Author + information
- Received March 22, 2011
- Revision received June 13, 2011
- Accepted June 20, 2011
- Published online September 20, 2011.
- Marc A. Miller, MD⁎,
- Srinivas R. Dukkipati, MD⁎,
- Alexander J. Mittnacht, MD†,
- Jason S. Chinitz, MD⁎,
- Lynn Belliveau, MD†,
- Jacob S. Koruth, MD⁎,
- J. Anthony Gomes, MD⁎,
- Andre d'Avila, MD, PhD⁎ and
- Vivek Y. Reddy, MD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Vivek Y. Reddy, Helmsley Electrophysiology Center, Mount Sinai School of Medicine, One Gustave L. Levy Place, Box 1030, New York, New York 10029
Objectives Our goal was to investigate the effects of percutaneous left ventricular assist device (pLVAD) support during catheter ablation of unstable ventricular tachycardia (VT).
Background Mechanical cardiac support during ablation of unstable VT is being increasingly used, but there is little available information on the potential hemodynamic benefits.
Methods Twenty-three consecutive procedures in 22 patients (ischemic, n = 11) with structural heart disease and hemodynamically unstable VT were performed with either pLVAD support (n = 10) or no pLVAD support (intra-aortic balloon pump counterpulsation, n = 6; no support, n = 7). Procedural monitoring included vital signs, left atrial pressure, arterial blood pressure, cerebral perfusion/oximetry, VT characteristics, and ablation outcomes.
Results The pLVAD group was maintained in VT significantly longer than the non-pLVAD group (66.7 min vs. 27.5 min; p = 0.03) and required fewer early terminations of sustained VT for hemodynamic instability (1.0 vs. 4.0; p = 0.001). More patients in the pLVAD group had at least 1 VT termination during ablation than non-pLVAD patients (9 of 10 [90%] vs. 5 of 13 [38%]; p = 0.03). There were no differences between groups in duration of cerebral deoxygenation, hypotension or perioperative changes in left atrial pressure, brain natriuretic peptide levels, lactic acid, or renal function.
Conclusions In patients with scar-related VT undergoing catheter ablation, pLVAD support was able to safely maintain end-organ perfusion despite extended periods of hemodynamically unstable VT. Randomized studies are necessary to determine whether this enhanced ability to perform entrainment and activation mapping will translate into a higher rate of clinical success.
A majority of patients with ventricular tachycardia (VT) in the setting of structural heart disease have unstable VT (1). As a result of this hemodynamic instability, less than one-third of patients are suitable for ablation guided only by mapping during VT (2), and with the greater use of effective coronary reperfusion during myocardial infarction, this number is further decreasing (3). Furthermore, most patients with scar-related VT have significant left ventricular dysfunction and other associated comorbidities that make them highly susceptible to the hemodynamic collapse and impairments in end-organ perfusion which can accompany unstable VT. As such, activation and entrainment mapping is difficult, if not impossible, for any extended period of time (1).
Theoretically, the use of mechanical cardiac assistance support allows for detailed mapping of VT in such patients, but there are limited data on the effects of these devices during VT mapping and ablation. The intra-aortic balloon pump (IABP) is the most commonly used temporary cardiac assist device, but this is a passive system that, in any event, has marginal effectiveness for hemodynamic support during VT. Specifically, low systolic blood pressure, elevated heart rate (>110 beats/min), and the need for accurate synchronization with the cardiac cycle (electrocardiographic gating) negatively affect IABP performance during VT (4). Conversely, continuous flow percutaneous left ventricular assist devices (pLVADs) do not appear to be limited by these factors and ultimately could facilitate mapping and ablation of ongoing VT, but there are limited data to support their use during VT ablation (5). We hypothesized that by maintaining near-normal hemodynamics and end-organ perfusion, a pLVAD would prove to be a viable option for mechanical cardiac assistance by allowing for a greater duration of mapping and ablation during ongoing VT.
All procedures were performed after obtaining written informed consent according to the institutional guidelines at the Mount Sinai Medical Center (New York, New York). This study was approved by our institutional review board.
