Author + information
- Received July 7, 2012
- Revision received January 8, 2013
- Accepted February 3, 2013
- Published online May 21, 2013.
- Stavros E. Mountantonakis, MD⁎,†,
- Robert E. Park⁎,
- David S. Frankel, MD⁎,
- Mathew D. Hutchinson, MD⁎,
- Sanjay Dixit, MD⁎,
- Joshua Cooper, MD⁎,
- David Callans, MD⁎,
- Francis E. Marchlinski, MD⁎ and
- Edward P. Gerstenfeld, MD⁎,‡,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Edward P. Gerstenfeld, Medicine/Cardiology, MU-East 4th Floor, University of California, San Francisco, 500 Parnassus Avenue, San Francisco, California 94143
Objectives The goal of this study was to determine the relationship of the ventricular tachycardia (VT) isthmus to channels of preserved voltage on an electroanatomic voltage map in postinfarction cardiomyopathy.
Background Substrate mapping in patients with postinfarction cardiomyopathy and VT may involve lowering the voltage cutoff that defines the scar (<1.5 mV) to identify “channels” of relative higher voltage within the scar. However, the prevalence of channels within the scar identified by using electroanatomic mapping and the relationship to the protected VT isthmus identified by entrainment mapping is unknown.
Methods Detailed bipolar endocardial voltage maps (398 ± 152 points) from 24 patients (mean age 69 ± 9 years) with postinfarction cardiomyopathy (ejection fraction 33 ± 9%) and tolerated VT were reviewed. Endocardial scar was defined according to voltage <1.5 mV. Isolated late potentials (ILPs) were identified and tagged on the electroanatomic voltage map. The baseline voltage cutoffs were then adjusted until all channels were identified. The VT isthmus was identified using entrainment mapping.
Results Inferior and anterior/lateral infarction was present by voltage mapping in 18 and 6 patients, respectively (scar area 44 ± 24 cm2). By adjusting voltage cutoffs, 37 channels were identified in 21 (88%) of 24 patients. The presence of ILPs within a channel was seen in 11 (46%) of 24 patients and 17 (46%) of 37 channels. A VT isthmus site was contained within a channel in only 11 of 24 patients or 11 of 37 channels. No difference in voltage characteristics was identified between clinical and nonclinical channels. Voltage channels with ILPs harbored the clinical isthmus with a sensitivity and specificity of 78% and 85%, respectively.
Conclusions Channels were identified in 88% of patients with VT by adjusting the voltage limits of bipolar maps; however, the specificity of those channels in predicting the location of VT isthmus sites was only 30%. The presence of ILPs inside the voltage channel significantly increases the specificity for identifying the clinical VT isthmus.
Reentrant ventricular tachycardia (VT) in postinfarction cardiomyopathy typically depends on a critical isthmus of slow conduction in the myocardial infarct that is electrically isolated from the rest of the myocardium (1). These isthmus sites can be characterized by entrainment maneuvers in hemodynamically stable VT (2). Previous studies have demonstrated that these corridors of slow conduction within the scar provide the substrate for reentrant VT and often contain isolated late potentials (ILPs) (3–7). When VT is not hemodynamically tolerated, “substrate” ablation consists of linear ablation through the scar border zone at sites of pacemaps that match the clinical VT (3), in addition to the ablation of ILPs that may indicate a critical zone of slow conduction. However, because these ILP sites are not specific to the clinical VT, this strategy often results in an extensive ablation, especially in patients with large myocardial infarctions.
The use of electroanatomic mapping to identify paths of relatively higher voltage within a scar that could serve as a zone of slow conduction for reentrant VT has also been proposed (8,9). Identification of those “channels” involves lowering the usual voltage cutoff that identifies scars (<1.5 mV) to identify regions of relatively preserved voltage within the denser scar that could serve as the protected VT isthmus. Ablation transecting these channels is another proposed strategy for substrate modification in poorly tolerated VT. However, although it has been demonstrated that a clinical VT may sometimes be associated with a higher voltage “channel,” a critical analysis of the number of identifiable channels and their relationship to the clinical VT has not been performed. In addition, there are several other possible explanations for relatively higher voltage channels within a myocardial scar, including undersampling of points within the scar, poor electrode contact or change in electrode orientation, and bystander or blind loop regions that are not critical to the VT circuit.
