Author + information
- Received September 27, 2002
- Revision received February 6, 2003
- Accepted February 13, 2003
- Published online July 16, 2003.
- Yuji Okuyama, MD, PhD*,
- Yasushi Miyauchi, MD*,
- Angela M Park, MD*,
- Akira Hamabe, MD*,
- Shengmei Zhou, MD*,
- Hideki Hayashi, MD*,
- Mizuho Miyauchi, MD*,
- Chikaya Omichi, MD*,
- Hui-Nam Pak, MD*,
- Lauren A Brodsky, BA*,
- William J Mandel, MD*,
- Michael C Fishbein, MD†,
- Hrayr S Karagueuzian, PhD* and
- Peng-Sheng Chen, MD*,* ()
- ↵*Reprint requests and correspondence:
Dr. Peng-Sheng Chen, Room 5342, CSMC, 8700 Beverly Boulevard, Los Angeles, California 90048-1865, USA.
Objectives The study examined the activations in the pulmonary veins (PVs) and the vein of Marshall (VOM) during atrial fibrillation (AF) in dogs with congestive heart failure (CHF).
Background The patterns of activation within the PVs and the VOM during AF in CHF are unclear.
Methods We induced CHF in nine dogs by rapid ventricular pacing. The patterns of activation during induced AF were studied one week after ceasing ventricular pacing.
Results The duration of induced AF averaged 80.7 ± 177.3 s. The termination of low-amplitude fractionated activity in the PVs preceded the termination of AF in 25 of 29 episodes. High-density mapping (1-mm resolution) showed that the PV was activated by a focal wave front independent of left atrial (LA) activation in 22 AF episodes. Frequent intra-PV conduction blocks and multiple wave fronts in the PVs were recorded during 10 AF episodes. Focal activations were observed within the VOM in 4 of 12 episodes of AF. Three atrial tachycardia (AT) episodes originated from a focus within a PV. Histological studies showed extensive fibrosis in the PVs and in the atria. The PVs in five normal dogs did not have focal or fractionated activity during induced AF.
Conclusions Atrial fibrillation in canine CHF is associated with independent focal activations in the PVs and the VOM, and with complex wave fronts within the PVs. The PVs may also serve as the origin of AT. These findings suggest that electrical and anatomical remodeling of the PVs and the VOM are important in the maintenance of AF and AT in dogs with CHF.
Congestive heart failure (CHF) increases the incidence of atrial tachyarrhythmias, resulting in a significant increase of morbidity and mortality (1). The mechanisms by which CHF increases atrial tachyarrhythmias have been explored by many investigators. In dogs with pacing-induced CHF, Li et al. (2)demonstrated that the duration of electrically induced atrial fibrillation (AF) is increased. Mapping stud- ies documented the presence of macroreentry and fibrillatory conduction in the atria (3). These proarrhythmic effects are associated with extensive atrial interstitial fibrosis andheterogeneous conduction delay. However, there was no study on the remodeling of muscle sleeve or muscle bundles surrounding the thoracic veins in CHF. Thoracic veins, such as the superior vena cava, the pulmonary veins (PVs) (4,5), and the vein of Marshall (VOM) (6,7), are known to have active muscle sleeves or muscle bundles that are electrically connected with the atrium. These muscles may contain node-like cells (8)and are capable of spontaneous diastolic depolarization (9), leading to automatic contraction (10,11). In addition to automaticity, the PV muscle cells are capable of rapid repetitive activity (12)and afterdepolarizations (13). The incidences of action potentials with an early afterdepolarization and of spontaneous tachycardia are much higher in dogs with chronic, rapid atrial pacing than in normal dogs (13). Many clinical studies showed that paroxysmal AF in humans may be initiated and maintained by rapid activity from the PVs (14), the VOM (15), and the superior vena cava (16). Our recent study (17)showed that the PVs and the VOM are also sources of rapid activation in a canine model of sustained AF. These data led us to hypothesize that significant proarrhythmic remodeling also occurs in the muscle sleeves of thoracic veins during CHF. The purpose of the present study was to test this hypothesis.
First surgery and rapid ventricular pacing to induce CHF
We performed sterile surgery in nine dogs (18 to 25 kg) to implant a pacemaker system for right ventricular pacing. Three days later, the pacemaker was programmed to pace at 150 beats/min for three days, at 200 beats/min for three days, and at 250 beats/min thereafter. When ventricular pacing resulted in CHF, the pacemaker was turned off.