From May 2010 to November 2010, a total of 23 consecutive VT ablations were performed in 22 patients with structural heart disease and who had at least 1 episode of hemodynamically unstable VT (defined as a mean arterial pressure <45 mm Hg). All patients were referred to us for further management of either sustained monomorphic ventricular tachycardia (MMVT) and/or recurrent implantable cardioverter-defibrillator (ICD) shocks. Patients underwent catheter ablation of VT with either pLVAD support (n = 10) or no pLVAD support (IABP counterpulsation, n = 6; no mechanical support, n = 7). Physician preference, with knowledge of the patient's underlying substrate and/or clinical VT characteristics, was the basis for using a particular percutaneous device (pLVAD vs. IABP). Peripheral arterial disease was the precluding factor in 57% (4 of 7) of the patients who did not receive any mechanical support; the remainder in this group did not receive mechanical support secondary to physician preference.
All procedures were performed under general anesthesia with endotracheal intubation. An indwelling arterial catheter was inserted into the radial artery for continuous blood pressure monitoring. Phenylephrine was used for blood pressure support as necessary. Cerebral tissue oxygen saturation (SctO2) was noninvasively measured and recorded with an absolute cerebral oximeter (FORE-SIGHT, CASMED, Branford, Connecticut). The technological background of SctO2 monitoring has been described elsewhere (6). Baseline SctO2 values were recorded 5 min after endotracheal intubation with an 80% to 100% fraction of inspired oxygen and then automatically every 2 s thereafter.
Noninvasive bispectral (BIS) electroencephalogram analysis (Aspect Medical Systems, Inc., Newton, Massachusetts) was used in all patients throughout the case. The BIS index is displayed as a numeric value ranging from 0 to 100; values of 40 to 60 are typically associated with adequate levels of anesthesia, but acute reductions from baseline may indicate cerebral hypoperfusion (7).
Cardiac assist devices
For patients treated with an IABP (Datascope Corporation, Mahwah, New Jersey), the catheter was inserted percutaneously through an 8-F introducer sheath into the femoral artery. Once correct positioning was confirmed (tip 2 to 3 cm from the left subclavian artery), the balloon was programmed to 1:1 assistance throughout the case. For patients treated with a pLVAD (Impella 2.5, ABIOMED, Inc., Danvers, Massachusetts), the catheter was inserted percutaneously through a 13-F sheath in the femoral artery (usually the left femoral artery). The device was positioned in the left ventricle as previously described (8), and the maximum performance level of 8 (1.9 to 2.5 l/min, 50,000 rpm) was maintained throughout the case, unless electromagnetic interference (EMI) (see Results section) necessitated temporary adjustments. In one case, the Impella LP 2.5 was implanted after programmed stimulation in a patient with overall preserved ventricular function but after inducing hemodynamically unstable VT; otherwise, in all cases of either pLVAD or IABP, the device was implanted at the beginning of the procedure.
The pre-close technique for arteriotomy site closure was used in all pLVAD cases (Perclose ProGlide, Abbott Laboratories, Abbott Park, Illinois) (9).
Electrophysiological evaluation and ablation strategy
Catheters were inserted in usual fashion and included a mapping catheter in the right ventricle and a 10-F linear phased-array intracardiac ultrasound probe (AcuNav, Biosense Webster, Inc., Diamond Bar, California). All mapping was performed with a hybrid magnetic/impedance-based electroanatomic mapping system (Carto 3, Biosense Webster, Inc.) and a 3.5-mm externally irrigated ablation catheter (Navistar ThermoCool, Biosense Webster, Inc.).
Left ventricular endocardial mapping was performed in 91% (21 of 23) of the procedures using a transseptal approach with an 8.5-F steerable sheath (Agilis, large curve, St. Jude Medical, St. Paul, Minnesota). Transseptal access was not obtained in 2 cases because endocardial mapping involved only the right ventricle. In the remaining 21 cases for which left atrial access was obtained, the mean left atrial pressure (LAP) was measured at the beginning and end of the case. If mechanical support was being used, the starting LAP was recorded at the beginning of the case after at least 5 min at the highest levels of support (IABP at 1:1 augmentation or Impella 2.5 at performance level 8 [50,000 rpm]). At the conclusion of the procedure, the mean LAP was again measured at the highest level of support. Epicardial mapping was performed in 12 cases (percutaneous subxiphoid puncture approach in 9 patients; surgical subxiphoid window in 3 patients) (10,11). Systemic anticoagulation with heparin was instituted to maintain an activated clotting time (ACT) of 250 to 300 s. The ACT target was not modified for the pLVAD group, as the recommended ACT level (>250 s) for the device used in this series falls within the same therapeutic range as our standard anticoagulation protocol for VT mapping and ablation (8). In cases in which epicardial access was obtained (always before the transseptal access or pLVAD insertion), anticoagulation was not administered until pericardial drainage/bleeding was absent.
