion with peak amplitudes of ≥0.25 mV interpreted to represent discrete activations. Review and editing of electrograms were performed by the operator. Three-dimensional activation maps were constructed from data acquired during programmed stimulation and during nonsustained and sustained VTs. Individual beat-to-beat transmural activation maps were hand-drawn on templates of four short-axis slices of the heart from 1.5 to 3 cm thick that were derived from scaled drawings of pathological specimens from patients with prior myocardial infarction.2 Measurements taken from still photographs of the heart made from video recordings during electrode insertion were used to localize electrode positions on the diagrams.2 After assignment of activation times to respective locations on the templates, isochronal lines were drawn in 20-ms increments.
Conduction block between two electrodes was defined by one of three criteria: (1) intervening electrodes exhibited no activation, (2) recordings from electrode sites distal to a block demonstrated low-voltage electrotonic activity followed closely by a larger-amplitude electrogram, or (3) large temporal gaps (>60 ms) existed between two electrodes, with adjacent electrodes in a less direct spatial path demonstrating sequential activation.2 3 4 5 6
The mechanism for a VT beat was defined as macroreentrant when (1) there was continuous depolarization from the preceding beat, (2) the site of initiation of a VT beat was adjacent to the site of termination of the preceding beat, and (3) the conduction velocity of the activation wave front from the site of termination of the preceding beat to the site of initiation of the VT beat was similar to the conduction velocity of the terminal portion of the activation wave front of the preceding beat.2 3 6 The mechanism was defined as focal when the site of initiation of a VT beat demonstrated radial spread of activation and was remote from the site of termination of the preceding beat with no intervening depolarizations despite multiple (four or more) intermediate recording sites.2 3 6 A focal mechanism does not exclude the possibility of microreentry. The CI of a beat n was defined as the difference in activation times measured at the earliest site initiating beat n and the earliest site initiating the preceding beat, n−1. The TA time of a beat was defined as the difference between the activation times recorded at the sites of latest and earliest activity. The DI of beat n was defined as the difference between the CI of beat n+1 and the TA time of beat n (DIn=CIn+1−TAn).
The two-sample F test for equality of variance was used to identify significant differences in variances of CI, TA time, and DI during nonsustained and sustained VTs. There were 43 possible comparisons of sustained and nonsustained VTs in five patients with the restriction that both VTs had to have occurred in the same patient. To provide a descriptive summary, the two-sample F test for equality of variance was used to identify a trend toward larger variances among nonsustained VTs. Because the same single sustained VT was used for many of the 43 F tests, the test results could not each be considered independent evidence. Therefore, average variances and average standard deviations across sustained and nonsustained VTs (fixing the number of beats evaluated) were computed for each patient. This data-reduction step resulted in two observations per patient, and these were evaluated with a single paired t test. Data are presented as mean±SD unless indicated otherwise.
A total of 44 nonsustained VTs were recorded; 3 were >25 beats and 41 were <10 beats in duration. Six sustained VTs (from five patients) were also recorded. A total of 151 extrastimuli preceding the initiation of nonsustained VT, 289 beats of nonsustained VT, 16 extrastimuli preceding the initiation of sustained VT (from four patients), and 130 beats of sustained VT were recorded and analyzed for this study and the companion study.3 Analyses of the CI, TA, and DI were performed for all beats of nonsustained VT and for up to the first 28 beats of each sustained VT. Analyses of the site of earliest activation, the three-dimensional activation sequence, and the arrhythmia mechanism were performed for all preceding extrastimuli (151 beats) and up to the first 5 beats of nonsustained VT (177 beats) and for all preceding extrastimuli (16 beats) and the first 10 beats of sustained VT (60 beats). In total, the three-dimensional activation sequences of 404 beats were defined on the basis of analysis of >36 000 electrograms. Analyses of the activation sequence and mechanism of beats of nonsustained VT beyond the fifth beat, the terminal 3 beats of each nonsustained VT, and the beats of sustained VT beyond the 10th were performed in a companion study.3
Analysis of Extrastimuli Initiating Nonsustained and Sustained VTs
Twenty-five nonsustained VTs were initiated by two extrastimuli, and 19 were initiated by three extrastimuli. Four sustained VTs were initiated with two extrastimuli, and 1 was initiated with three extrastimuli. Analysis of activation maps during programmed stimulation demonstrated that 79% of extrastimuli that initiated nonsustained VT exhibited less conduction delay than those initiating sustained VT. The TA times for the last drive train beat, T2, and for the extrastimuli S2-S4 initiating nonsustained VT were 118±13, 136±19, 166±29, and 211 ms, respectively. The overall differences, however, were not statistically significant compared with those initiating sustained VT (130±30, 160±34, 206±49, and 213 ms, respectively). Thus, activation times leading to nonsustained VT were comparable to those leading to sustained VT.
