Background The goal of this study was to define the mechanisms of spontaneously occurring ventricular arrhythmias in the setting of nonischemic heart failure.
Methods and Results Three-dimensional cardiac mapping from 232 intramural sites was performed in four rabbits with heart failure induced by combined aortic regurgitation and aortic stenosis and in four control rabbits. During the development of heart failure, serial echocardiographic examination demonstrated a progressive increase in left ventricular (LV) chamber dimensions and a decrease in LV systolic function over 19±2 months. Serial Holter monitoring demonstrated spontaneously occurring premature ventricular complexes (PVCs) (up to 13 000 per day) and couplets in all four rabbits with heart failure, and runs of nonsustained ventricular tachycardia (VT) up to 26 beats long in three. Mapping of spontaneous rhythm was performed for up to 60 minutes. None of the control rabbits demonstrated spontaneous arrhythmias during mapping. Three rabbits with heart failure demonstrated isolated PVCs, and two demonstrated couplets and runs of nonsustained VT up to 4 beats long. The three-dimensional activation sequence of 50 sinus beats (42 from rabbits with heart failure; 8 from control rabbits), 19 PVCs, and 37 beats of couplets and nonsustained VT was determined and the mechanism of arrhythmia defined for all ventricular ectopic beats analyzed. Normal sinus beats from the failing rabbits activated rapidly, with a total activation time of 28±1 ms (P=.18 versus sinus beats from control hearts, 26±1 ms). Sinus beats preceding PVCs in the rabbits with heart failure activated in a similar fashion, with a total activation time of 26±1 ms. In each case, these PVCs initiated in the subendocardium by a nonreentrant mechanism based on the absence of intervening electrical activity between the termination of the preceding beat and the initiation of the next (225±7 ms), despite the presence of multiple intervening electrode recording sites. Couplets and monomorphic and polymorphic VTs were due to repetitive nonreentrant activation at the same or different subendocardial sites. Total activation time of beats of VT averaged 44±1 ms and did not differ from that of isolated PVCs (43±2 ms, P=.65). Pathological analysis of tissue demonstrated myocardial fiber hypertrophy, degenerative changes, and interstitial fibrosis throughout the failing hearts.
Conclusions Spontaneously occurring PVCs, couplets, and VT in a model of nonischemic heart failure are due to nonreentrant mechanisms such as triggered activity or abnormal automaticity. Approaches to the treatment of spontaneously occurring ventricular arrhythmias in patients with nonischemic heart failure should be directed at nonreentrant mechanisms.
Congestive heart failure, whether caused by ischemic or nonischemic processes, is associated with a very poor prognosis, with a 2-year mortality approaching 50% in patients in New York Heart Association classes III and IV.1 In nearly one half of the cases, death is sudden,2 usually caused by ventricular tachycardia (VT) that degenerates to ventricular fibrillation (VF).3 4 Currently available antiarrhythmic treatment of these malignant ventricular arrhythmias is ineffective5 and potentially proarrhythmic.6 The development of effective antiarrhythmic therapy will be facilitated by delineation of the underlying electrophysiological mechanisms, ie, determination of whether VT is due to reentry or to a focal nonreentrant mechanism such as triggered activity or abnormal automaticity.
Unfortunately, very little is known about the electrophysiological alterations and the underlying electrophysiological mechanisms responsible for ventricular arrhythmias in the failing heart. This is due primarily to a lack of experimental animal models of heart failure that demonstrate spontaneously occurring arrhythmias and in which arrhythmia mechanisms can be elucidated by use of state-of-the-art mapping procedures. Delineation of electrophysiological mechanisms of arrhythmias occurring in vivo requires simultaneous recording from multiple (several hundred) sites throughout the entire heart, ie, high-resolution three-dimensional (3D) cardiac mapping.7
Three-dimensional mapping of spontaneously occurring ventricular arrhythmias was performed recently in an experimental model of ischemic cardiomyopathy.8 Dogs with heart failure induced by multiple intracoronary microembolizations demonstrated a progressive decrease in left ventricular (LV) systolic function,9 pathological alterations comparable to human ischemic cardiomyopathy,8 9 and spontaneously occurring premature ventricular complexes (PVCs), couplets, and runs of VT.8 9 10 Mapping from 232 sites throughout the left and right ventricles and the interventricular septum has demonstrated that PVCs and VT initiate by a focal mechanism arising primarily from the subendocardium, with no evidence of macroreentry.8 Spontaneously occurring monomorphic and polymorphic VTs are due to repetitive focal activation.8
Determination of whether arrhythmias associated with nonischemic heart failure are initiated by reentry or a nonreentrant mechanism also requires mapping to be performed in an experimental animal preparation that demonstrates frequent spontaneously occurring arrhythmias. End-stage nonischemic heart failure in humans, resulting from cardiomyopathy or chronic pressure and/or volume overload, represents a final common outcome characterized by LV dilatation, LV systolic dysfunction, severe but relatively nonspecific histological alterations,11 12 and frequent spontaneously occurring ventricular arrhythmias; however, none of the currently available models of nonischemic heart failure demonstrates all of these characteristics.
Gilson et al13 recently developed a rabbit model of nonischemic heart failure that is induced by aortic regurgitation followed by aortic constriction. After 4 months, these rabbits demonstrated marked hypertrophy and a moderate degree of LV systolic dysfunction. However, the incidence of ventricular ectopy was not assessed.13 While this preparation did not demonstrate severe heart failure at that time, chronic pressure and volume overload of greater duration could potentially lead to a much greater degree of LV dilatation and systolic dysfunction, marked histological alterations, and spontaneously occurring ventricular arrhythmias.