Electrophysiology testing and radiofrequency catheter ablation were performed using standard techniques and included a combination of entrainment and/or activation mapping (if possible), pace mapping, and sinus rhythm–guided substrate modification. In all patients, routine laboratory tests and brain natriuretic peptide levels were measured immediately before and 1 day after the procedure. Arterial lactate levels were measured immediately before and at the conclusion of the procedure. The VT burden was defined as the total duration of all episodes of MMVT, lasting 10 s or more, during the course of the procedure. Burst ventricular pacing that did not result in MMVT was not included. Early terminations of MMVT, while mapping tachycardia, with either pace-termination or direct current cardioversion, were also quantified. Episodes of ventricular fibrillation and polymorphic VT (which were universally terminated by immediate cardioversion) were excluded. Recurrent VT during follow-up was defined as either sustained MMVT or VT/ventricular fibrillation–requiring ICD therapy within 3 months of the procedure.
Comparisons of continuous variables between groups with or without pLVAD support were made using the Student t test (mean ± SD) with the Levene test for equality of variance used to confirm homogeneity. Variables with non-normally distributed data were expressed as median (interquartile range), and statistical comparisons for these variables were made using the Mann-Whitney U test. Categorical variables were compared using the chi-square test, or, when fewer than 5 expected outcomes were expected per cell, the Fisher exact test. Integrated SctO2 (area under the curve) for SctO2 <55% was calculated as: minutes × desaturation points <55%. A relative decrease in SctO2 from baseline was calculated as: percent decrease from baseline = (baseline – current value)/baseline × 100. Two-sided p < 0.05 was considered indicative of statistical significance. Statistical calculations were performed using SPSS version 12.0 (SPSS Inc., Chicago, Illinois).
The baseline characteristics of the 2 groups are shown in Table 1. There was no significant difference between the groups with regard to age, history of hypertension, diabetes, ischemic cardiomyopathy, coronary artery bypass grafting, left ventricular ejection fraction, left ventricular end-diastolic diameter or baseline renal function. There was no difference between the 2 groups in pre-procedural antiarrhythmic medication use: Class III antiarrhythmic medications (most commonly, amiodarone) and beta-blockers were used in 90% and 70% of the entire patient cohort, respectively. There was no significant difference in pre-procedural heart failure symptoms between the pLVAD and non-pLVAD groups (NYHA functional class III or IV: 6 [60%] vs. 3 [23%], p = 0.10).
Electrophysiological study and VT burden
As shown in Table 2, epicardial access was obtained in 48% (11 of 23) of all cases and was similar between the 2 groups (n = 6 [60%] in the pLAVD group vs. 5 [39%] in the non-pLVAD group; p = 0.41). The duration of the procedure was significantly longer in the pLVAD group compared with the non-pLVAD group (528.3 min vs. 407.3 min; p = 0.01). Two or more VT morphologies were induced in 96% (22 of 23) of the procedures, and there was no significant difference between the pLVAD and non-pLVAD groups with regard to the number of inducible MMVTs (3.4 vs. 2.8; p = 0.26) or mean tachycardia cycle length (395.1 ms vs. 395.8 ms; p = 0.99) (Table 2). The pLVAD group had a significantly higher VT burden (defined as the total time in VT during the procedure) compared with the non-pLVAD group (66.7 min vs. 27.5 min; p = 0.03) and required fewer early terminations of sustained MMVT for hemodynamic instability during entrainment/activation mapping (1.0 vs. 4.0; p = 0.001). Representative examples of MMVT with and without pLVAD support are shown in Figures 1 and 2⇓, respectively.