In the current study, our goal was to examine the relationship between channels identified on an electroanatomic map, ILPs, and the location of the VT isthmus identified by entrainment mapping during VT occurring in patients with postinfarction cardiomyopathy. We hypothesized that the mere presence of relative higher voltage channels within the scar area would not be specific for the VT isthmus and that channels containing ILP would be more likely to identify an isthmus associated with the clinical VT. To be sure we identified the critical VT isthmus, we only included patients with mappable VT in which an isthmus could be identified with entrainment mapping and where ablation at the isthmus site terminated the VT.
We retrospectively reviewed endocardial electroanatomic maps (CARTO XP or CARTO 3, Biosense Webster, Inc., Diamond Bar, California) from patients with postinfarction cardiomyopathy undergoing catheter ablation for recurrent, drug-resistant VT at the University of Pennsylvania between March 2007 and March 2010. Only patients with hemodynamically stable VT whose isthmus site was identified with entrainment and where VT was terminated with ablation at this site were included in the analysis. We also only included patients with complete and homogenous maps (>300 points, fill threshold <10 mm). All procedures were performed following the institutional guidelines of the University of Pennsylvania Health System.
Patients were brought to the electrophysiology laboratory after providing written informed consent. Antiarrhythmic drugs were discontinued before the procedure. Surface electrocardiogram (ECG) leads were placed in the standard positions. Bipolar electrograms were recorded with a bandpass filter at 30 to 500 Hz by using a 3.5-mm open irrigated tip ablation catheter (ThermoCool, Biosense Webster, Inc.). A 6-F quadripolar catheter was placed in the right ventricular apex for pacing. An 8-F intracardiac echocardiography catheter (ACUSON AcuNav, Siemens Medical Solutions, Mountain View, California) was advanced to the right ventricle to assist with catheter positioning, monitor lesion formation, and assess for complications. Electroanatomic maps were initially created during sinus rhythm in all cases before the induction of VT. In pacemaker-dependent patients or patients with biventricular pacemakers, the voltage map was created during continuous right or biventricular pacing.
VT was then induced with programmed stimulation from the right ventricular apex. If VT was not initiated, programmed stimulation from the left ventricle near the area of scar was performed. We included only patients in whom the clinical tachycardia was induced, based on ECG morphology, cycle length, and stored device intracardiac electrograms. Once the hemodynamically stable VT(s) were induced, entrainment mapping was performed, and the VT isthmus site(s) were identified and tagged by the operator on the electroanatomic map. The VT isthmus was identified by using classic techniques, including: entrainment with concealed fusion, stimulus to QRS <70% and >30% VT cycle length, postpacing interval and tachycardia cycle length <30 ms, and a stimulus to QRS and electrogram to QRS interval within 30 ms. Additional entrainment sites (entrance, exit) were also tagged on the map at the operator's discretion. Once the critical isthmus was identified, radiofrequency ablation was performed at the isthmus site, and upon termination of the tachycardia a new point was tagged in sinus rhythm. If VT did not terminate after more than 3 radiofrequency ablations at this site, mapping continued and the site was not counted as a critical isthmus. After VT termination, additional ablation lesions were delivered at the isthmus sites as well as additional substrate-based ablation at the operator's discretion. Programed ventricular stimulation was then repeated to assess inducibility of the clinical arrhythmia.
Analysis of electroanatomic maps
The electroanatomic maps were analyzed offline for identification of ILPs and voltage channels. Points that were obtained after initiation of ablation were removed. At baseline, the nominal setting for the color display of voltage maps had an upper threshold of 1.5 mV (purple) and a lower threshold of 0.5 mV (red). All points within the scar or border zone (<1.5 mV) were reviewed, and all ILPs were identified and tagged on the map. We then attempted to uncover channels of contiguous, relatively higher voltage amplitude compared with adjacent tissue within the area of scar by adjusting the voltage cutoffs in a systematic fashion. First, the upper threshold of the color display was decreased in decrements of 0.05 mV until the value of 0.55 mV was reached. Both the upper and the lower color thresholds were then decreased in decrements of 0.01 mV until 0.05/0.01-mV values were reached. We then readjusted the voltage levels to 0.54/0.50 mV and decreased both upper and lower voltage cutoffs in 0.01 mV decrements, maintaining a difference of 0.04 mV between the upper and lower voltage cutoffs. The same method was used to maintain a 0.03-, 0.02-, and 0.01-mV difference between the upper and lower voltage cutoffs. When a voltage channel appeared, minor adjustments of the voltage cutoff limits were made to bring out the voltage difference between the channel and surrounding scar.
