Author + information
- Received December 20, 2001
- Revision received March 20, 2002
- Accepted April 3, 2002
- Published online June 19, 2002.
- Steven M Markowitz, MD, FACC*,
- Richard F Brodman, MD, FACC†,
- Kenneth M Stein, MD, FACC*,
- Suneet Mittal, MD, FACC*,
- David J Slotwiner, MD*,
- Sei Iwai, MD*,
- Mithilesh K Das, MD* and
- Bruce B Lerman, MD, FACC*,* ()
- ↵*Reprint requests and correspondence:
Dr. Bruce B. Lerman, Division of Cardiology-Starr 4, The New York Hospital-Cornell Medical Center, 525 East 68th Street, New York, New York 10021, USA.
Objectives The purpose of this study was to define the anatomic distribution of electrically abnormal atrial tissue and mechanisms of atrial tachycardia (AT) after mitral valve (MV) surgery.
Background Atrial tachycardia is a well-recognized long-term complication of MV surgery. Because atrial incisions from repair of congenital heart defects provide a substrate for re-entrant arrhythmias in the late postoperative setting, we hypothesized that atriotomies or cannulation sites during MV surgery also contributed to postoperative arrhythmias.
Methods In 10 patients with prior MV surgery, electroanatomic maps were constructed of 11 tachycardias (6 right atrium [RA], 4 left atrium [LA] and 1 biatrial). Activation and voltage maps were used to identify areas of low voltage, double potentials and conduction block.
Results Lesions were present in the lateral wall of the RA (six of seven maps) and in the LA along the septum adjacent to the right pulmonary veins (four of five maps). In 8 of 10 patients, these findings corresponded to atrial incisions or cannulation sites. Arrhythmia mechanisms were identified for 9 of 11 tachycardias. A macro–re-entrant circuit was mapped in six cases, three involving lesions in the lateral wall of the RA and three involving the LA septum and right pulmonary veins. In three of these cases figure-of-eight re-entry was demonstrated, and in the other three a single macro–re-entrant circuit was observed. In three other cases, a focal origin was identified adjacent to abnormal tissue in the RA (two cases) or within a pulmonary vein (one case).
Conclusions Surgical incisions for MV surgery provide a substrate for atrial arrhythmias. Both macro–re-entrant and focal mechanisms contribute to AT after MV surgery.
Atrial tachyarrhythmias are common in the early and late convalescent periods after mitral valve (MV) surgery (1). Early postoperative arrhythmias are often attributed to the acute effects of cardiac surgery, such as pericarditis, atrial ischemia and surgical trauma, and late arrhythmias are ascribed to longstanding myopathic processes in the atria. However, a paucity of data exists regarding the arrhythmogenic potential of atrial incisions used for exposing the MV. It is well-established that atrial incisions for repair of congenital heart disease may provide a substrate for arrhythmias in the venous atrium (2–6). Analogous to the situation with congenital heart disease, it is plausible that lesional atrial tachycardia (AT) occurs after MV surgery.
Several approaches are commonly used to provide surgical access to the MV, including the right lateral left atriotomy, access through the interatrial groove, the trans-septal approach and the superior trans-septal approach (Fig. 1). The former two involve incisions in the left atrium (LA) only, whereas the latter two involve both right and left atrial incisions. The purpose of this study was to use detailed electroanatomic mapping to define the distribution of atrial lesions after different surgical approaches to the MV and identify the origin of AT after MV repair or replacement.
The study population consisted of 11 consecutive patients (age 62 ± 12 years; 6 women) who had prior MV repair or replacement and underwent electroanatomic mapping of AT. The clinical characteristics of these patients and the surgical approaches to the MV are described in Table 1. The interval from the most proximate MV surgery to the electrophysiologic study was 50 ± 67 months (range: 2 to 204 months).
Seven patients also had atrial fibrillation (AF) after MV surgery. Six of these seven patients had AF documented within two months of surgery, and three of these patients also had late AF (3, 11 and 17 years postoperatively, respectively). Six patients had at least one class I or class III antiarrhythmic drug failure (median: 1, range: 1 to 3).