The purpose of the present study was to define the electrophysiological mechanisms underlying spontaneously occurring VT in the setting of nonischemic end-stage heart failure. Accordingly, the following experiments were performed: (1) rabbits with combined aortic insufficiency and aortic stenosis were followed over a long term and subjected to serial two-dimensional (2D) echocardiography (to assess LV chamber size and systolic function) and Holter monitoring (to assess the incidence and frequency of spontaneously occurring ventricular arrhythmias); (2) 3D cardiac mapping of the spontaneous ventricular arrhythmias that developed was performed; and (3) histological analysis of tissue was performed to characterize the structure of regions participating in initiation or maintenance of the electrophysiological alterations.
Studies were performed on healthy adult New Zealand White rabbits of either sex (2.9 to 3.5 kg). Heart failure was produced by induction of aortic insufficiency, which was followed 14 days later by constriction of the abdominal aorta.13 Three-dimensional cardiac mapping was performed in four rabbits in which heart failure was induced (HF rabbit group) and four control rabbits. The Washington University Animal Studies Committee reviewed and approved the experimental protocol.
Induction of Heart Failure
For the induction of aortic insufficiency, rabbits were anesthetized with ketamine (30 mg/kg), acepromazine (0.75 mg/kg), and xylazine (7.5 mg/kg IM). After the rabbit was shaved and the neck prepped, a 4F sheath was inserted into the left carotid artery. The lead II surface ECG was monitored on a Lifepak 4 (Physio-Control). A beveled polyethylene catheter (4F) connected to a pressure transducer was then introduced into the carotid artery and pushed abruptly through the aortic valve several times. Aortic insufficiency was considered adequate when the aortic pulse pressure increased by at least 50%.13 The severity of aortic insufficiency was subsequently assessed by 2D echocardiography with color-flow mapping (see below). The carotid artery was then repaired, and the incision was closed.
Aortic constriction was performed 14 days later under the same anesthetic regimen. The abdomen was shaved and prepped with betadine. The lead II surface ECG was monitored. A midline incision was made, and the abdominal aorta was isolated proximal to the renal arteries. A silk ligature was tightened around both the aorta and an adjacent piece of polyethylene catheter (2.42-mm OD). The catheter was withdrawn immediately, producing a reduction of aortic diameter of approximately 45%.13 The incision was then closed.
Before the creation of aortic insufficiency (baseline) and at approximately 3- to 6-month intervals, each rabbit underwent echocardiographic examination. Rabbits were sedated with ketamine (35 mg/kg IM) and placed in the right lateral decubitus position on a specially designed table that had a 6×6-in square hole. This allowed placement of a 5-MHz transducer on the right lateral thorax from below. Standard short-axis views were obtained below the level of the mitral valve leaflets. M-mode echocardiographic measurements of LV end-diastolic (LVEDD) and end-systolic (LVESD) dimensions were obtained, following American Society of Echocardiography recommendations.14 Fractional shortening (FS) was calculated as FS (%)=(LVEDD−LVESD)/LVEDD.
The severity of aortic insufficiency (AI) was assessed by color-flow Doppler examination of the diastolic LV outflow tract jet in the parasternal long-axis view (obtained with the rabbit in a supine position). A ratio of jet height to LV outflow tract height of ≥65% was considered severe AI; 25% to 64%, moderate AI; and <25%, mild AI.15
In addition, aortic regurgitant fraction (RF) was determined from the time-velocity integral of the pulsed Doppler tracing of right ventricular (TVI-RVOT) and LV (TVI-LVOT) outflow tract flow and measurement of the RV (Area-RVOT) and LV (Area-LVOT) outflow tract cross-sectional area (πr2, where r is half of outflow tract diameter) based on the following formula16 :
At baseline (before induction of aortic regurgitation) and at an average of once every 2 weeks after aortic constriction, Holter monitor recordings were obtained from conscious rabbits. The chest was shaved and five ECG electrodes were applied to the chest. For the first 12 months of the study, a specially designed nylon jacket (Alice King Chatham) was put on, and the Holter monitor was strapped to the rabbit’s side. Beginning 12 to 15 months after aortic constriction, Holter recordings were obtained continuously by placing the nylon jacket on the rabbit and tunneling the lead wires through a long, specially designed tether made of flexible metal tubing to a freely rotating Holter recorder placed on top of the cage. Recordings were made for 24-hour periods without the animal’s movement being restricted. Cassette tapes with two channels of ECG recordings were analyzed by use of a computer-assisted Marquette Series 8500 Holter analysis system to determine the frequency of PVCs, couplets, and runs of VT. Arrhythmia frequency was verified by manual counting.
Three-dimensional cardiac mapping was performed in control rabbits and in HF rabbits (20±2 months after aortic constriction) as described previously.17 18 19 Briefly, rabbits were anesthetized with ketamine (10 mg/kg IM) and pentobarbital (40 mg/kg IV, with additional doses of 16 to 32 mg IV as needed), intubated, and mechanically ventilated. Body temperature was maintained at 37°C by a thermostatic esophageal probe controlling an infrared lamp. Intramural recordings were obtained from specially designed plunge-needle electrodes that were described previously.17 18 19 Twenty-five plunge electrodes, each containing eight bipolar electrode pairs separated by 500 μm (200 sites), were placed in the left ventricle. Eight electrodes, each containing two bipolar pairs separated by 500 μm (16 sites), were placed in the right ventricle. Four septal electrodes, each containing four bipolar pairs separated by 2.5 mm (16 sites), were placed in the anterior and posterior septum under 2D echocardiographic guidance. All electrodes had an interbipole spacing of 500 μm. The distance between plunge electrodes was 3 to 9 mm. During the experiment, warm (37°C) saline was applied to the heart intermittently to prevent surface cooling and to moisten the epicardium.