Second surgery for mapping studies
In our experience, dogs at the peak of their CHF symptoms often do not tolerate general anesthesia and open chest surgery. To avoid premature death of the animals during the procedure, and to ensure a stable preparation, mapping studies were performed a week after the cessation of pacing. Five normal dogs were also studied for comparison. The right atrium and pulmonary artery pressure was measured by a balloon-tipped catheter in six CHF dogs and in three control dogs. The chests were opened via either a median sternotomy (n = 7) or a left lateral thoracotomy (n = 2). The pericardium was removed to expose the PVs. Bipolar hook electrodes were inserted into the right atrium and left atrium appendage for recording atrial electrograms and for burst pacing to induce atrial tachyarrhythmias.
Computerized mapping studies
We used a high-density computerized mapping system (Unemap, Uniservices) (18)for mapping studies. The electrodes were distributed over four patches; each patch records 420 bipolar electrograms. One patch was sutured to the right atrium lateral wall, one to the left atrium anterior wall, and one to the left atrium posterior wall. The fourth patch was placed manually on one of the PVs to map these veins one at a time. The VOM was mapped either by the electrodes on the left atrium posterior wall or when the fourth electrode patch was moved to map the left inferior pulmonary vein (LIPV).
Each recording consists of 8 s of data. In each dog, the left superior pulmonary vein (LSPV), LIPV, right superior pulmonary vein (RSPV), and right middle pulmonary vein (RMPV) were mapped in random order. The right inferior pulmonary vein (RIPV) was also mapped when a left thoracotomy was employed. Because of technical difficulties, the RIPV was not studied when median sternotomy was performed. The signals were filtered with a 0.05-Hz high-pass filter and were digitized at 1,000 samples/s with 12 bits of resolution.
Bipolar electrode mapping studies
After completing the computerized mapping studies, a pair of bipolar hook electrodes (5–6-mm interpolar distance) was inserted into each PV within 1 cm of the PV–left atrium junction in six CHF dogs. We recorded atrial electrograms and four PV electrograms simultaneously with a high-pass filter of 30 Hz and a low-pass filter of 500 Hz. The data were continuously digitized at >1,000 times/s with 16 bits of accuracy (Axon Instruments, Union City, California).
We analyzed 77 runs of atrial tachyarrhythmias with the computerized mapping system according to the methods reported previously (17,19). The atrial cycle length was calculated as the mean of all activation cycle lengths within one electrode patch. A focal activation is defined as an activation propagating centrifugally from a central site. Alternatively, if the distal end of the muscle sleeve was not included in the mapped region, an activation originating from the distal end of the muscle sleeve and propagating only in the direction toward the left atrium was also considered a focal activation.
After we concluded the mapping studies, we harvested the heart for histopathological studies. Data are presented as mean ± SD. Two-tailed Student ttests were used to compare the means between two groups. The null hypothesis was rejected if a p value was ≤0.05.
Induction of CHF
All nine dogs showed clinical manifestation of CHF after 24.8 ± 4.7 days of pacing and were in sinus rhythm at the beginning of the second surgery. The mean right atrium pressure in CHF dogs (3.0 ± 1.8 mm Hg) did not differ significantly from that of the controls (2.3 ± 0.6 mm Hg, p = 0.41), but the mean pulmonary artery diastolic pressure in CHF dogs (6.9 ± 2.5 mm Hg) was significantly higher than that of the controls (3.7 ± 0.6 mm Hg, p < 0.05). The mean heart weight of CHF dogs (274 ± 43 g) was significantly heavier that that of the controls (189 ± 10 g, p < 0.01).
Bipolar hook electrode mapping of AF: two types of activations in PVs
Thirty-one AF episodes and 17 atrial tachycardia (AT) episodes were induced and recorded with hook electrodes in six CHF dogs. The mean duration of AF in CHF dogs was 80.7 ± 177.3 s, which was significantly longer than that in the controls (6.1 ± 11.5 s, p < 0.05). All, except one, of the AF episodes terminated spontaneously. Sixteen AT episodes terminated spontaneously with a mean duration of 28.8 ± 48.8 s. One episode was terminated with overdrive pacing.
Three episodes of AT to AF transition were recorded. In all three episodes, the onset of fractionated activity within the PV (asterisks in Fig. 1) was followed by similar activities in the RSPV, LSPV, and LIPV. The occurrence of the fractionated activity was associated with a shortening of activation cycle length in the left atrium, leading to AF. This sequence of events was similar to that reported in human patients (20).