At the time of the initial electrophysiologic study, all patients were off antiarrhythmic medications for at least five half-lives, except two patients with a history of AF who were taking amiodarone or sotalol at the time of the study.
Informed consent was obtained before each electrophysiologic study. Patients were locally anesthetized with 0.25% bupivacaine and sedated with midazolam and morphine. Diagnostic electrophysiology catheters included a 6F quadripolar catheter positioned at the His bundle, a 7F duodecapolar catheter around the tricuspid annulus and a 6F decapolar catheter in the coronary sinus. Left atrial mapping was performed through a trans-septal sheath.
Complete electroanatomic maps were obtained of the spontaneous clinical AT (eight patients; eight tachycardias) or induced AT (two patients; three tachycardias). In each case, electroanatomic data were collected during AT. In patients with bioprosthetic valves and MV repairs, mapping was performed in either the right atrium (RA) or LA, as guided by activation of the coronary sinus and right atrial catheters as well as P-wave morphology. In patients with mechanical MVs, mapping was limited to the RA to avoid potential complications related to the mechanical prosthesis.
The electroanatomic mapping system (CARTO, Biosense Webster, Diamond Bar, California) consists of a quadripolar catheter with a 4-mm distal electrode (Navistar, Biosense Webster) and a spatial reference patch on the patient’s back. A coronary sinus electrogram was used as a reference for local activation. Activation times were assigned based on the onset of bipolar electrograms, which were obtained with filter settings of 30 and 400 Hz. In addition to color-coded activation maps, isochronal maps were analyzed to identify the direction of wavefront propagation and areas of slow conduction.
Conduction block was inferred if there were adjacent regions with wavefront propagation in opposite directions, separated either by a line of double potentials (two components with a clear isoelectric interval) or dense isochrones (typically >100 ms difference in activation over <2 cm distance).
Voltage maps were displayed by manually adjusting the color range such that areas with bipolar electrograms ≤0.3 mV were designated red, and areas ≥0.4 mV were designated purple (Figs. 2 to 5). ⇓⇓⇓⇓This mapping technique emphasized areas of low voltage relative to surrounding tissue. Discrete regions of lower voltage compared to surrounding tissue were considered to represent either scar or areas of depressed conduction (“abnormal tissue” or “lesions”).
Entrainment mapping was performed from a “protected isthmus,” as identified on the electroanatomic map, by pacing 20 to 50 ms less than the tachycardia cycle length. Sites with concealed entrainment and a post-pacing interval ≤30 ms longer than the AT cycle length were considered to lie in protected zones within the tachycardia circuit. Entrainment mapping was performed in all cases of macro–re-entry in the RA. Entrainment mapping was not routinely performed with LA tachycardias to avoid termination of AT with rapid pacing.
In cases of macro–re-entry, linear lesions were created with sequential ablations to connect anatomical barriers that defined a critical isthmus. Focal tachycardias were targeted at the sites of earliest activation. Power was adjusted between 10 and 50 W and ablation was continued for 30 to 60 s at each site unless a rise in impedance occurred. Ablation was acutely successful if the tachycardia terminated during ablation and was not inducible with programmed stimulation.
Operative reports were obtained for all patients in this series and were reviewed by two authors (R. F. B. and S. M. M.). Surgeries were classified according to the surgical approach: 1) left atriotomy, 2) dissection in the interatrial groove or 3) superior trans-septal approach (Fig. 1). No patient in this series had the basic trans-septal approach using a right atriotomy followed by a separate atrial septal incision, although five patients had the superior trans-septal approach. In addition, note was made of venous cannulation sites and the use of retrograde cardioplegia.
Distribution of abnormal voltage
Electroanatomic maps were obtained during 11 episodes of AT (Table 2). Each patient underwent mapping of a single AT, with the exception of one patient who had maps of two tachycardias. Six tachycardias were mapped in the RA, four in the LA, and one in both atria. Of the seven RA maps, low voltage was present in the lateral or posterolateral wall in six patients. In all but one patient, these regions were contiguous with the superior vena cava (SVC), the inferior vena cava (IVC), or both. Other areas of low voltage included the interatrial septum (two cases) and the anterior RA between the tricuspid valve (TV) and the SVC, at the site of the RA appendage (one case). Of the six patients with lesions in the right lateral wall, three had left atriotomies (two through a standard left atriotomy and one through the interatrial groove), two had the superior trans-septal approach, and one had a left atriotomy followed by the superior trans-septal approach during reoperation.
