Author + information
- Received April 19, 2005
- Revision received May 26, 2005
- Accepted July 4, 2005
- Published online November 15, 2005.
- Nitish Badhwar, MBBS, FACC⁎,
- Jonathan M. Kalman, MBBS, PhD, FACC†,
- Paul B. Sparks, MBBS, PhD†,
- Peter M. Kistler, MBBS†,
- Mehran Attari, MD‡,
- Marcie Berger, MD, FACC‡,
- Randall J. Lee, MD, PhD, FACC⁎,
- Jasbir Sra, MD, FACC‡ and
- Melvin M. Scheinman, MD, FACC⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Melvin M. Scheinman, Section of Cardiac Electrophysiology, University of California, San Francisco, 500 Parnassus Avenue, MU-East 434, Box 1354, San Francisco, California 94143.
Objectives We sought to describe the electrophysiological features and long-term outcome after radiofrequency catheter ablation (RFCA) of atrial tachycardia (AT) arising from the coronary sinus (CS) musculature.
Background Atrial tachycardia requiring RFCA deep within the CS has been described in isolated case reports. However, the mechanism and exact site of origin of this tachycardia have not been well elucidated.
Methods The study included 8 patients (5 men) of a consecutive series of 283 patients undergoing RFCA for focal AT.
Results In sinus rhythm, a discrete potential (P) was noted after the CS atrial electrogram and during tachycardia, the CS (P) preceded the surface P-wave by 30 to 50 ms. The CS (P) always preceded the earliest electrogram in the left atrium (LA). Three-dimensional electroanatomical mapping was available in four patients, and in one case it showed earliest activation in the CS with rapid spread to the proximal CS and then to the LA. Ablation of the AT initially attempted from the earliest site in the LA in three patients was unsuccessful. In all patients the tachycardia was safely and successfully ablated at a site 3.6 cm within the CS. There has been no recurrence over a follow-up of 37 ± 13 months.
Conclusions Focal AT emanating deep within the CS musculature can be recognized by a discrete potential associated with the CS atrial signal both during sinus rhythm and tachycardia. Long-term success without complications can be accomplished by ablating within the CS in close proximity to the CS (P).
Focal atrial tachycardia (AT) usually arises from well-defined anatomic regions like the crista terminalis, tricuspid annulus, coronary sinus (CS) ostium in the right atrium (RA), and from the pulmonary veins and mitral annulus in the left atrium (LA). Radiofrequency catheter ablation (RFCA) of focal AT is achieved by targeting sites of earliest atrial activation that precede the “P-wave” during tachycardia. In the prior case reports of AT requiring RFCA within the CS (1–5), there is no definitive evidence that the origin of the tachycardia is from CS muscle itself or from atrial muscle bundles in the ligament of Marshall (LOM) or from a part of the LA contiguous to the site of ablation. In addition, there is no data relative to tachycardia mechanism. The purpose of our study is to describe the clinical and electrocardiographic features, mechanism, and site of origin of this tachycardia, as well as the long-term outcome after RFCA.
This is a retrospective study of all patients undergoing focal AT ablation at three medical centers (University of California, San Francisco; Royal Melbourne Hospital, Melbourne; and Aurora Sinai/St. Luke’s Medical Centers, University of Wisconsin Medical School-MCC, Milwaukee) between September 2000 and July 2003. A total of 283 patients undergoing focal AT ablation were evaluated, and eight patients (3%) who underwent radiofrequency ablation deep within the CS are the subjects of this study.
All patients had symptomatic documented tachycardia and proved refractory to at least one antiarrhythmic agent. None received amiodarone.
Each patient signed a written informed consent. Patients were studied in the postabsorptive state using a combination of intravenous midazolam and fentanyl. All antiarrhythmic medications were discontinued at least five half-lives before the study. Conventional 12-lead surface electrocardiogram (ECG) and bipolar intracardiac electrogram recordings (filtered between 30 and 500 Hz) were amplified and displayed on a Bard Electrophysiology Lab System Duo (C.R. Bard Inc., Lowell, Massachusetts) in three patients, Cardiolab System (Prucka Engineering, Sugarland, Texas) in four patients, and Real Time Position Mapping (RPM) System (EP Technologies, Boston Scientific, Natick, Massachusetts) in one patient. Two quadripolar catheters were placed via the right femoral vein in the high RA and His-bundle position. A decapolar catheter was placed in the CS with the proximal pole at the ostium. Intravenous isoproterenol and atrial overdrive or programmed stimulation were used for arrhythmia induction if spontaneous tachycardia was not present at baseline.