Progressive decrease in the CI of the terminal extrastimulus usually led to greater conduction delay. As illustrated by the data acquired from patient 1 and shown in Fig 1⇓, a decrease in the T2-S2 interval by the use of double extrastimuli led to 4- and 3-beat VTs. Sustained VT was then induced and exhibited a greater degree of conduction delay of the terminal extrastimulus (268 ms) than that preceding any of the prior nonsustained VTs. Likewise, with the use of three extrastimuli, progressive shortening of the S3-S4 or S2-S3 intervals led to nonsustained VT, but the extent of conduction delay of the triple extrastimuli did not exceed that achieved with double extrastimuli. As a result, sustained VT was not induced with triple extrastimuli, even though it was induced with double extrastimuli. In this case, differences in the extent of conduction delay of the last extrastimulus did contribute to the development of sustained VT.
Although the terminal extrastimuli preceding the initiation of either nonsustained or sustained VT demonstrated variable degrees of conduction delay, the sites activated last by these extrastimuli occurred in the same region in an individual patient. Sites of conduction delay induced by the extrastimulus initiating nonsustained VT or sustained VT were identical in 7 of 29 cases (24%), and were within one recording site in the other 22 cases (76%).
Concordance of Sites Initiating Nonsustained and Sustained VTs
The first beat (T1) of nonsustained VT initiated in the subendocardium, midmyocardium, or subepicardium. Despite the proximity of the sites of latest activation of the extrastimuli initiating nonsustained VT and sustained VT, the site of initiation of T1 of nonsustained VT and the site of initiation of T1 of sustained VT were concordant in only 17% of cases. In the patients in whom there was a concordance of sites (patients 1, 3, and 5), only five nonsustained VTs initiated at the same site as sustained VT. Discordant sites were categorized as adjacent or distant (separated by more than one recording site) to sites initiating sustained VT. Among discordant sites, 68% were remote (two or more electrode sites away) and up to 3 to 6 cm from the tissue initiating sustained VT (Fig 2⇓).
The site of initiation of nonsustained VT changed from the first to the second beat in 27 cases (61%). A similar change was observed in four of five sustained VTs. In fact, the site of earliest activation could shift between as many as five subendocardial, midmyocardial, or subepicardial sites during the first beats of a sustained VT, producing an initial polymorphic phase (Fig 3⇓).
Despite the shift to different intramural sites during the initiation of sustained VT, the sites of earliest activation after the initiation of nonsustained VT were concordant with these initial beats of sustained VT in only 30% of cases, and beats of nonsustained VT rarely demonstrated activation sequences that were similar to those of beats of sustained VT. As also shown in Fig 3⇑, beats of nonsustained VT arose from multiple sites throughout the left ventricle and the interventricular septum and were usually distant (more than two sites away) from sites of earliest activation during sustained monomorphic VT, even during its initial polymorphic phase. Thus, nonsustained VT most often initiated at sites discordant from those responsible for sustained VT.
Oscillations in CI, TA Time, and DI
An analysis of TA times and CIs of the initial beats of nonsustained and sustained VTs suggested that nonsustained VT was associated with a greater degree of oscillation of the CI than was sustained VT. In Fig 4⇓, CI, TA time, and DI are displayed as functions of extrastimulus and beat number for a sustained VT (left) and a nonsustained VT (right) recorded from patient 2. Oscillations in these intervals were evident both during the initial beats of the sustained VT and during the initial and subsequent beats of nonsustained VT.To determine whether nonsustained VT is characterized by greater oscillations in CI, TA time, and DI, ventricular activation was analyzed during VTs in four patients in whom both nonsustained and sustained VTs were recorded. For each patient, the mean, median, and variance of the CI, TA time, and DI for each extrastimulus and beat of nonsustained and of sustained VT were compared. There was no significant difference between mean and median CIs, TA times, or DIs of nonsustained versus sustained VTs. No specific pattern of the direction of change in any of these variables during the last one or two beats of nonsustained VTs was identified. By use of the two-sample F test for identification of significant differences in variance, ≤10 beats of nonsustained VT were compared with corresponding beats of sustained VT from the same patient (Table 1⇓). Significantly higher variance (P<.05) in CI was observed in 17, in TA time in 16, and in DI in 11 of the 43 VT comparisons performed. Only one, three, and two nonsustained VTs had lower variances in CI, TA time, and DI, respectively. The mean variance of the nonsustained VTs <10 beats in duration was compared with that of corresponding beats of sustained VTs from the same patients by use of paired t tests (Table 2⇓). The mean variance in CI, but not in TA time or DI, was significantly greater in nonsustained VTs (P=.027).