In the LA, low voltage was present in the septum contiguous to the right pulmonary veins (four maps); two of these patients also had low voltage in the posterior wall. Of these four patients, two had the superior trans-septal approach and two had left atriotomies (one through a standard left atriotomy and one through the interatrial groove).
An example of simultaneous RA and LA maps is depicted in Figure 2. This patient, who had a superior trans-septal approach, demonstrates a region of low voltage involving the lateral wall of the RA, the superior RA anterior to the SVC, and the RA septum. In the LA, low voltage is present in the septum and extends to the dome of the LA.
Overall, seven patients demonstrated low voltage in regions known to have surgical incisions or cannulation sites. The other three patients had left atriotomies but did not undergo mapping of the LA (because of mechanical prostheses or RA origin). Five patients demonstrated low voltage in anatomical regions remote from the expected atriotomy or cannulation sites.
Five tachycardias originated in the RA and six originated in the LA. In two patients with LA tachycardia, electroanatomic mapping was limited to the RA because of mechanical mitral prostheses, and LA origins were inferred from activation patterns in the coronary sinus and RA. In both patients, passive activation of the RA was demonstrated through the low atrial septum (Patient 9) or Bachmann’s bundle (Patient 10). Therefore, mechanisms could be classified for nine tachycardias. Six were found to be macro–re-entrant and three were focal, as summarized in Figures 6 and 7. ⇓⇓
Complete macro–re-entrant circuits were identified in both the RA (three cases) and the LA (three cases), based on mapping 90 ± 8% of the tachycardia cycle length (range 79% to 99%) and demonstrating adjacent regions of early and late activation. In each case of macro–re-entry, anatomical barriers could be identified, involving at least one lesion.
In the RA, abnormal tissue in the lateral wall provided the substrate for re-entry. In two cases, figure-of-eight re-entry was present with a central isthmus between two lesions (Fig. 3). In each case, double potentials were present within or adjacent to the low voltage regions. One case of macro–re-entry in the RA involved a single circuit around a line of low voltage and conduction block in the lateral wall, with a zone of slow conduction in the superior RA adjacent to the SVC (Fig. 2). This patient had prior ablation of counterclockwise atrial flutter in the IVC-tricuspid isthmus, resulting in bidirectional block.
In three cases of macro–re-entry in the RA, rapid pacing was performed from a protected isthmus as defined by the electroanatomic map (Fig. 6). Pacing at these sites resulted in concealed entrainment in two cases and termination of the tachycardia in one case. In each case of concealed entrainment, the post-pacing interval was ≤10 ms longer than the tachycardia cycle length, confirming participation of these sites in the tachycardia circuit.
In the LA, abnormal tissue in the septum adjacent to the right pulmonary veins served as an anatomical barrier for reentry. In two patients, single re-entrant circuits were present around the septal lesion and right pulmonary veins (Fig. 4). Activation around the MV was passive, demonstrating bifurcation of a single wavefront at the annulus or fusion of clockwise and counterclockwise wavefronts. In each case, a protected isthmus was identified between the septal lesion/pulmonary veins and the mitral annulus. Figure-of-eight re-entry was demonstrated in one case (Patient 6; Fig. 5), involving a clockwise loop around an area of low voltage in the posterior wall and a counterclockwise loop around the mitral annulus, with a central isthmus between two areas of low voltage. A bystander corridor was identified on the electroanatomic map between the septal lesion and MV.
Two focal tachycardias originated in the RA adjacent to linear lesions characterized by low voltage, double potentials and dense isochrones. These lesions were present in the lateral wall between the vena cavae (Patient 5) or the anterior RA between the TV and SVC (Patient 4). The sites of earliest activation were 81 and 94 ms before the onset of the surface P-wave, respectively. One focal AT originated in the left inferior pulmonary vein.