Mapping of tachycardia
Focal AT was diagnosed based on standard electrophysiological criteria (6). Activation mapping during tachycardia to look for sites of earliest endocardial activation relative to the surface P-wave was done with a 4-mm tip quadripolar steerable ablation catheter in the RA, LA, and CS. A transeptal puncture was performed using conventional techniques to map in the LA. Intravenous heparin was used to maintain the activation clotting time at >250 ms after LA access was obtained. The 12-lead ECG was analyzed to look for the earliest P-wave onset during tachycardia. Activation time was measured from the onset of the first sharp component of the bipolar electrogram on the distal mapping catheter to the earliest P-wave. Three-dimensional (3D) electroanatomical mapping using the CARTO System (Biosense Webster, Diamond Bar, California) was performed in three patients, and the RPM system was used in one patient. The 3D mapping system was used to tag the earliest endocardial activation as well as ablation spots; CS venography was performed in five patients.
Surface 12-lead P-wave morphology was assessed as per Tang et al. (7). The P waves were described on the basis of the deviation from baseline during the T-P interval as being: 1) positive (+); 2) negative (−); 3) biphasic: if there were both positive and negative (± or −/+) deflections from baseline; and 4) isoelectric: arbitrarily defined when there was no P-wave deflection from baseline of >0.05 mV. Figure 1illustrates the P-wave morphology during tachycardia.
RFCA and long-term follow-up
Radiofrequency ablation was performed in the CS with continuous temperature feedback control of the power output to achieve a target temperature of 50°C. Initial power setting was 5 W, and the maximum power used was 30 W for a maximum of 60 s. Three patients also had attempted ablation in the LA, where a temperature of 60°C and maximum power of 50 W were used. Acute procedural success was defined by the inability to induce tachycardia 30 min after ablation despite aggressive burst atrial pacing and the use of isoproterenol. The patients were followed in our respective clinics in order to assess return of symptoms or documented tachycardia.
The study population included eight patients (five men) with a mean age of 34 ± 14 years (Table 1).One patient had rheumatic heart disease (postaortic valve replacement); the remainder had no structural heart disease. The patients had paroxysmal tachycardia that was present despite treatment with at least one antiarrhythmic drug.
Electrophysiological characteristics of tachycardia
Tachycardia occurred spontaneously in most patients and could be easily induced with atrial burst pacing (Table 2).Isoproterenol was required for tachycardia maintenance in five patients. Tachycardia cycle length varied from 250 to 450 ms (330 ± 79 ms). Tachycardia cycle length decreased with isoproterenol infusion in three patients. Spontaneous or adenosine-induced atrioventricular block was present in all patients, and atrioventricular dissociation was noted during tachycardia with ventricular overdrive pacing in all of them. Both of these observations exclude atrioventricular reentrant tachycardia using an accessory pathway. The eccentric activation pattern in the CS electrograms argued against atrioventricular nodal reentrant tachycardia. Atrial overdrive pacing (AOD) during tachycardia was performed in all patients, and no evidence of entrainment was found. Tachycardia terminated with AOD in two patients. Adenosine administered in conventional doses caused atrioventricular block without termination of tachycardia.