After a stabilization period of 30 minutes, bipolar electrogram information was acquired simultaneously from each of the 232 transmural sites and individually amplified, filtered (40 to 500 Hz), and converted from analog to digital at a 2-kHz sampling rate. The digital data, along with the lead II surface ECG tracing, were stored on tape by use of a Sangamo Sabre IV high-density recorder (Fairchild Weston Systems). Spontaneous rhythm was recorded continuously for up to 60 minutes (see the “Results” section). At the end of the experiment, each animal was killed with KCl. The heart was then excised and rinsed in normal saline. Detailed electrode localization was performed as described previously.17 18 19 Each plunge-needle electrode was removed and replaced by a labeled pin. The heart was then immersed and fixed in formalin for at least 24 hours. The pins were replaced with small plastic brush bristles to facilitate sectioning of the heart transversely into four transmural sections approximately 6 mm thick. Each electrode was localized precisely as to its exact insertion site and the direction at which it entered the myocardium. The outline of each section was traced, showing the location of each recording site (Fig 1⇓). Plunge-needle electrodes that lay along the plane of sectioning were represented on sections both apical and basal to the plane of sectioning. The tracings were enlarged for later 3D construction of isochronal maps as described below.
Analysis of Electrograms and Construction of Isochronal Maps
Electrogram data were analyzed off-line by use of a Micro-VAX computer (Digital Equipment Corp) with interactive color graphics. Details of the mapping system were given previously.20 Initially, the tape containing the electrogram data was played back, and the accompanying surface lead II tracing was reviewed to find the beat(s) of interest (eg, sinus beat, PVC, or VT). Electrograms were displayed, and the activation times, assigned by the computer on the basis of a peak criterion,21 were reviewed and manually reassigned if required. An amplitude threshold of 0.25 mV was considered indicative of activation of tissue by the depolarizing wavefront.17 18 19 After review of all bipolar electrograms, the activation times for all sites were printed and assigned to their respective intramyocardial location indicated by the detailed 3D localization described previously. The 3D isochronal maps were then constructed in 10-ms increments.
Serum Electrolytes and Arterial Blood Gases
Peripheral venous blood samples for measurement of serum potassium (K+) and magnesium (Mg2+) were obtained from conscious HF rabbits within 3 weeks of the mapping study and were within the normal range for all rabbits. During all 3D mapping studies, serial arterial blood samples were obtained for arterial blood gas and K+ measurements. Any hypokalemia before mapping was corrected with intravenous KCl so that mapping was always performed in rabbits with serum values of K+ within the normal range.
After mapping analysis, transmural fixed tissue samples (up to 1.5×1.5×1.5 cm) from selected areas of interest from failing hearts (sites of initiation of ectopic beats and other sites) and control hearts were dehydrated, embedded in paraffin, and sectioned at a thickness of 5 μm. Transmural tissue sections were stained with hematoxylin and eosin or Masson’s trichrome stain and examined by light microscopy.
Holter monitors were generously donated by Scole Engineering and Delmar-Avionics.
Data Analysis and Statistics
The total activation time for each beat was the difference in activation time between the earliest and latest sites of activity. The mechanism of a particular beat was defined as reentrant when (1) there was continuous depolarization from the preceding beat, (2) the site of initiation of a premature beat was adjacent to the site of termination of the preceding beat, and (3) the conduction velocity of the activation wavefront from the site of termination of the preceding beat to the site of initiation of the ectopic beat was similar to the conduction velocity of the terminal portion of the activation wavefront of the preceding beat.17 18 19 A mechanism was defined as nonreentrant when the site of initiation of a premature beat was remote from the site of termination of the preceding beat with no intervening depolarizations, despite multiple intermediate recording sites.17 18 19 Data are expressed as mean±SEM. Comparisons of values for LVEDD, LVESD, FS, hemodynamics, total activation time, and coupling interval data were done by Student’s t test (for paired or unpaired data), and a probability value of <.05 was considered indicative of a significant difference.
Induction of Aortic Insufficiency and Stenosis
Perforation of the aortic valve with a beveled catheter resulted in an increase in the aortic pulse pressure from 16±1 to 30±3 mm Hg (P=.01). Color-flow Doppler echocardiography performed immediately afterward demonstrated that the degree of aortic insufficiency was severe in all four rabbits.
At 19±4 months after aortic constriction, color-flow Doppler echocardiography demonstrated severe aortic regurgitation in all four HF rabbits, and the aortic regurgitant fraction averaged 44±14%.
LV Chamber Dimension and Systolic Function
Over the course of 20±2 months, there was a progressive increase in LV dimensions and a decrease in LV systolic function (Fig 2⇓). LVEDD increased from 1.26±0.07 to 2.07±0.03 cm (P=.003); LVESD increased from 0.74±0.06 to 1.62±0.02 cm (P=.001). Sixty-five percent of the increase in LVEDD and 59% of the increase in LVESD occurred after the first 3 months. The most rapid increase in LV size occurred in HF rabbit 4, in which LVEDD increased from 1.08 to 1.96 cm and LVESD increased from 0.57 to 1.47 cm within 3 months (85% and 83% of the total changes, respectively). After the first 3 months, there was little additional chamber enlargement in this rabbit.
FS decreased from 0.42±0.02 to 0.21±0.02 (P=.004). Sixty percent of this change occurred in the first 3 months. The most severe depression of LV systolic function occurred in HF rabbit 2, in which FS decreased from 0.39 to 0.17 within 15 months (Fig 2⇑). The most rapid change in LV function was in HF rabbit 4, in which FS decreased from 0.47 to 0.25 within 3 months.
None of the rabbits demonstrated ventricular ectopy at baseline. After aortic regurgitation and aortic constriction, Holter monitoring in conscious rabbits demonstrated a progressive increase in the frequency and complexity of ventricular ectopic activity with the development of PVCs (usually one to three distinct morphologies, which were consistent from Holter to Holter, for each rabbit), couplets, and runs of monomorphic and polymorphic VTs (Figs 3⇓ and 4⇓). Frequent (>100 per day) PVCs occurred consistently in HF rabbits 1, 3, and 4 after 15 to 18 months. More frequent PVCs (>1000 per day) occurred in HF rabbits 1 and 4 after 21 months. HF rabbit 2 never developed more than 30 PVCs per day (Fig 4⇓, top).