The presence of postinfarction cardiomyopathy was confirmed by either the presence of either a history of myocardial infarction with >70% narrowing of an epicardial coronary artery, Q waves on the 12-lead ECG, or the presence of a perfusion defect on myocardial perfusion imaging in a typical coronary distribution corresponding to a low-voltage zone on electroanatomic mapping.
ILPs were defined as split electrograms, with the second component occurring after the terminal QRS and >20 ms after the end of the local ventricular electrogram (7). For a voltage path to be characterized as a “channel,” it had to be defined by at least 3 contiguous points of higher voltage amplitude than that of the surrounding area and transect the scar. This difference in voltage amplitude had to be significant enough to create a corridor of purple color transecting an area of red color. The voltage setting that resulted in demonstration of a channel was recorded. The first channel to appear when decreasing the upper voltage cutoff was noted. The geometrical characteristics (length and width) of each channel were then measured by using the CARTO XP or CARTO 3 software. The width of each channel was measured at 3 different sites along the channel, and the average is presented. The distance of the isthmus site recorded in sinus rhythm to the nearest channel was also noted. Voltage channels that harbored a VT isthmus were defined as clinical, and all other channels were defined as nonclinical.
The infarct area was circumscribed by using the surface area tool on the CARTO mapping system and compared with the endocardial surface area of the complete left ventricular map.
Continuous data are expressed as mean ± SD. The Student 2-tail t test was used to compare voltage differences between channels that harbor isthmuses and those that did not. The Fisher exact test was used to assess the relationship between the presence of ILPs and presence of a clinical channel. Statistical analyses were performed by using SPSS version 16.0 (IBM SPSS Statistics, IBM Corporation, Armonk, New York). A p value <0.05 was considered significant.
Between March 2007 and March 2010, a total of 140 patients with postinfarction VT were referred to the University of Pennsylvania for catheter ablation. Of those, 24 patients (23 males; mean age 67 ± 9 years) (Table 1) had a hemodynamically stable clinical VT that was induced in the laboratory and permitted identification of the critical isthmus by entrainment mapping. In 18 (75%) of 24 patients, the 12-lead ECG of the clinical VT was available, and a direct confirmation of the induced VT was possible. In the remaining 6 patients, the clinical VT was characterized based on matching the cycle length and morphology of the intracardiac stored electrograms. The mean left ventricular ejection fraction was 33 ± 9%. Eight patients had undergone previous ablation for VT. The hemodynamically stable VT that allowed identification of an isthmus site had a mean cycle length of 410 ± 63 ms. After ablation of the clinical arrhythmia, repeat programmed stimulation failed to induce the clinical tachycardia in all patients. An additional 27 “nonclinical,” poorly tolerated tachycardias were induced in 19 of the 24 patients with aggressive programmed ventricular stimulation after the clinical tachycardia was rendered noninducible. Sixteen of those tachycardias were targeted with additional substrate ablation at the presumed exit site based on pacemapping and ECG morphology. In all cases, the exit of the nonclinical tachycardia was not in proximity to the identified exit of the clinical VT.
The complete maps that are included in this analysis contained a mean of 398 ± 152 points. The majority of patients had inferior/inferoseptal scar (n = 18) and less frequently anterior/anteroapical (n = 4) or anterolateral/lateral (n = 2) scar. Mean scar area was 18 ± 9% of the total left ventricular endocardial surface area. The presence of ILPs was identified in the majority of patients (19 of 24 [79%]).