The earliest endocardial activation was recorded in the CS with an average distance of 3.6 cm from the CS ostium (Table 2). This site always preceded the earliest LA endocardial site during tachycardia. All patients were noted to have a sharp potential in the CS that preceded the CS atrial signal during tachycardia (Fig. 2B).The CS potential (P) preceded the surface P-wave by 30 to 50 ms (average 38 ms) during tachycardia (Table 2). This potential was noted to occur after the CS atrial electrogram (A) during sinus rhythm in six of the eight patients as shown in Figure 2A. After successful ablation this CS (P) was still visible in two patients. Catheter ablation performed from the earliest site at the mitral annulus in one patient (Fig. 3A)led to prolongation of tachycardia cycle length as well as prolongation of the interval between the CS (P) and the CS (A) (Fig. 3B). Tachycardia terminated with dissociation of the CS (P) from the LA electrogram (Fig. 3C). The interval between the dissociated CS (P) was identical to the tachycardia cycle length. However, tachycardia could be reinitiated when the sharp potential again became coupled to the body of atrial myocardium and could not be ablated from the LA. Catheter ablation at the site of the earliest CS (P) (in the CS) terminated the tachycardia. In this case, the tachycardia focus was clearly dissociated from the atrium and proves that the early sharp potential is driving the tachycardia. In addition, the LA radiofrequency application must have interfered with conduction along the CS muscle because it failed to activate the RA. The observation that remote ablation (LA) caused exit block from the tachycardia circuit together with successful ablation at the earliest CS (P) strongly suggests a site of origin from the CS muscle rather than the tracts in the LOM.
LA mapping during tachycardia
All patients underwent transeptal puncture with mapping of the LA. The early LA sites preceded the surface P-wave by 15 to 20 ms, but ablation at these sites in three patients proved to be ineffective. The earliest CS electrogram and specifically CS (P) always preceded the earliest LA electrogram (Fig. 4).
Electroanatomic mapping of the tachycardia
For the group as a whole, both LA as well as RA activation maps were available in four patients (CARTO in three, RPM in one). The earliest endocardial LA site was localized to the posterolateral mitral annular region. The LA maps were consistent with a focal origin (Fig. 5)as manifested by a radial spread of activation, and the endocardial activation time was 32 ± 14% of the tachycardia cycle length. Also of note was a relatively broad area of early LA activation.
In one patient simultaneous LA and CS 3D endocardial activation was obtained. The 3D propagation map showed earliest activation from a focus in the CS (3.6 cm from the CS os) with rapid spread along the CS musculature, followed by sequential activation of the LA (Figs. 6 and 7).⇓⇓Catheter ablation at the earliest site in the CS terminated the tachycardia.
Coronary sinus venography was performed in five patients. It did not show any CS abnormalities. Selective cannulation of the vein of Marshall was not done. It has been shown that the muscle bundles of LOM insert at the proximal CS musculature or posterior LA free wall superior to the CS (8). The earliest CS electrogram in our study was recorded in the distal CS, arguing against the LOM as the cause of the tachycardia.
The P-wave morphology on the surface ECG during tachycardia is shown in Table 3.The 12-lead ECG from one patient showing tachycardia with adenosine-induced atrioventricular block is shown in Figure 1. The P waves were consistently negative in inferior leads and positive in V1and avR. The precordial leads showed a transition to negative P waves in V4to V6. The P-wave in lead I was negative or isoelectric, and lead avL consistently showed an initial positive component.
Radiofrequency catheter ablation was performed at the site of earliest atrial activation in the LA at the mitral annulus in three patients. There was acute success followed by recurrence within minutes in two patients and no response in one. The CS (P) was transiently dissociated from the atrial electrogram in the third patient. However, tachycardia was reinducible. Tachycardia slowing before termination was noted in one patient.
Tachycardia termination associated with loss of the CS (P) was documented transiently in three patients during mapping in the CS and was attributed to mechanical trauma. A 3D mapping system was used to mark the site where earliest CS (P) was recorded during tachycardia allowing for RFCA in sinus rhythm (three patients); RFCA performed in sinus rhythm in one patient initiated AT followed by prompt termination. In all cases the earliest endocardial activation was located in the CS, where a potential preceded the surface P-wave by a mean of 38 ± 7 ms. Acute success was achieved by RFCA in all patients at this site. The mean number of RFCA lesions was 2.5. Low powers (5 W to 30 W) were used, and maximum temperature was set to 50° C.
Adverse effects and follow-up
One patient had clinical symptoms of pericarditis without pericardial effusion that resolved with nonsteroidal drugs. No long-term complications were observed. These patients were followed at three to six months at the individual clinics, and long-term success without any complications was noted in all patients at a mean follow-up of 37 ± 13 months.