To determine the first beats of nonsustained VT most important in producing variations in CI and conduction, CI and TA time from the last extrastimulus to the fifth cycle for nonsustained and sustained VTs were compared. The standard deviation of the CI measured from the first three or four cycles was significantly different for the two types of VTs (Table 3⇓). In addition, there was a trend toward a progressive decrease in the standard deviation of the CI as more beats of sustained VT were analyzed, an observation not seen during nonsustained VT. There was no significant difference in the standard deviation of TA time during the initiating beats of sustained and nonsustained VTs. Thus, nonsustained VT was characterized by oscillations in CI, especially during the first four cycles, that could not be explained by a greater variance in TA time.
Mechanisms of Nonsustained VT
Mapping was of sufficient density to identify the mechanisms of initiation of 36 beats of nonsustained VT and of 48 beats of sustained VT.
In eight cases, initiation of the first beat of nonsustained VT arose by intramural reentry at a CI of 225±16 ms, with the TA times of the preceding extrastimuli averaging 204±6 ms, and with a DI of 34±17 ms. In these cases, conduction delay during the last extrastimulus contributed to the initiation of VT. The first beat of VT arose in the subendocardium in four cases, in the midmyocardium in three cases, and in the epicardium in one case. In two patients, sustained VT also initiated by intramural reentry arising in the subendocardium or midmyocardium.
An example of nonsustained VT initiating by reentry is shown in Fig 5⇓. Activation during the third extrastimulus, S4, was first detected at an apical site in level IV (red dot) and propagated basally in the anteroseptal region around a line of transmural conduction block in levels II and III and then laterally in level I and apically in level II. Activation then proceeded both clockwise (180-ms isochrone; arrow) in level II as well as apically and laterally to level IV. Although activation at the latest site was due to conduction in a counterclockwise direction in the apex around a line of midmyocardial block (220-ms isochrone), it was the delayed activation in level II (180-ms isochrone) that propagated basally to initiate T1 in the epicardium of level I by intramural reentry.
The initiation of sustained VT with double extrastimuli in the same patient is shown in Fig 6⇓. Activation during the last extrastimulus, S3, which was first detected in the basal anteroseptum (*), propagated laterally and apically. Despite marked conduction delay in the apex (TA time=268 ms), it was the late activation in level III that propagated basally and led to the initiation of sustained VT in the subendocardium of level I by intramural reentry. The sustained VT in patient 1 was the only VT initiating at that site. Thus, termination sites for the last extrastimuli initiating nonsustained and sustained VTs were in close proximity, but there was discordance of sites of initiation of the first beat of VT due to intramural reentry along different pathways that were distant from the sites of terminal activation.
In one patient, nonsustained VT initiated in the midmyocardium by intramural reentry that occurred along the same reentrant pathway that initiated sustained VT (Figs 7⇓ and 8⇓). The last extrastimulus, S4, initiated in the epicardium of the lateral left ventricle and propagated in a clockwise (Fig 8⇓, A and B) and apical direction (Fig 8⇓, A, C, and D) but encountered nontransmural or transmural conduction block in levels II, III, and IV. There was very delayed activation in the subendocardium in the anteroseptal region of level III (220-ms isochrone), but it was the slow conduction in levels III and IV (Fig 8⇓, E and F) that led to initiation of T1 29 ms later at midmyocardial site B in level II (*) by intramural reentry. As shown in Fig 8⇓, initiation of T1 of the nonsustained VT and the sustained VT occurred by activation along the pathway A-C-D-E-F-B.
The first beat of nonsustained VT initiated by a focal mechanism in 10 cases, on the basis of the absence (112±16 ms) of any intervening electrical activity from the terminal activation of the last extrastimulus to the initiation of the first beat of VT, despite the presence of multiple intervening electrodes. Focal initiation occurred in the subendocardium in all cases and did so at a CI of 264±11 ms and with an average TA time of the preceding extrastimulus of 148±9 ms. No T1 of the sustained VTs was found to initiate by a focal mechanism.