Changes of PV activation were also seen prior to AF termination. Figure 2shows a typical example of non-sustained AF, which converted to AT (arrows) before termination. During AF, there were two types of PV activations. The first type is a more regular activation pattern that had a longer activation cycle length (RSPV) than that in the left atrium. The second type is the fractionated electrograms registered in two of the PVs (asterisks in Fig. 2). Among all dogs studied, fractionated activity was observed in 29 of 31 AF episodes. The most apparent fractionated activity was seen in the LSPV (n = 11), LIPV (n = 4), RSPV (n = 3), and RMPV (n = 11). In the remaining two episodes, the PVs showed only the first (slower) type of activation without fractionated activity. The fractionated activity was found in only one vein in 2 dogs, in two veins of 1 dog, in three veins of 2 dogs, and in four veins of 1 dog. In 17 of 29 episodes that included one or more PVs with fractionated activity, slow activations were also registered in the other PVs. Termination of the fractionated activity was associated with an AF to AT transition, followed by AT termination in 25 of 29 episodes, and coincided with the termination of AF in the remaining 4 episodes.
The mean cycle length of the induced AT was 169.0 ± 30.9 ms. During AT, the activation cycle lengths of all the PVs were the same, and there was no fractionated activity.
High resolution mapping of AF: two types of activations in the PVs
Forty-four AF episodes and 33 AT episodes were induced and recorded with the computerized mapping methods in seven CHF dogs. The mean activation cycle length of the left atrium during AF (142.4 ± 12.0 ms) was shorter than that of the right atrium (158.3 ± 11.8 ms, p < 0.01). Just as we had observed two types of activation patterns in bipolar hook electrode mapping, we also observed two types of activation patterns in the PVs during AF.
Focal activation pattern
Focal activation in the PVs was observed in 22 of 44 AF runs. Among these 22 focal activations, 10 showed an activation propagating centrifugally from a central site, whereas 12 episodes showed activations originating from the distal end of the muscle sleeve and propagating only in the direction toward the left atrium. In the remaining 22 episodes, the mapped PV was activated by wave fronts from the left atrium. Figure 3shows a typical example of independent focal activation in the PV. Despite the longer cycle length, Figure 3Bshows that a focal source activated most of the PV. This wave front collided with a wave front that originated from the left atrium. Figure 3Cshows that the entire PV was activated by a focal wave front that originated in the PV, independent of the left atrium. In Figure 3D, the entire PV was activated by a wave front coming from the left atrium.
Complex activations in the PVs during AF
In 10 of the 22 AF episodes that included focal activations, complex activations were also observed. Figure 4depicts an example of complex activations. Dynamic display revealed conduction blocks and the coexistence of multiple wave fronts in the same PV. Sometimes these wave fronts did not seem to be related to one another, with some wave fronts activating as though they were from a focal source (propagating away from a focal site). However, some wave fronts could be compatible with reentry with slow conduction, or could be triggered by activity induced by a preceding wave front. Sustained and organized reentry was not observed, as the patterns of activation changed from beat to beat. The termination of AF was recorded in four episodes. In three of these four episodes, cessation of the complex activations in a PV was followed by focal activations in the PV and then termination of all atrial tachyarrhythmias (Fig. 5). On average, the time lapse between cessation of the complex activations to termination of all atrial tachyarrhythmias was 0.6 ± 0.2 s.
Focal activations within the VOM during AF
Figure 6shows that during sinus rhythm and during some AF activations (activations 1 and 2), the entire VOM was activated by a wave front that originated from the proximal VOM (the portion near the coronary sinus). The propagation velocity within the VOM was 1.1 m/s in sinus rhythm and 1.2 m/s in AF. However, during activations 3 through 7, the VOM was activated from an independent focus near electrode site e.Activation 5 showed a collision of wave fronts from site eand from the proximal VOM. Figure 6shows that the VOM is capable of independent focal activation. We registered VOM potentials in 12 of 44 episodes of AF. During 4 of 12 episodes, focal activations were identified within the VOM.
Activation patterns of AT
We induced 33 episodes of AT with a mean activation cycle length of 158.1 ± 12 ms. Focal activations in the PVs served as a source of AT in three episodes (Fig. 7). Seven AT runs had focal activation originating from the crista terminalis. In the remaining 23 AT episodes, all electrode patches recorded wave fronts that invaded the mapped region from the edge. This activation pattern indicates that either the earliest activation site was outside the mapped area, or that the electrode patches recorded a portion of the macroreentrant circuit. The VOM potential was recorded in 6 of 33 episodes; all 6 episodes showed passive activation of the VOM by atrial wave fronts, with intermittent conduction blocks (Fig. 7A, Panel c).