Although earlier isolated case reports have described an unusual AT arising from CS musculature (1–5), we report the first series to systematically describe the electrocardiographic and electrophysiological features of this tachycardia. In addition we provide clear-cut evidence that the tachycardia focus is independent of the LA. The LOM has been shown to have variable and extensive connections to the LA (8) that were never observed in any of our 3D maps. The tachycardia mechanism is consistent with triggered activity or abnormal automaticity, and the long-term safety and efficacy of catheter ablation from within the CS is documented.
Tachycardia was easily induced by AOD, and isoproterenol was required for AT maintenance in the majority of the cases. This is compatible with a triggered activity or abnormal automaticity. The results of AOD with respect to lack of evidence for entrainment as well as the electroanatomical maps weigh against a reentrant mechanism. Clearly, the 3D maps rule out macroreentry but do not exclude microreentry. Intravenous adenosine in doses that caused atrioventricular block failed to terminate tachycardia, but graded increases of dose were not given. Atrial tachycardia termination with adenosine has previously been used to classify AT as focal or triggered (9). However, the majority of the focal AT in that study were localized to the crista terminalis or the tricuspid annulus, and they did not include any AT arising from within the CS. Lack of response to adenosine in our patients could either signify different tissue properties of the CS musculature or lack of adequate adenosine dose.
We identified a distinct CS (P) during tachycardia in all patients. The CS (P) was found to precede the surface P-wave by 30 to 50 ms (Table 2). A similar potential was described in three prior case reports (2,3,5). A novel finding was the observation that the CS (P) followed the P-wave in sinus rhythm in six of eight patients while in two the potential was apparently merged with LA electrogram and was thus inapparent until tachycardia supervened. In all instances the CS (P) preceded the P-wave. This finding appears to be pathognomic of CS AT, as similar findings have been used as evidence of AT and atrial fibrillation arising from pulmonary veins (10). Another unique finding was the ability to dissociate the CS (P) from the LA electrogram during ablation from the LA site contiguous to the earliest CS site. This observation proves that the potential is independent of the LA and hence localized to the CS muscle. Successful ablation from CS muscle and not from the earliest LA site in three of our patients and in four patients from the literature (1,3–5) also supports the notion that the tachycardia originated from the CS musculature.
The four propagation maps were of interest from several perspectives. First, the LA endocardial maps showed a broad activation pattern compatible with a source emanating from a contiguous structure. This finding was found in two previous case reports (3,5). A unique finding in our study was the observation of dissociation of CS (P) and LA activation in one patient. Another patient showed a clear-cut activation pattern emanating from the CS musculature with rapid sequential spread to the CS os region, followed by LA activation. This essentially excludes a site of activation from muscle tracts within the LOM. The available anatomic and electrophysiological data suggest widespread communication of tracts from LOM to the LA (8). Hence, a site of activation from LOM should result in rapid initial or parallel activation of the LA that was not observed in our patient. The value of using advanced imaging techniques for these patients is emphasized by the frequency of tachycardia termination with mechanical trauma. The ability to mark early areas allowed for ablation in sinus rhythm in three patients in our series and in two patients (3,5) in the literature.
Previous studies have used the surface P-wave morphology during AT to predict the site of origin (7). It has been shown that the P-wave morphology is largely determined by the direction of septal and left atrial activation (11); hence, the morphology may vary among patients with the same AT depending on the dominant route of LA activation. Nevertheless, the morphology was remarkably consistent in patients with CS muscle tachycardia. All patients had a positive P-wave in V1and avR and negative P waves in the leads II, III, and avF suggestive of origin from the left side and the inferior part of the atrium. The precordial leads were consistently positive in V1and predominantly negative in V6with a transition at V3or V4. This transition is later than that observed for AT originating at the CS ostium where transition usually occurs between V1and V2. Lead I was isoelectric or minimally negative suggesting a left-sided origin. The P-wave in avL in all patients showed an initial positive component followed either by a negative or isoelectric component. These characteristic findings were also noted in the previous case reports (1–5). In summary, the finding of positive P waves in V1and avR associated with negative P waves in inferior leads and a positive P-wave in avL is highly suggestive of AT arising from the CS musculature.