Five beats of nonsustained VT were maintained by a focal mechanism arising in the subendocardium. This was associated with a TA time of the preceding beat of only 123±22 ms, which was shorter than that preceding reentrant beats during the maintenance of nonsustained VT. Eighteen beats during the maintenance of sustained VT initiated by a focal mechanism arising in the subendocardium, but in no case did focal initiation occur at the same site during both nonsustained VT and sustained VT.
An example of nonsustained VT initiating by a focal mechanism is shown in Fig 9⇓. Activation during the last extrastimulus, S4, was first detected in the posterior septum of section II. Activation proceeded rapidly through the septum, encountering lines of conduction block in the anterior and posterior septa of the apical section of level IV and terminated with a TA time of 163 ms. T1 initiated 198 ms later in the endocardium of the anteroseptal region of section II by a focal mechanism, with no intervening activity between the latest site of S4 activation and the initiation site of T1 (Fig 10⇓). Activation of T1 proceeded uniformly without marked conduction delay and terminated in the apical septum with a TA time of 92 ms. The second beat, T2, initiated in the basal posterolateral endocardium 246 ms later via a focal mechanism. Fig 10⇓ displays the electrograms from the initiating and surrounding sites from these 2 beats of VT that demonstrate the absence of intervening electrical activity and initiation by a focal mechanism.
More than one mechanism was often documented within the same nonsustained VT. Fig 11⇓ illustrates the initiation and maintenance of a 7-beat VT. The last extrastimulus, S4, initiated in the anterior epicardium of the left ventricle and encountered areas of block in sections III and IV. It terminated in the posterior septum in level II with a TA time of 132 ms. T1 initiated by a focal mechanism 100 ms later in the anterior subendocardium in level III. Activation proceeded uniformly, and T2 initiated by a focal mechanism at the same site that initiated T1 after a slightly longer DI of 116 ms. During T2, moderate conduction slowing was observed in the anteroseptal region of section II. Beats T3 and T4 also initiated via a focal mechanism at a site similar to that of the initiation site of T1 and T2. Areas of conduction slowing and block were similar to those of T2. However, further conduction slowing that developed in the anteroseptal region during T4 (TA time of 167 ms) led to initiation of T5 by intramural reentry. Delayed activation in level III during T5 ultimately led to initiation of T6 by intramural reentry as well. The TA times of beats T1-T5 (100, 100, 105, 167, and 197 ms, respectively) were not reflected by the QRS durations of these beats (165, 133, 156, 147, and 128 ms, respectively).
As shown in Figs 11⇑ and 12⇓, the transition from a focal to a reentrant mechanism was associated with the development of marked conduction delay in level III. T3 demonstrated delayed activation at site H (106 ms) in level IV, but it did not propagate to adjacent site I in level III. However, slower activation during T4 resulted in later activation at site H (122 ms) and delayed activation at site I (170 ms) that then led to the initiation of T5 by intramural reentry at site A. Thus, a nonsustained VT could be initiated and maintained by two different mechanisms.
Analysis of the data acquired demonstrates that nonsustained VT induced by programmed stimulation in patients with healed myocardial infarction is characterized by comparable conduction delay of the terminal extrastimulus but is initiated and maintained at sites discordant from those initiating and maintaining sustained VT. Not only are oscillations in CI greater for nonsustained VT, but these changes are evident within the first four cycles of the arrhythmia.
In this study, a decrease in the CI of the extrastimuli was usually associated with increased levels of conduction delay, supporting the use of closely coupled extrastimuli. Nevertheless, the patterns of activation and the sites activated last by extrastimuli initiating nonsustained and sustained VTs were similar, suggesting that the induction of sustained and nonsustained VTs is not related to differences in the activation sequence elicited by changes in the CI of the extrastimuli.
A critical determinant of whether VT was sustained or nonsustained was the location of the site of earliest activation during initiation and maintenance. Sustained VT usually initiated at one site and then changed to another initiation site by the second beat. Sometimes the site of initiation could change a number of times before stable monomorphic VT developed. Despite the multiplicity of initiation sites of sustained VT, nonsustained VT was initiated and maintained at sites that were discordant from those for sustained VT in 74% of cases.