Although VT occurred as early as after 6 months in HF rabbits 1 and 3 (Fig 4⇑, bottom), rabbits 1, 3, and 4 consistently demonstrated spontaneous runs of VT after 15 to 18 months. The length of runs of VT could vary considerably from day to day. The maximal length of VT for HF rabbits 1, 3, and 4 was 8, 26, and 22 beats, respectively (Figs 3⇑ and 4⇑). HF rabbit 2 never demonstrated VT. The Table⇓ lists the maximal frequency of ectopy over 1 hour of Holter monitoring.
Comparison of the Holter data (Fig 4⇑) with the echocardiographic data (Fig 2⇑) revealed that development of VT in HF rabbits 1, 3, and 4 at 15 months occurred when FS had decreased to 0.27±0.02 and LVEDD and LVESD had increased to 2.07±0.04 and 1.50±0.04 cm, respectively. However, HF rabbit 2 did not demonstrate VT after 15 months, despite a comparable increase in LV chamber dimensions and a greater decrease in FS (to 0.17). Thus, as in humans,4 development of VT did not appear to be related only to the level of cardiac enlargement or LV dysfunction.
Mapping of Spontaneously Occurring Ventricular Arrhythmias
Three-dimensional mapping of spontaneously occurring ventricular arrhythmias from 232 intramural sites was performed for up to 60 consecutive minutes. Three HF rabbits (1, 3, and 4) demonstrated frequent isolated PVCs, two demonstrated runs of nonsustained VT up to 4 beats long (the Table⇑), while one (rabbit 2) exhibited no ventricular ectopy during mapping. The frequency, degree of complexity, and morphology of the ventricular ectopy exhibited during mapping were comparable to those observed during previous Holter monitor recordings in conscious rabbits (the Table⇑ and Fig 3⇑). Mapping of the four control rabbits for up to 60 consecutive minutes demonstrated no spontaneous ectopy.
A total of 98 representative mapped beats from HF rabbits (42 sinus beats, 19 isolated PVCs, 18 beats of couplets, and 19 beats of VT from six runs of nonsustained VT) and 8 sinus beats from control rabbits were analyzed on the basis of a review of >24 000 electrograms obtained from intramural sites throughout the left and right ventricles and the interventricular septum. The mechanism of arrhythmia was defined for all the ventricular ectopic beats.
Sinus Rhythm in Control Rabbits
Sinus beats in the control rabbits initiated in the interventricular septum and spread rapidly to the apex and base and from endocardium to epicardium. The coupling interval of sinus beats averaged 290±11 ms, and the total activation time averaged 26±1 ms. Fig 5⇓ (left) gives an example of the 3D activation of a control sinus beat. Initiation occurred in the interventricular septum in level IV (asterisk). The depolarizing wavefront spread rapidly throughout the heart, terminating in the epicardium of the anterior septum in level I (the plus sign) with a total activation time of 31 ms.
Sinus Rhythm in Rabbits With Heart Failure
Sinus rhythm in HF rabbits initiated in the septum and spread rapidly throughout the heart, with a coupling interval of 316±9 ms (P=.12 versus control rabbits) and a total activation time of 28±1 ms (n=8). The total activation times in the three rabbits that demonstrated ventricular ectopy (27±1 ms, n=6) were comparable to the time in rabbit 2 (29 ms, n=2), which did not demonstrate ventricular ectopy during mapping, and to that in the four control rabbits (26±1 ms, n=8). Fig 5⇑ (middle) shows an example of the activation of a sinus beat in the failing rabbit heart. The sinus beat (NS) initiated in the septum in level II and was followed by early breakthrough in the lateral LV (10-ms isochrones in levels I through IV). Activation then spread rapidly throughout the heart, terminating in the epicardium of the right ventricle in level II, with a total activation time of 29 ms.
Sinus beats preceding PVCs demonstrated activation sequences and total activation times (26±1 ms) that were identical to sinus beats not preceding PVCs or VT. All PVCs initiated in the subendocardium by a nonreentrant mechanism, based on the lack of intervening electrical activity (225±7 ms) between the termination of the preceding sinus beat and the initiation of the PVC, despite the presence of multiple intervening electrode sites. Six PVCs (32%) initiated in the left ventricle, 1 (5%) in the right ventricle, and 12 (63%) in the interventricular septum. PVCs initiated with a coupling interval of 253±18 ms (n=19) and conducted with a total activation time of 43±2 ms.
An example is shown in Fig 5⇑ (right). The sinus beat (NS) initiated in the septum (level II) and spread rapidly throughout the heart, with an activation sequence identical to that of a sinus beat not preceding a PVC or VT (Fig 5⇑, middle). After termination of the sinus beat in the basal portion of the right ventricle (level II, the plus sign) with a total activation time of 29 ms, there was no electrical activation recorded anywhere in the heart for 267 ms, after which the PVC initiated by a nonreentrant mechanism at a subendocardial site in the endocardium (level IV). This is shown in further detail in Fig 6⇓. After termination of the sinus beat at site A, there was no electrical activity at sites B through O or at any other recording site for 267 ms, after which the PVC initiated at a distant site P by a nonreentrant mechanism.
Couplets and Nonsustained VT
Sinus beats preceding couplets and runs of VT exhibited similar activation sequences and identical total activation times (28±1 ms) to those of sinus beats not preceding PVCs or VT (28±1 ms). Beats of monomorphic and polymorphic couplets all initiated in the subendocardium by a nonreentrant mechanism. The coupling interval of the initiating beat of ventricular couplets averaged 291±3 ms; the total activation of the beats of couplets averaged 41±2 ms.