Activation patterns during AF in normal dogs
A total of 47 AF episodes but no AT episodes were induced and mapped in five normal dogs. Each episode of AF was short-lived and spontaneously terminated. In all the recordings of AF, the activation cycle lengths in the PVs (162.3 ± 39.3 ms) were slower than those in the left atrium (141.7 ± 9.8 ms, p < 0.05). The PVs were always activated passively and periodically by the wave fronts coming from the left atrium. Neither focal activation nor fractionated activity was observed in the PVs.
In CHF dogs, there was extensive interstitial fibrosis in the atria and near ventricular pacing sites (Fig. 8). Bundles of myofibers were separated by thick layers of fibrous tissue. The percentage of fibrosis in PVs was significantly greater in CHF dogs (6.4 ± 2.1%, n = 5) than in control dogs (0.5 ± 0.5%, n = 5, p < 0.01). There were no significant differences in the thickness of the PV muscle sleeves between CHF (600 ± 118 μm, n = 5) and control (780 ± 208 μm, n = 5) dogs sampled 3 mm away from the PV–left atrium junction. There were no significant differences in the amount of fibrosis in the PVs among CHF dogs with or without complex activity.
In this study we demonstrated that some AF episodes in CHF are characterized by focal activations in the PVs and the VOM, and by complex, fractionated wave fronts within the PVs. The termination of fractionated activity in the PVs preceded the termination of AF. The PVs may also serve as the origin of AT and play a role in the AT to AF transition. These findings suggest that the PVs and the VOM are important in the maintenance of AF and AT in dogs with CHF.
Fractionated activity and the mechanisms of AF in CHF
We found that the termination of the fractionated activity registered by hook electrodes within the PVs preceded the termination of AF in dogs with CHF. This fractionated activity is probably the same electrical phenomenon as the complex pattern of activation recorded by high-density mapping. The increased fibrosis and a heterogeneous change of conduction velocity might contribute to this complex activity.
A second possible mechanism is the triggered activity (21). It is known that the PV muscle cells are capable of rapid repetitive activity (12)and afterdepolarizations (13). Congestive heart failure is known to cause increased sympathetic activity (22)and an increased transient inward Na+-Ca2+exchanger current (23). These changes might result in the increased propensity for triggered arrhythmias.
It is also possible that these complex activities are due to a combined effect of reentry and triggered activity. An increased atrial activation rate could increase cellular Ca2+-loading (24)and facilitate the development of triggered activity (21). Triggered activity, in turn, might provide a source of rapid activity that facilitates reentry.
Focal activation in the thoracic veins
The focal activations in the thoracic veins are independent of electrical activity elsewhere in the atria. Transmural reentry might cause focal activity on the epicardial surface; however, the thickness of the PV muscle sleeves averaged 0.6 mm, which is probably too thin for a complete reentrant circuit (25). Furthermore, the VOM is a very small vessel. The muscle bundles associated with the VOM averaged 0.7 mm in diameter in humans (26)and roughly half that size in dogs (6). This small diameter makes the muscle bundles within the VOM highly unlikely to accommodate a reentrant circuit. Therefore, these focal activations within the thoracic veins are more likely to be due to non-reentrant mechanisms. Automaticity or triggered activity (11,13)might be the underlying mechanism of this focal activity.
Atrial tachycardia (AT)
In patients with preexcitation syndrome, successful surgical ablation of the accessory pathways prevented future occurrence of AF (27). These observations suggest that accessory pathway-mediated tachycardia is the trigger for AF. Effective accessory pathway ablation prevented tachycardia, thereby preventing AF. Other investigators (20)reported that fractionated PV activations preceded the transition from atrial flutter to AF. In our study we showed that PVs may serve as the origin of AT and play a role in the AT to AF transition. This is another example of tachycardia-induced AF via PV mediation.