Intramural reentry was found to underlie the initiation and maintenance of both nonsustained VT and sustained VT. In each case, reentry was dependent on the development of marked conduction delay of the preceding beat. Reentrant activation arose at endocardial, midmyocardial, or epicardial sites. The finding of marked midmyocardial conduction delay and nontransmural conduction block leading to initiation at either subendocardial, midmyocardial, or epicardial sites suggests that intramural activation plays a critical role in the development of nonsustained VT in the infarcted heart, as we have demonstrated in studies of sustained VT both in patients with healed myocardial infarcts2 and in experimental animals.5
Reentry was dependent on the development of a critical degree of conduction delay. Such delay often developed in response to programmed stimulation and led to the development of reentry at different sites. However, even in the presence of marked conduction delay, sites of reentrant initiation were not always adjacent to sites of greatest conduction delay. As a result, slow conduction, primarily in midmyocardial regions, could propagate in a number of different directions and exit via different sites, results similar to those reported from studies of patients by Downar et al.7 Although Downar et al concluded that multiple morphologies of VT could result from alterations in the reentrant pathway, we found that this could also be due to focal mechanisms arising from any of multiple subendocardial sites that are not dependent on the development of a critical degree of conduction delay.
Nonsustained VT also initiated by a focal mechanism. This focal activation was found to arise only from the subendocardium and was associated with less conduction delay of the preceding beat than beats initiating by reentry. Although microreentry cannot be excluded as a mechanism because of the resolution of the mapping performed in the present study, we believe it is unlikely. Focal initiation always arose from sites that were distant from sites of terminal activation or marked conduction delay from the preceding beat, and multiple intervening electrode recording sites failed to demonstrate any intervening electrical activity. In addition, our group has demonstrated, using high-density grid electrode mapping with an interelectrode distance of 350 mm, that microreentrant circuits in infarcted canine epicardium occur in an area as small as 0.05 cm2.8 In all of these cases, microreentry was associated with conduction delay on the order of 100 ms at sites 1 to 2 cm away. However, this was not the case for any of the beats of nonsustained VT initiating by a focal mechanism in the present study or for any focal beats of sustained VT in our previous studies in humans.2 3
The contribution of focal activation in the development of VT has only been recently appreciated. However, we have previously reported that the maintenance of sustained monomorphic VT in a similar group of patients with healed myocardial infarcts (4 months to 13 years) was due to a focal mechanism in one half of the cases in which a mechanism could be identified.2 Thus, focal mechanisms appear to play a critical role in the development of both sustained and nonsustained VTs. Why focal activation occurs repetitively at certain sites and leads to sustained monomorphic VT in some cases while in others there is focal activation from one site only for 1 or several beats remains to be determined.
Oscillations of CI and TA Time
In contrast to the results of other investigators,9 10 11 we found that nonsustained VT was characterized by greater oscillations in CI, but not TA, than was sustained VT. Duff et al9 found that nonsustained VT was associated with oscillations in CI, conduction time, and refractoriness. However, their results were based solely on analysis of the extent of TA from the QRS duration of the surface ECG. In this and our previous studies in the feline heart,12 13 we found that TA time measured in this way was not accurate, as determined by comparisons with results of three-dimensional cardiac mapping. This finding suggests that the marked conduction delay that may play a critical role in the development of reentrant excitation may occur in localized regions of the heart that are too small to affect the surface ECG.
Our findings also differ from those of Frame et al,10 11 who demonstrated greater oscillations in conduction, action potential duration, and refractoriness in nonsustained VTs than in sustained VTs. These studies were performed in a canine model of reentry around a single fixed anatomic barrier in the atria. The results of our study of nonsustained VT in the infarcted human heart and of our previous study of sustained VT in a similar patient population2 indicate that the initiation and maintenance of VT are due to focal as well as reentrant mechanisms and that reentry could occur over a number of different intramural pathways.
The cause of the oscillation in CI remains unknown. It could be related to dispersion of refractoriness. Marchlinski14 demonstrated oscillations in ventricular refractoriness in response to abrupt shortening of paced cycle lengths in the human heart, and Rosenbaum et al15 reported on the spatial characteristics of similar oscillations in action potential duration in isolated, perfused guinea pig hearts. Oscillations in CI during nonsustained VT could also be due to changes in the sites of initiation and the mechanism of initiation of beats of VT. The results of the present study demonstrate not only discordance of initiation sites for nonsustained and sustained VTs but also that multiple sites throughout the heart—subendocardial, midmyocardial, and epicardial—can be activated by a focal or reentrant mechanism. The nature of CI oscillation and its contribution to the termination of VT are explored in a companion study.3
The major limitation of the present study was the resolution of maps of the human heart. A greater density of electrodes would have allowed delineation of arrhythmia mechanism for more beats of nonsustained and sustained VT. However, the resolution was sufficient to (1) define the mechanism for 36 beats of nonsustained VT and 48 beats of sustained VT, (2) define intramural reentrant circuits even in the interventricular septum, and (3) define the site of arrhythmia initiation sufficiently to allow successful surgical ablation of VT in all of the patients studied.