Runs of VT that were mapped included four runs of 3-beat monomorphic VT from HF rabbit 3 and two runs of polymorphic VT (one 3-beat run and one 4-beat run) from HF rabbit 1. Nonsustained VT initiated at a coupling interval of 300±2 ms and was maintained at a coupling interval of 257±14 ms (P<.01). All beats of VT initiated in the subendocardium by a nonreentrant mechanism.
Fig 7⇓ shows 3D activation maps of a sinus beat and 3 beats (X1 through X3) of a 3-beat run of nonsustained VT. The sinus beat (NS) initiated in the septum in level II and spread rapidly, with an activation sequence similar to that of the isolated sinus beat and the sinus beat preceding the PVC in Fig 5⇑. After termination of the sinus beat in the RV epicardium in level II (the plus sign), there was no electrical activity recorded anywhere in the heart for 270 ms, after which the first beat of VT, X1, initiated by a nonreentrant mechanism at the same apical endocardial site in level IV at which the PVC in Fig 5⇑ initiated. After initiation, activation of X1 was rapid, terminating at an epicardial site in level I with a total activation time of 37 ms. Beats X2 and X3 initiated at the same apical subendocardial site in level IV by a nonreentrant mechanism. The total activation time of each of these beats was 46 ms.
The activation sequences of X2 and X3 were identical and resembled the activation of X1. The only difference was that X1 demonstrated earlier activation in the anterior regions of levels I through III (20- to 40-ms isochrones), most likely from the breakthrough of simultaneous sinus activation (arising in level II) that resulted in X1 (and the PVC in Fig 5⇑) being a fusion beat. This accounted for the slight difference in QRS morphology of the PVC and beat X1 versus beats X2 and X3.
As Fig 8⇓ shows, X2 and X3 initiated by a nonreentrant mechanism, with no intervening activity between the termination of X1 and X2 (site A) and the initiation of X2 or X3 (site M), despite the presence of multiple intervening electrode recording sites (eg, sites B through L). In addition, once initiation at site M occurred, there was rapid (within 45 ms) activation to adjacent sites B through L.
To further define factors responsible for the development of monomorphic VT, analysis was performed of two PVCs, three couplets, and three runs of monomorphic VT from rabbit 3, all of which demonstrated similar morphology and initiated at the same site. As Fig 9⇓ shows, isolated PVCs (P1 and P2) could vary in their total activation times and coupling intervals. The initiating beats of couplets C1 through C3 all demonstrated total activation times and coupling intervals similar to those of PVC P1. The second beats (X2) of all three couplets demonstrated total activation times comparable to those of the first beats (X1), but they occurred at a shorter coupling interval. The initiating two beats (X1 and X2) of the three runs of VT demonstrated coupling intervals and total activation times similar to those of the couplets. The terminal beats (X3) of the three runs of VT were similar to the preceding X2 beats. Thus, PVCs, beats of monomorphic couplets, and runs of VT could demonstrate nearly identical activation sequences (Figs 5⇑ and 7⇑) and comparable total activation times and coupling intervals. Therefore, the difference among PVCs, monomorphic couplets, and monomorphic VT appeared to be the number of times a nonreentrant site fired rather than any disparities in activation sequence, conduction delay, or degree of prematurity of the initial ectopic beats.
Seven beats of polymorphic VT were mapped in HF rabbit 1. All beats arose in the subendocardium by a nonreentrant mechanism. Fig 10⇓ shows 3D maps of a sinus beat (NS) followed by the 4 beats (X1 through X4) of a 4-beat run of polymorphic VT. The sinus beat originated in the septum in level I and spread rapidly through the heart, with an activation sequence and total activation time (25 ms) identical to those of sinus beats not preceding PVCs or VT in this rabbit. After termination of the sinus beat at a posterior epicardial site in level I, there was no electrical activity anywhere in the heart for 286 ms, after which the first beat of VT, X1, initiated at a septal site in level III by a nonreentrant mechanism. Beats X2 through X4 initiated by nonreentrant mechanisms but at different sites (asterisk).
As Fig 11⇓ shows, X3 initiated by a nonreentrant mechanism, with no intervening electrical activity between the termination of X2 (site A) and the initiation of X3 (site X, arrow), despite the presence of multiple immediately adjacent (0.5 to 4 mm) recording sites (sites B through W). Once X3 initiated at site X, there was very rapid (≤20 ms) activation to adjacent sites B through X. The fourth beat of VT, X4, initiated by a nonreentrant mechanism at septal site Y.
As Fig 12⇓ shows, three of the four sites of initiation for X1 through X4 (A through C) were also sites of initiation of isolated PVCs or beats of couplets and demonstrated comparable total activation times and coupling intervals. Thus, the factor distinguishing between multiform PVCs and polymorphic couplets or VT appears to be the number of consecutive firings of the different nonreentrant sites. Total activation times for all ectopic beats mapped never exceeded 58 ms. The slowest conduction velocity, 20 cm/s, was found during a beat of nonsustained VT in an HF rabbit. There was no evidence of late activation (even <0.25 mV in amplitude) or any slow conduction at nonreentrant initiation sites or at immediately adjacent recording sites located 0.5 to 5 mm away in the left and right ventricles (see Fig 11⇑) and 0.5 to 6 mm away in the septum, making the presence of microreentry very unlikely.
All four failing hearts demonstrated marked diffuse histological alterations, including myocardial fiber hypertrophy and degenerative changes, including myofibrillar loss and vacuolization and interstitial fibrosis (Fig 13⇓, middle and bottom); these alterations were not evident in the control hearts (Fig 13⇓, top).