CS anatomy and atrial arrhythmias
The CS is defined as the portion of the cardiac venous system starting from the insertion of the oblique vein of Marshall and terminating at the CS ostium in the RA (12,13). In an elegant post-mortem study of 240 human hearts Lùdinghausen et al. (12) have shown that the CS is an intramural structure. In this study myocardial coverings were noted to extend 2 to 11 mm beyond the opening of the vein of Marshall. Chauvin et al. (13) have also demonstrated a muscular cuff surrounding the CS and extending, on average, 4 cm from the ostium. The myocardial sleeve around the CS has been shown to be composed of bands of muscle from the LA as well as the RA (12–14). Wit and Cranefield (15) in an elegant series of experiments involving direct microelectrode recording from both the region of the CS os as well as from muscle deep within the CS elucidated the mechanism of arrhythmias induced from the CS. They found that arrhythmias could be initiated by norepinephrine and/or pacing. They found that arrhythmias originating from the region of the os were due to abnormal automaticity while those arising from deep within the CS were triggered rhythms that could be abolished with verapamil. Our clinical findings are compatible with this experimental data.
Our findings are somewhat similar to reports of AT (or even fibrillation) originating from the muscle sleeves extending from the LA into and surrounding the pulmonary veins (10). Interestingly, a recent report from Haissaguerre et al. (16) described the involvement of CS foci in the mechanism of paroxysmal atrial fibrillation.
While CS venography in five patients showed no CS abnormalities, the procedure was not performed in three patients. We also did not attempt to selectively cannulate the vein of Marshall. It could be possible that the origin of this tachycardia could be from the LOM (8) in some patients. However, the LOM is known to have extensive connections in the LA that may require ablation in the LA (8,17,18). The dissociation of the CS (P) by RFCA in LA (Fig. 3) and termination of AT by RFCA in CS strongly suggests an origin from CS musculature. Similarly, data from Patient #7 showing earliest activation of the CS musculature (Figs. 6 and 7) is not compatible with tachycardia from LOM. The present study does not conclusively prove the mechanism of the tachycardia. It is suggestive of abnormal automaticity or triggered activity and argues against macroreentry, but microreentry cannot be conclusively excluded. Verapamil was not used in our patients. A 3D mapping system was used to select the ablation site in sinus rhythm (at the site of earliest activation during AT) in three patients in whom catheter manipulation led to tachycardia termination. Acute and long-term success was obtained in all these patients. We acknowledge that catheter position at identical site(s) can differ in 3D maps acquired during AT versus during sinus rhythm, and one should consider using other tools like pace mapping to confirm the appropriate ablation site.
Focal AT emanating deep within the CS musculature can be recognized by a discrete potential associated with the CS atrial signal both during sinus rhythm and tachycardia. It is associated with a distinctive surface P-wave morphology during tachycardia. Ablation performed from the LA is unsuccessful in these patients. Three-dimensional mapping is useful in marking sites of early activation in the CS for ablation, as the tachycardia is sensitive to mechanical pressure during catheter manipulation and may terminate. Long-term success without complications can be accomplished by ablating in close proximity to the CS (P).
- Abbreviations and Acronyms
- atrial overdrive pacing
- atrial tachycardia
- coronary sinus
- CS (A)
- coronary sinus atrial electrogram
- CS (P)
- coronary sinus potential
- left atrium
- ligament of Marshall
- right atrium
- radiofrequency catheter ablation
- Received April 19, 2005.
- Revision received May 26, 2005.
- Accepted July 4, 2005.
- American College of Cardiology Foundation
- Saoudi N.,
- Cosio F.,
- Waldo A.,
- et al.
- Tang C.W.,
- Scheinman M.M.,
- Van Hare G.F.,
- et al.
- Kim D.T.,
- Lai A.C.,
- Hwang C.,
- et al.
- Iwai S.,
- Markowitz S.M.,
- Stein K.M.,
- et al.
- Okumura K.,
- Plumb V.J.,
- Page P.L.,
- Waldo A.L.
- Chauvin M.,
- Shah D.C.,
- Haissaguerre M.,
- Marcellin L.,
- Brechenmacher C.
- Rotter M.,
- Sanders P.,
- Takahashi Y.,
- et al.
- Polymeropoulos K.P.,
- Rodriguez L.M.,
- Timmermans C.,
- Wellens H.J.
- Tai C.T.,
- Hsieh M.H.,
- Tsai C.F.,
- et al.