The present study allowed a unique opportunity to compare the three-dimensional activation of nonsustained VT and sustained VT in patients with ischemic heart disease. The marked discordance of sites initiating and maintaining nonsustained VT supports the findings that the use of nonsustained VT induced by programmed extrastimuli in patients with a history of VT is not an adequate end point with which to assess electrophysiological instability or guide antiarrhythmic therapy.16 Furthermore, the induction of only nonsustained VT in a patient who previously demonstrated sustained VT and in whom antiarrhythmic therapy has been started has been useful as a therapeutic end point.17 18 The results of this study suggest that this may reflect alteration of activation at sites that are critical to sustained VT but that are not involved in the induction of nonsustained VT.
Selected Abbreviations and Acronyms
This work was supported in part by the Michael Bilitch Fellowship in Cardiac Pacing and Electrophysiology, the North American Society of Pacing and Electrophysiology (Dr Chung), and National Institutes of Health grants HL-46929 (Dr Pogwizd) and R01-HL-50295 (Dr Cain). The authors gratefully acknowledge the expert technical assistance of Dieter Ambos and Jerome Peirick, surgical mapping by Drs James L. Cox and T. Bruce Ferguson, and secretarial assistance from Barbara Donnelly.
- Received November 20, 1995.
- Revision received November 25, 1996.
- Accepted December 16, 1996.
- Copyright © 1997 by American Heart Association
Pogwizd SM, Hoyt RH, Saffitz JE, Corr PB, Cox JL, Cain ME. Reentrant and focal mechanisms underlying ventricular tachycardia in the human heart. Circulation. 1992;86:1872-1887.
Pogwizd SM, Chung MK, Cain ME. Termination of ventricular tachycardia in the human heart: insights from three-dimensional mapping of nonsustained and sustained ventricular tachycardias. Circulation. 1995:95:2528-2540.
Kramer JB, Saffitz JE, Witkowski FX, Corr PB. Intramural reentry as a mechanism of ventricular tachycardia during evolving canine myocardial infarction. Circ Res. 1985;56:736-754.
Pogwizd SM. Focal mechanisms underlying ventricular tachycardia during prolonged ischemic cardiomyopathy. Circulation. 1994;90:1441-1458.
Downar E, Saito J, Doig JC, Chen TCK, Sevaptsidis E, Masse S, Kimber S, Mickleborough L, Harris L. Endocardial mapping of ventricular tachycardia in the intact human ventricle, III: evidence of multiuse reentry with spontaneous and induced block in portions of reentrant path complex. J Am Coll Cardiol. 1995;25:1591-1600.
Zuanetti G, Hoyt RH, Corr PB. β-Adrenergic–mediated influences on microscopic conduction in epicardial regions overlying infarcted myocardium. Circ Res. 1990;67:284-302.
Duff HJ, Mitchell LB, Gillis AM, Stieldon RS, Chudleigh L, Cassidy P, Chiamvimonvat N, Wyse DO. Electrocardiographic correlates of spontaneous termination of ventricular tachycardia in patients with coronary artery disease. Circ Res. 1993;88:1054-1062.
Frame LH, Simson MB. Oscillations of conduction, action potential duration, and refractoriness: a mechanism for spontaneous termination of reentrant tachycardia. Circulation. 1988;78:1277-1287.
Frame LH, Rhee EK. Spontaneous termination of reentry after one cycle or short nonsustained runs: role of oscillations and excess dispersion of refractoriness. Circ Res. 1991;68:493-502.
Pogwizd SM, Corr PB. Reentrant and nonreentrant mechanisms contribute to arrhythmogenesis during early myocardial ischemia: results using three-dimensional mapping. Circ Res. 1987;61:352-371.
Pogwizd SM, Corr PB. Electrophysiologic mechanisms underlying arrhythmias due to reperfusion of ischemic myocardium. Circulation. 1987;76:404-426.
Marchlinski FE. Characterization of oscillations in ventricular refractoriness in man after an abrupt increment in heart rate. Circulation. 1987;75:550-556.
Rosenbaum DS, Kaplan DT, Kanai A, Jackson L, Garan H, Cohen RJ, Salama G. Repolarization inhomogeneities in ventricular myocardium change dynamically with abrupt cycle length shortening. Circulation. 1991;84:1333-1345.