Seven of 11 tachycardias were targeted for ablation (five macro–re-entrant and two focal). Reasons for not performing ablation in four patients included left atrial tachycardias in patients with mechanical MVs (two patients), a nonclinical tachycardia arising from a pulmonary vein (one patient), and recurrent AF (one patient). Acute success, defined as termination of the tachycardia and noninducibility, was achieved in five cases (three macro–re-entrant and two focal). Ablation of reentrant tachycardias was achieved by targeting a critical isthmus in the reentrant circuit, between two areas of scar in the RA or between the right pulmonary veins and the mitral annulus. In these cases, AT terminated after one to five applications of RF energy (36 to 97 s). Two focal tachycardias in the RA terminated after one and seven RF applications, respectively (33 and 203 s). In two cases of macro–re-entry (one RA and one LA), ablation failed to interrupt AT or prevent reinducibility. In these patients, post-hoc analysis indicated that ablation was performed in bystander regions.
In this study we demonstrated consistent patterns of low voltage and conduction block in both atria of patients with prior MV surgery. In the RA, most patients had abnormal tissue in the lateral or posterolateral wall, and in the LA abnormal tissue was present in the septum adjacent to the right pulmonary veins. Re-entrant tachycardias in the form of single circuits and figure-of-eight were demonstrated involving these lesions and other anatomical structures. Focal tachycardias were also demonstrated arising adjacent to abnormal tissue.
Atriotomy incisions versus atrial cardiomyopathy
Areas of low voltage and conduction block may be related to atriotomy scars or myopathic processes in the atria. In most patients, these abnormal areas corresponded to expected atriotomy locations or cannulation sites. The size and characteristics of abnormal tissue varied, with some patients demonstrating a narrow line of conduction block and others with broader areas of low voltage. In some patients, areas of low voltage were present remote from the expected atriotomy sites. For example, in three patients with LA approaches, large areas of low voltage were present in the lateral wall of the RA or the septum. Two patients were also found to have low voltage in the posterior wall of the LA, distant from the left atrial incisions. The presence of large areas of abnormal tissue suggests that factors other than a simple suture line contribute to the myopathic process. These might include interruption of blood supply to neighboring tissue after atrial incisions, atrial ischemia or trauma at the time of surgery and mechanical stresses pre- and postoperatively.
Surgical approaches to the MV
The conventional left atriotomy involves an incision in the right lateral wall of the LA anterior to the right pulmonary veins (Fig. 1A). Consistent with this anatomy, we identified abnormal tissue adjacent to the right pulmonary veins in these patients. The relationship between the LA incision and the right pulmonary veins is noteworthy, as scar contiguous with the pulmonary veins may form a substrate for reentry. A modification of this approach involves dissection in the interatrial groove and a left atriotomy near the limbus of the interatrial septum (Fig. 1B). Typically, LA approaches also involve bicaval cannulation through separate purse-strings and incisions in the RA, and retrograde cardioplegia via the coronary sinus may also require a third purse-string suture. Whether RA cannulation causes scar large enough to support atrial arrhythmias has not been established. However, we demonstrated abnormal RA tissue in several patients with a left atrial approach, including an example of macro–re-entry in the RA following a left atriotomy (Patient 2).
Over the past 10 years, the superior trans-septal approach has gained favor as a means of providing exposure to the mitral annulus, particularly in cases of reoperation (Fig. 1, C1and C2). This involves incisions in the lateral RA wall, extending anterior to the SVC, the interatrial septum and the dome of the LA. Patients with the superior trans-septal approach were found to have discrete areas of low voltage within these regions and tachycardias arising from either the right or left atria.
There are conflicting data regarding arrhythmia complications related to the different surgical approaches to the MV (7,8). Because the superior transseptal approach compromises blood supply to the sinus node, some authors have noted higher frequencies of sinus node dysfunction after this operation (7). However, it is uncertain if the superior trans-septal approach, with its more extensive atrial incisions, results in a higher frequency of atrial tachyarrhythmias.