Comparing PV activations in different models of AF
In an animal model of atrial pacing-induced sustained AF, there are persistent rapid focal discharges from the PV (18). The activation rates in the PV and in the VOM are higher than the activation rate in the left atrium (17). In contrast, the present study reports AF activations in a canine model of CHF with only unsustained AF being inducible. The activations in the PVs and in the VOM could be either faster or slower than in the left atrium. These differences could be due to differing processes of electrophysiological remodeling. The mRNA expression of the L-type Ca2+channel alpha-1c gene and Kv1.5 potassium channel gene were reduced in pacing-induced AF (28). No significant changes were found in the mRNA levels of the rapid Na+channel, the Na+/Ca2+exchanger, or the Kv4.2/4.3 channels responsible for I(to). In contrast, Li et al. (23)reported that CHF in dogs causes a lengthening of action potential duration and refractoriness, associated with selectively decreased atrial I(to), I(Ca), I(Ks), and increased Na+/Ca2+exchanger current, while leaving other currents unchanged. A differential cellular electrophysiological remodeling process caused by CHF and by rapid atrial pacing might be responsible for the different PV activations observed in these canine models of AF.
The findings in the present study suggest that future research in AF among patients with CHF should include thoracic vein activations in their study design. For example, a study on the pharmacological agents should consider the effects of those agents on the electrophysiology of the muscle sleeves and muscle bundles within the thoracic veins. By focusing on the activations in the thoracic veins, it is possible to better define antiarrhythmic drug action in patients with CHF, thereby improving patient care.
We mapped only the anterior portion of the PV. Therefore, the activation patterns in the posterior surface of the PV could not be determined. Because we did not perform transmembrane potential recordings, it was unclear whether automaticity or triggered activity was responsible for the focal activity in the PVs. Another limitation is that we did not perform PV and VOM ablation. Without those data, it is unclear whether PV and VOM ablation can lead to the termination or the prevention of reinduction of AF.
The Unemap mapping system was developed by Drs. Peter Hunter, David Bullivant, and David Budgett. We thank Adam Cates, PhD, of Guidant Inc., for donating the pacemaker. We also thank Nina Wang, Avile McCullen, Katherine Fu, and Elaine Lebowitz for their assistance.
☆ This study was supported by grants from Irvine Biomedical and Bayer Yakuhin, Ltd. (Y.O.); the Cardiac Arrhythmia Research Enhancement Support Group (CARES) (Y.M.); a NASPE Fellowship grant (S.Z.), an ECHO Foundation Award (H.S.K.), a Pauline and Harold Price Endowment (P.-S.C.), NIH grants HL71140, HL66389, P50-HL52319, AHA grants 0255937Y, 9950464N, UCTRDRP 11RT-0058, and the Ralph M. Parsons Foundation, Los Angeles, California.
- atrial fibrillation
- atrial tachycardia
- congestive heart failure
- left inferior pulmonary vein
- left superior pulmonary vein
- pulmonary vein
- right inferior pulmonary vein
- right middle pulmonary vein
- right superior pulmonary vein
- vein of Marshall
- Received September 27, 2002.
- Revision received February 6, 2003.
- Accepted February 13, 2003.
- American College of Cardiology Foundation
- Li D.,
- Fareh S.,
- Leung T.K.,
- Nattel S.
- Nathan H.,
- Eliakim M.
- Scherlag B.J.,
- Yeh B.K.,
- Robinson M.J.
- Doshi R.N.,
- Wu T.-J.,
- Yashima M.,
- et al.
- Chen Y.J.,
- Chen S.A.,
- Chen Y.C.,
- et al.
- Hwang C.,
- Wu T.-J.,
- Doshi R.N.,
- Peter C.T.,
- Chen P.-S.
- Tsai C.F.,
- Tai C.T.,
- Hsieh M.H.,
- et al.
- Wu T.-J.,
- Ong J.J.C.,
- Chang C.-M.,
- et al.
- Zhou S.,
- Chang C.-M.,
- Wu T.-J.,
- et al.
- Lee J.J.,
- Kamjoo K.,
- Hough D.,
- et al.
- Hasking G.J.,
- Esler M.D.,
- Jennings G.L.,
- Burton D.,
- Johns J.A.,
- Korner P.I.
- Li D.,
- Melnyk P.,
- Feng J.,
- et al.
- Sun H.,
- Chartier D.,
- Leblanc N.,
- Nattel S.
- Spach M.S.,
- Dolber P.C.
- Kim D.T.,
- Lai A.C.,
- Hwang C.,
- et al.
- Chen P.-S.,
- Pressley J.C.,
- Tang A.S.L.,
- Packer D.L.,
- Gallagher J.J.,
- Prystowsky E.N.