Histological sections from sites of nonreentrant initiation (n=14) from HF rabbits 1, 3, and 4 were examined in detail and compared with sections from regions in these hearts subtending recording sites (n=224) at which nonreentrant initiation did not occur. They were also compared with sections subtending recording sites (n=40) from the heart of HF rabbit 2, which never demonstrated VT by Holter recording or during mapping, and with sections from control hearts. His-Purkinje tissue was not evident in sections from sites of nonreentrant activation, although this does not rule out a spatial relation between Purkinje tissue and nonreentrant initiation. Tissue from regions in the vicinity of nonreentrant initiation sites demonstrated areas of patchy interstitial and replacement fibrosis, predominantly in the subendocardium. The extent of local fibrosis varied from moderate (Fig 14⇓, rows 1 and 2, right) to minimal (row 3, right) and was never observed to occupy more than 15% of the transmural wall thickness. In HF rabbits that demonstrated PVCs and VT (rabbits 1, 3, and 4), myocardium from regions in which nonreentrant initiation did not occur also demonstrated patchy interstitial and replacement fibrosis, primarily in the subendocardium (Fig 14⇓, bottom left). The distribution and severity of the fibrosis were comparable to those in the vicinity of nonreentrant initiation sites and sometimes were even more extensive (Fig 14⇓, bottom left). Surprisingly, the most extensive diffuse interstitial fibrosis was noted in HF rabbit 2 (Fig 14⇓, bottom right), which never demonstrated spontaneous VT. Thus, patchy interstitial fibrosis was present in the failing hearts, but its presence or degree did not appear to be related to nonreentrant sites of initiation or to the extent of conduction delay.
Three-dimensional mapping of spontaneously occurring ventricular ectopy in rabbits with severe nonischemic heart failure induced by combined aortic regurgitation and aortic stenosis has demonstrated that PVCs, couplets, and runs of VT all initiate by a nonreentrant mechanism arising in the subendocardium. Monomorphic and polymorphic VTs are due to repetitive nonreentrant activation from the same sites that are responsible for isolated PVCs.
Determination of a nonreentrant mechanism for the initiation of PVCs and VT was based on the absence of any electrical activity between the termination of the preceding beat and the initiation of the next, despite the presence of multiple intervening intramural recording sites.17 18 19 Mapping from the endocardial and/or epicardial surfaces would have left a large portion of the heart unmapped and would have precluded determination of the mechanism of arrhythmia. In addition, mapping from 232 sites in the rabbit heart, which is approximately 1/12 the size of the canine heart, would be comparable to mapping from >2700 sites in the canine heart. Thus, the enhanced resolution possible with studies in the rabbit heart further support the concept that nonreentrant mechanisms are critical in the development of arrhythmias in this model of nonischemic cardiomyopathy.
Bundle-branch reentry can initiate focally and has been found to underlie spontaneous monomorphic VT in approximately 6% of patients with cardiomyopathy.22 23 Although there were no recordings of His-bundle activation in the present study to definitively exclude this possibility, the finding that focal initiation occurred at the same site as the earliest sites during sinus rhythm in only one isolated PVC makes bundle-branch reentry an unlikely mechanism.
Although the possibility of microreentry cannot be definitively excluded by results of the present study, it is extremely unlikely for the following reasons. First, there was no intervening electrical activity between termination and initiation sites, which often were at opposite ends of the heart (see Figs 6⇑ and 8⇑). Second, on the basis of the VT cycle lengths, a microreentrant circuit small enough to escape detection by mapping with the resolution of that in the present study would have required a conduction velocity more than an order of magnitude slower than the slowest velocity measured, even during VT. Third, there was no evidence of any slow conduction in and around initiation sites (see Fig 11⇑). Even with microreentrant circuits as small as those demonstrated in the chronically infarcted canine epicardium in vitro by high-density grid electrode mapping (with an interelectrode distance of 350 μm),24 there was evidence of conduction delay on the order of 100 ms between sites 1 to 2 cm apart. That was never seen in the present study. In fact, no ventricular ectopic beat demonstrated a total activation time exceeding 58 ms.
Nature of Nonreentrant Mechanisms
The nature of the nonreentrant mechanism is unknown but may be caused by triggered activity arising from delayed afterdepolarizations (DADs), early afterdepolarizations,25 or altered automaticity. The repetitive firing from single subendocardial sites with decreasing coupling intervals is consistent with experimental studies of triggered activity.25 DADs and triggered activity have been demonstrated in hypertrophied myocardium in vitro26 and in human ventricular myocardium from patients with ischemic cardiomyopathy.27 Whether DADs could occur in subendocardial myocytes or Purkinje cells by turning on of the transient inward current (Iti)28 remains to be determined. However, increased levels of intracellular calcium,25 which are critical for the development of Iti and in turn DADs, have been demonstrated in myocytes from experimental animal preparations of nonischemic heart failure29 30 and in myocytes from hearts of patients with dilated cardiomyopathy.31 32
Myocardium from hypertrophied33 or failing34 hearts demonstrates prolongation of action potential duration, which could potentially lead to early afterdepolarizations.35 However, the occurrence of spontaneous arrhythmias at short coupling intervals makes this a less likely mechanism. Enhanced automaticity would also be unlikely in light of the short coupling intervals of ectopic activity observed in the present study and the degree of irregularity (Figs 10⇑ and 12⇑).
Although there was no evidence of slow conduction or the development of reentry in the genesis of spontaneous ventricular arrhythmias in this preparation, reentry could potentially contribute to the development of nonsustained and/or sustained VT. The presence of marked interstitial fibrosis such as that in HF rabbit 2 (Fig 13⇑), while insufficient to lead to conduction delay during sinus rhythm, could, in the setting of premature stimulation, lead to functional conduction delay caused by altered anisotropic conduction36 37 and the development of reentry. Such a finding has been demonstrated in the human heart in some patients with idiopathic dilated cardiomyopathy undergoing localized epicardial mapping at the time of heart transplantation.38 Furthermore, the development of VF from VT may also involve reentry because intramural reentry appears to be the common mechanism for VF in a number of pathological settings.18 19
Animal Preparation of Nonischemic Cardiomyopathy
The animal preparation of combined aortic regurgitation and aortic stenosis demonstrates both volume and pressure overload of the left ventricle13 and leads to a greater degree of LV hypertrophy than either aortic regurgitation or aortic stenosis alone.39 After 4 months, rabbits demonstrate a moderate decrease in LV systolic function, along with a decrease in β-adrenergic responsiveness resulting from a decrease in β-receptor number.13 In the present study, after 15 to 26 months, there was severe depression of LV function along with marked histological alterations and spontaneously occurring ventricular arrhythmias.