By adjusting voltage cutoffs, we were able to identify 37 putative channels in 21 (88%) of 24 patients. The presence of both ILPs and channels was seen in 19 (79%) of 24 patients, and the presence of ILPs within a voltage channel was seen in 11 (46%) of 24 patients and 17 (46%) of 37 channels. ILPs in sites adjacent to the isthmus were seen in 19 (79%) of 24 patients. A VT isthmus site was contained within a channel in 11 (30%) of 37 channels and in 11 (46%) of 24 patients. As the voltage cutoffs on the electroanatomic maps were decreased in the prespecified fashion, the identified clinical channel was the first channel to appear in 8 patients, the second channel in 2 patients, and the third channel in 1 patient (Figs. 1 to 3).⇓⇓⇓ Isthmuses for the remaining 13 patients were located a mean distance of 24.4 ± 8.5 mm from the nearest channel (Figs. 4 and 5).⇓⇓ Patients with a VT isthmus that was located within a channel had faster VT, a larger number of identified channels, and more channels harboring ILPs (Table 2). No difference in voltage characteristics was identified between clinical (including a VT isthmus) and nonclinical (no VT isthmus) channels.
No difference was seen in channel location, voltage, or geometrical characteristics between channels that harbor ILPs and those that did not. However, channels with ILPs were more likely to harbor the clinical isthmus. The sensitivity and specificity of any channel harboring an isthmus site regardless of presence of ILPs were 46% and 33%, respectively. On the contrary, channels with ILPs harbored a clinical isthmus with a sensitivity and specificity of 78% and 85%. In addition, the first channel to appear had a sensitivity of 73% but a specificity of only 20% for harboring a clinical VT isthmus.
The clinical VT terminated during ablation in all 24 patients. Additional nonclinical induced VTs were also targeted for ablation at the operators' discretion. After a mean follow-up period of 11.7 ± 5.0 months, 18 patients were arrhythmia free, 1 patient had a recurrence of the clinical VT 2 months after ablation and underwent a repeat ablation that targeted the same VT circuit, and 4 patients were lost to follow-up. Lastly, 1 patient had recurrence of a different morphology VT 9 months after the initial ablation. This VT was not induced or targeted for ablation during the first procedure.
Classic entrainment mapping allows the operator to localize the protected VT isthmus with a single catheter and has a high success rate for catheter ablation; however, a minority of VTs are stable enough to allow entrainment mapping. Substrate-based ablation has expanded our ability to target unstable VTs by using approaches such as pacemapping, targeting ILPs, and targeting “channels” present on an electroanatomic map. However, the mere presence of channels within myocardial scar does not mean that such a channel is capable of supporting VT. We found that: 1) only 11 (30%) of 37 “channels” that could be identified with electroanatomic mapping contained a clinical VT isthmus; 2) only 11 (44%) of 25 mappable VTs were associated with any identifiable channel; and 3) the presence of ILPs within a channel was most specific for a VT isthmus. Our study is in agreement with previous reports which showed that voltage channels could be identified in many patients with postinfarction VT (7–9). Nevertheless, the use of these channels alone in identifying the clinical isthmus has low specificity, and therefore their ability to accurately guide ablation is poor. We demonstrate that the presence of ILPs within a channel increases the likelihood that the channel will be important for maintaining VT.
Demonstration of channels of relatively preserved voltage within a scar using electroanatomic mapping has been reported in previous studies (7–9). This finding has been recently supported by other imaging modalities, including position emission tomography/computed tomography and magnetic resonance imaging (10,11). In the original study by Arenal et al. (8), electroanatomic maps from 26 patients with postinfarction cardiomyopathy were analyzed for presence of voltage channels and their relationship with clinical or induced VTs. The authors found that 20 of 23 identified channels were associated with a VT isthmus. The reason for the lower number of clinically relevant channels in our study is unclear but is likely related to a greater use of entrainment mapping for identifying the critical VT isthmus and the different methods for voltage map analysis used in our study. Arenal et al. used 5 prespecified voltage cutoffs with only a 0.01-mV difference between the upper and lower voltage limits. It is unclear whether such a small difference is truly important. The mean difference between the voltage limits in our study was 0.22 ± 0.20 mV (median 0.10 mV), which in combination with our high-density mapping, may represent a more physiological voltage difference between a channel and surrounding scar. Hsia et al. (9) examined the relationship of “channels” to the VT isthmus in 26 patients with VT. They found evidence of conducting channels associated with 18 of 32 VTs, with VT termination during ablation in 16 of 18 channels. However, the authors only reported on the number of conducting channels that were related to mapped VTs; therefore, the true prevalence of conducting channels and their importance as an ablation target were unclear.