Figure-of-eight re-entry and single circuits
In the RA, lesions in the lateral wall typically formed the substrate for figure-of-eight re-entry. One example of single-circuit reentry involving a lateral lesion occurred in a patient with prior ablation in the IVC-tricuspid isthmus, suggesting that figure-of-eight might have occurred if conduction was present in the isthmus. This highlights the importance of identifying dual loop tachycardias and the potential limitations of confining ablation to the cavotricuspid isthmus.
Figure-of-eight reentry in the RA consisted of two limbs around superior and inferior lesions in the lateral wall, with a central isthmus between these two abnormal areas. Equal activation times were observed around each limb of the circuit, thus supporting a figure-of-eight mechanism. A relatively narrow isthmus permitted limited ablation of these arrhythmias. The location of atrial lesions may play a role in determining whether typical atrial flutter develops. In an animal model of lesional reentry in the RA, circuits confined to the lateral wall were demonstrated with anterior lesions, whereas posterior lesions gave rise to typical atrial flutter circuits (9).
In the LA, lesions in the septum adjacent to the right pulmonary veins provided the common substrate for reentry. The pulmonary veins and associated lesions served as central obstacles for single reentrant circuits or figure-of-eight reentry. Figure-of-eight reentry in the LA took the form of one loop around an area of scar and one loop around the mitral annulus. In the case of single reentrant circuits, more complex conduction patterns were observed around the mitral annulus, including fusion or bifurcation of wavefronts. Jais et al. (10)have recently reported their experience in mapping atypical flutter in the LA and have identified similar tachycardia circuits involving the mitral annulus and the pulmonary veins. They also identified “silent areas” characterized by low amplitude signals that determined the reentrant circuits.
We also demonstrated that AT can arise from focal sources after MV surgery. In one case of focal AT from the left inferior pulmonary vein, the pathogenesis of this arrhythmia is probably unrelated to MV surgery, as tachycardias arising from the pulmonary veins occur in patients with and without structural heart disease. However, focal origins of AT were also demonstrated in the RA adjacent to abnormal tissue, characterized by low voltage, double potentials and conduction block. A variety of mechanisms might account for focal AT in this setting, including micro–re-entry, triggered activity and abnormal automaticity.
These data do not permit a distinction between anatomical and functional block, because maps were acquired during AT. The large size of some low-voltage regions suggests a fixed abnormality, although the existence of functional block cannot be excluded. It is possible that lines of conduction block form adjacent to fixed lesions, as demonstrated in a recent animal study of right atrial lesional tachycardias (9).
The identification of figure-of-eight re-entry was based on electroanatomic maps that show equal activation times in each circuit. More definitive evidence for this mechanism would include entrainment or ablation in each limb independently.
We demonstrate distinct patterns of low voltage and conduction block in patients with prior MV surgery. Abnormal tissue, which forms the substrate for atrial arrhythmias, may be found in both atria and, in most cases, appear to be related to surgical incisions. Mechanisms of arrhythmogenesis include single circuit reentry, figure-of-eight re-entry and focal tachycardias.
☆ This work was supported in part by grants from the National Institute of Health (RO1 56139), the American Heart Association, New York City Affiliate Grant-in-Aid, the Maurice and Corinne Greenberg Arrhythmia Fibrillation Research Grant and the Raymond and Beverly Sackler Foundation.
- atrial fibrillation
- atrial tachycardia
- inferior vena cava
- left atrium
- mitral valve
- right atrium
- superior vena cava
- tricupsid valve
- Received December 20, 2001.
- Revision received March 20, 2002.
- Accepted April 3, 2002.
- American College of Cardiology Foundation
- Kalman J.M.,
- VanHare G.F.,
- Olgin J.E.,
- Saxon L.A.,
- Stark S.I.,
- Lesh M.D.
- Shah D.,
- Jais P.,
- Takahashi A.,
- et al.
- Love B.A.,
- Collins K.K.,
- Walsh E.P.,
- Triedman J.K.
- Nakagawa H.,
- Shah N.,
- Matsudaira K.,
- et al.
- Delacretaz E.,
- Ganz L.I.,
- Soejima K.,
- et al.
- Tomita Y.,
- Matsuo K.,
- Sahadevan J.,
- Khrestian C.M.,
- Waldo A.L.
- Jais P.,
- Shah D.C.,
- Haissaguerre M.,
- et al.