Nonetheless, the frequency and complexity of the spontaneous ventricular arrhythmias did not appear to be related only to the degree of LV dysfunction. This was also found to be the case in dogs with ischemic cardiomyopathy induced by multiple intracoronary microembolizations10 and in humans.4 These findings suggest a potential role for neurohumoral mechanisms,2 hemodynamic factors such as LV stretch,40 and other factors yet to be determined. In addition, the presence and complexity of nonreentrant ventricular arrhythmias were not related only to the presence of interstitial fibrosis.
The delineation of the mechanism of spontaneously occurring ventricular arrhythmias in nonischemic heart failure was made possible by the development of a model demonstrating frequent ventricular ectopy and by the use of a computerized mapping system with a large data storage capacity. Three of four rabbits were studied at the time when serial Holter monitoring demonstrated ventricular ectopic activity that was frequent enough that 3D mapping with continuous data recording from 232 simultaneous sites for up to 60 minutes would yield occasional PVCs, couplets, and runs of VT. The ventricular ectopic activity that was mapped demonstrated a frequency, level of complexity, and QRS morphology comparable to those of spontaneous rhythms recorded by Holter monitoring in conscious rabbits (Fig 4⇑). These spontaneously occurring arrhythmias were not induced by plunge electrode placement, as evidenced by the complete lack of ectopic activity during recording for up to 60 minutes in the four control rabbit hearts. Three-dimensional mapping with specially designed electrodes has been validated extensively in a number of animal species and in a variety of pathological states8 17 18 19 41 and shown not to alter activation during sinus rhythm or the inducibility of ventricular arrhythmia by programmed stimulation.
Although mapping was performed in only four rabbits with heart failure, the extensive characterization of LV function and arrhythmia development over a period of 15 to 27 months, the consistent findings between mapped beats and spontaneously occurring ectopic beats in conscious rabbits, the extensive number of spontaneously occurring sinus and ectopic beats mapped (representing analysis of >24 000 individual electrograms), and the extraordinary consistency with which every mapped ectopic beat demonstrated both nonreentrant initiation and the absence of significant conduction delay suggest that the findings of this study are representative of this nonischemic heart failure model.
The findings of nonreentrant initiation of spontaneous PVCs and VT in this rabbit model of nonischemic heart failure are very similar to the recent findings of focal initiation of spontaneously occurring PVCs and VT in dogs with ischemic cardiomyopathy induced by multiple intracoronary embolizations.8 These two studies demonstrate that nonreentrant mechanisms underlie VT in the setting of both ischemic and nonischemic heart failure and that slow conduction and the development of reentry play minor roles, if any, in the genesis of spontaneous ventricular ectopy in these settings. The contribution of reentry and nonreentrant mechanisms to the development of sustained VT induced by programmed electrical stimulation in the setting of nonischemic heart failure was not assessed in the present study and remains to be determined. Sustained ventricular arrhythmias are inducible in up to 80% of patients with nonischemic heart failure and a history of sustained VT42 and 28% of those with a history of nonsustained VT.43 Thus, it is likely that sustained VT may be induced in this preparation of nonischemic heart failure that demonstrates spontaneously occurring VT (up to 26 beats long, Fig 3⇑) in the setting of severe LV systolic dysfunction and marked pathological alterations that resemble human nonischemic cardiomyopathy.
Hearts from rabbits with end-stage nonischemic heart failure demonstrate marked histological alterations, including marked hypertrophy, degenerative cellular changes, and interstitial fibrosis. Although induction of the combination of LV pressure and volume overload does not directly cause alterations in myocytes such as those observed in cardiomyopathies (eg, induced by toxins), the result of this long-term overload is marked cellular alterations that are common to heart failure of a wide variety of causes, including nonischemic and idiopathic dilated cardiomyopathy in humans.11 12 Thus, the pathological alterations in this model, the marked systolic dysfunction, and the spontaneous ventricular arrhythmias are similar to those associated with nonischemic heart failure in man.
The implications of the study findings are multiple. The role of nonreentrant mechanisms in the development of VT in the setting of nonischemic heart failure could explain the lack of therapeutic benefit to patients treated with antiarrhythmic agents that alter conduction and the development of reentry.5 The facts that the response to programmed electrical stimulation is a poor predictor of efficacy of antiarrhythmic therapy44 45 and that late potentials occur infrequently in patients with idiopathic dilated cardiomyopathy46 further suggest a potential role of nonreentrant mechanisms in the development of sustained VT in this setting. Recent 3D intraoperative mapping studies in patients with ischemic heart failure demonstrated that inducible sustained monomorphic VT is due to a focal mechanism in 50% of cases,47 further supporting a potential role for nonreentrant mechanisms in the failing human heart.
Spontaneous ventricular arrhythmias in a rabbit model of nonischemic heart failure are due to nonreentrant mechanisms. Approaches to the treatment of ventricular arrhythmias in the human heart with nonischemic heart failure should be directed at such mechanisms.
Work from the Washington University School of Medicine Laboratory was supported in part by NIH grant HL-46929. I would like to thank William Petty, Teng-Xian Liu, and Jerome Peirick for technical assistance; Ava Ysaguirre for preparation of the manuscript; Dr Jeffrey Saffitz for assistance with pathological analysis; Dr Kathryn A. Yamada for review of the manuscript; and Dr Peter B. Corr for invaluable support and advice throughout the project.
Reprint requests to Steven M. Pogwizd, MD, Cardiovascular Division, Box 8086, Washington University School of Medicine, 660 S Euclid Ave, St Louis, MO 63110.