Relative high-voltage channels within scar can often be identified in patients with postinfarction cardiomyopathy and reentrant VT (7,11); however, the majority of channels are “bystanders” or are artifacts. A strategy of empiric ablation of all identified channels in noninducible/poorly tolerated VT is therefore not likely to be useful. A corroborative finding of a functional measure of slow conduction, such as the presence of ILPs within a channel, can increase the specificity of predicting a clinical channel; however, it should be recognized that this will not include the clinical VT isthmus in some patients because the clinical isthmus may still be remote from the identified channels. In addition, because the clinical VT isthmus was identified in the first 2 channels to appear with voltage adjustment in 10 of 11 patients, further voltage adjustment after identification of the first 2 channels is unlikely to be of benefit.
This was a retrospective analysis of electroanatomic maps that were generated before induction of clinical VT. Patients with documented VTs were more likely to have a denser map at the areas of expected VT exit. Differential point density could certainly affect demonstration of voltage channels. Nevertheless, we only included complete maps with a high point density. The presence of ILPs was demonstrated in sinus rhythm or ventricular pacing.
Even though more than 1 VT was induced in some patients, we only used VTs characterized by entrainment mapping as the gold standard for identifying VT isthmus sites (2). In addition, we did not report the relation between entrainment-identified entrance and exit sites to the voltage channels because these sites were not identified in all cases. Pacing within the channels in sinus rhythm to identify which exhibited slow conduction (long stimulus–QRS) was not performed because the ablation was performed during VT. Our findings were based on hemodynamically stable VT; although the extrapolation of these results to hemodynamically unstable VT is reasonable, prospective validation of this ablation strategy is needed.
Finally, after ablation of the clinical VT, other unstable VTs were targeted for ablation using substrate modification at the operator's discretion. Therefore, it is possible that some identified channels were important for other induced, poorly tolerated VTs. This, however, would not alter the low specificity of channels for identifying the clinical VT isthmus.
Channels could be identified in 88% of patients with postinfarction VT by adjusting the voltage limits of bipolar maps; however, the specificity of those channels in predicting the location of VT isthmus sites was only 30%. The presence of ILPs inside the voltage channel significantly increased the specificity for identifying the clinical circuit. When VT is not inducible or hemodynamically tolerated, voltage channels containing ILPs could be targeted as a proxy for VT isthmus sites. This ablation strategy needs to be evaluated prospectively.
Dr. Cooper has received honorarium for education from Boston Scientific, Medtronic, Inc., St. Jude Medical, Biotronik, and Spectranetics; and consulting fees from Medtronic, Inc. Dr. Marchlinski is on the scientific advisory board and has received lecture honorarium from Biosense Webster, Inc. Dr. Gerstenfeld has received research grant support from Biosense Webster, Inc. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- isolated late potential
- ventricular tachycardia
- Received July 7, 2012.
- Revision received January 8, 2013.
- Accepted February 3, 2013.
- American College of Cardiology Foundation
- de Bakker J.M.,
- van Capelle F.J.,
- Janse M.J.,
- et al.
- Stevenson W.G.,
- Khan H.,
- Sager P.,
- et al.
- Marchlinski F.E.,
- Callans D.J.,
- Gottlieb C.D.,
- Zado E.
- Soejima K.,
- Suzuki M.,
- Maisel W.H.,
- et al.
- Arenal A.,
- Gonzalez-Torrecilla E.,
- Ortiz M.,
- et al.
- Bogun F.,
- Good E.,
- Reich S.,
- et al.
- Haqqani H.M.,
- Kalman J.M.,
- Roberts-Thomson K.C.,
- et al.
- Arenal A.,
- del Castillo S.,
- Gonzalez-Torrecilla E.,
- et al.
- Dickfeld T.,
- Lei P.,
- Dilsizian V.,
- et al.
- Perez-David E.,
- Arenal A.,
- Rubio-Guivernau J.L.,
- et al.