- Received November 14, 1994.
- Revision received January 23, 1995.
- Accepted February 10, 1995.
- Copyright © 1995 by American Heart Association
Packer M. Sudden unexpected death in patients with congestive heart failure: a second frontier. Circulation. 1985;72:681-685.
Pogwizd SM. Focal mechanisms underlying ventricular tachycardia in a model of ischemic cardiomyopathy. Circulation. 1994;90:1441-1458.
Sabbah HN, Stein PD, Kono T, Gheorghiade M, Levine TB, Jafri S, Hawkins ET, Goldstein S. A canine model of chronic failure produced by multiple sequential coronary microembolizations. Am J Physiol. 1991;260(Heart Circ Physiol 29):H1379-H1384.
Sabbah HN, Goldberg AD, Schoels W, Kono T, Webb C, Brachmann J, Goldstein S. Spontaneous and inducible ventricular arrhythmias in a canine model of chronic heart failure: relation to haemodynamics and sympathoadrenergic activation. Eur Heart J. 1992;13:1562-1572.
Unverferth DV, Fetters JK, Unverferth BJ, Leier CV, Magorien RD, Arn AR, Baker RB. Human myocardial histologic characteristics in congestive heart failure. Circulation. 1983;68:1194-1200.
Scott TM, Hackel DB. Heart. In: Kissane JM, ed. Anderson’s Pathology. 8th ed. St Louis, Mo: CV Mosby Co; 1985:560-662.
Gilson N, El Houda Bouanani N, Corsin A, Crozatier B. Left ventricular function and β-adrenoceptors in rabbit failing heart.Am J Physiol. 1990;258(Heart Circ Physiol 27):H634-H641.
Sahn DJ, DeMaria A, Kisslo J, Weyman A, for the Committee on M-Mode Standardization of the American Society of Echocardiography. Recommendation regarding quantitation in M-mode echocardiography: results of a survey of echocardiographic measurements. Circulation. 1978;58:1072-1083.
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. Mechanisms underlying the development of ventricular fibrillation during early myocardial ischemia. Circ Res. 1990;66:672-695.
Pogwizd SM, Corr PB. Electrophysiologic mechanisms underlying arrhythmias due to reperfusion of ischemic myocardium. Circulation. 1987;76:404-426.
Witkowski FX, Corr PB. An automated simultaneous transmural cardiac mapping system. Am J Physiol. 1984;247(Heart Circ Physiol 16):H661-H668.
Caceres J, Jazayeri M, McKinnie J, Avitall B, Denker ST, Tchou P, Akhtar M. Sustained bundle branch reentry as a mechanism of clinical tachycardia. Circulation. 1989;79:256-270.
Akhtar M, Damato AN, Batsford WP, Ruskin JN, Ogunkelu B, Vargas G. Demonstration of re-entry within the His-Purkinje system in man. Circulation. 1974;50:1150-1162.
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.
Wit AL, Rosen MR. Afterdepolarizations and triggered activity: distinction from automaticity as an arrhythmogenic mechanism. In: Fozzard HA, Haber E, Jennings RB, Katz AM, Morgan HE, eds. The Heart and Cardiovascular System: Scientific Foundations. 2nd ed. New York, NY: Raven Press; 1992:2113-2163.
Aronson RS. Afterpotentials and triggered activity in hypertrophied myocardium from rats with renal hypertension. Circ Res. 1981;48:720-727.
Perreault CL, Meuse AJ, Bentivegna LA, Morgan JP. Abnormal intracellular calcium handling in acute and chronic heart failure: role in systolic and diastolic dysfunction. Eur Heart J. 1990;11:8-21.
Wikman-Coffelt J, Stefenelli T, Wu ST, Parmley WW, Jasmin G. [Ca2+]i transients in the cardiomyopathic hamster heart. Circ Res. 1991;68:45-51.
Gwathmey JK, Copelas L, MacKinnon R, Schoen FJ, Feldman MD, Grossman W, Morgan JP. Abnormal intracellular calcium handling in myocardium from patients with end-stage heart failure. Circ Res. 1987;61:70-76.
Beuckelmann DJ, Näbauer M, Erdmann E. Intracellular calcium handling in isolated ventricular myocytes from patients with terminal heart failure. Circulation. 1992;85:1046-1055.
Aronson RS. Characteristics of action potentials of hypertrophied myocardium from rats with renal hypertension. Circ Res. 1980;47:443-454.
Spach MS, Kootsey JM, Sloan JD. Active modulation of electrical coupling between cardiac cells of the dog. Circ Res. 1982;51:347-362.
Spach MS, Miller WT III, Geselowitz DB, Barr RC, Kootsey JM, Johnson EA. The discontinuous nature of propagation in normal canine cardiac muscle. Circ Res. 1981;48:39-54.
Anderson KP, Walker R, Urie P, Ershler PR, Lux RL, Karwandee SV. Myocardial electrical propagation in patients with idiopathic dilated cardiomyopathy. J Clin Invest. 1993;92:122-140.
Hansen DE, Craig CS, Hondeghem LM. Stretch-induced arrhythmias in the isolated canine ventricle. Circulation. 1990;81:1094-1105.
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.
Poll DS, Marchlinski FE, Buxton AE, Doherty JU, Waxman HL, Josephson ME. Sustained ventricular tachycardia in patients with idiopathic dilated cardiomyopathy: electrophysiologic testing and lack of response to antiarrhythmic drug therapy. Circulation. 1984;70:451-456.
Middlekauff HR, Stevenson WG, Woo MA, Moser DK, Stevenson LW. Comparison of frequency of late potentials in idiopathic dilated cardiomyopathy and ischemic cardiomyopathy and advanced congestive heart failure and their usefulness in predicting sudden death. Am J Cardiol. 1990;66:1113-1117.
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.