VT ablation outcome
There were a total of 34 VTs in 10 patients and 38 VTs in 13 patients in the pLVAD and non-pLVAD groups, respectively. Of these, 21 VTs (13 in the pLVAD group and 8 in the non-pLVAD group) were terminated by radiofrequency energy delivery during ongoing VT (i.e., ablation during VT). On a per-patient basis, at least 1 VT was terminable by energy delivery during ongoing VT in 9 of 10 (90%) patients and 5 of 13 (39%) patients in the pLVAD and non-pLVAD groups, respectively (p = 0.03) (Table 2). There was no difference in total duration of radiofrequency ablation in the pLVAD and non-pLVAD groups (27.9 ± 13.6 min vs. 21.8 ± 9.2 min [p = 0.23], respectively). Inducibility for VT was assessed by programmed ventricular stimulation at the end of the procedure in 87% (20 of 23) of the entire cohort (8 of 10 of the pLVAD group and 12 of 13 of the non-pLVAD group). There was no difference in the number of patients who remained inducible for MMVT in the pLVAD and non-pLVAD groups (2 of 8 [25%] vs. 4 of 12 [33%]; p ≥ 0.05, respectively). During follow-up, 3 (30%) patients in the pLVAD group and 4 (31%) patients in the non-pLVAD group had recurrences of sustained MMVT and/or appropriate ICD therapy (p = 0.97).
Hemodynamics and surrogates of end-organ perfusion
Although the mean arterial pressure was similar in both groups, systolic blood pressure was lower and diastolic pressure was higher in the pLVAD patients compared with the non-pLVAD patients (105.5 mm Hg vs. 117.9 mm Hg [p = 0.04] and 62.8 mm Hg vs. 50.5 mm Hg [p = 0.002], respectively) (Table 3). Baseline mean LAP was increased for the entire cohort (14.2 ± 4.8 mm Hg), and there was no statistical difference between the pLVAD group and the non-pLVAD group with regard to the post-procedure elevations in mean LAP (2.5 ± 4.4 mm Hg vs. 5.4 ± 7.7 mm Hg; p = 0.32). Similarly, there was no statistical difference between the 2 groups in the peri-procedural changes in brain natriuretic peptide levels (505 [−73 to 732] vs. 221 [152 to 730]; p = 0.76), arterial lactate levels (0.5 ± 0.7 mmol/l vs. 0.4 ± 0.5 mmol/l; p = 0.58), or renal function (glomerular filtration rate −3 ± 13 vs. 5 ± 11; p = 0.11).
Duration of cerebral tissue deoxygenation was similar for both groups, whether comparing the integrated (area under the curve) SctO2 <55%, total duration (in minutes) of SctO2 <55%, or relative reductions of ≥20% from postintubation baseline values (Table 3). There was no difference between the mean BIS values in the pLVAD group compared with the non-pLVAD group (52.6 vs. 49.6; p = 0.40) or the duration the BIS value was <30 (3.6 min vs. 6.7 min; p = 0.69).
EMI during VT entrainment mapping was classified as follows: mild if noted by the electroanatomic mapping system but no intervention required; moderate if it necessitated reducing the pLVAD performance level; or severe if it required temporarily turning off the pLVAD. Mild EMI occurred in 6 cases, moderate in 3, and severe in 1. In cases with mild EMI, entrainment mapping was continued. For the moderate EMI cases, pLVAD performance level was titrated down, generally from 50,000 rpm (1.9 to 2.5 l/min) to 35,000 rpm (0.4 to 10.0 l/min), which either resolved or decreased EMI enough to continue mapping (Fig. 3). In the isolated case of severe interference, the pLVAD was temporarily turned off, which resolved the EMI. Magnetic interference did not have a discernible effect on the pLVAD performance.
There were 2 major complications (9%) in the entire cohort, both related to pericardial tamponade in patients who had undergone epicardial mapping/ablation. One patient in the pLVAD group developed pericardial tamponade approximately 6 h into the procedure, due to a 5-mm laceration on the anterior surface of the right ventricle. Although the pLVAD itself was not the cause of the myocardial laceration, one cannot rule out the possibility that a contributing factor to the pericardial effusion was some degree of hesitancy by the operator to rapidly reverse systemic anticoagulation due to the pLVAD. Conversely, 1 patient in the IABP group also developed pericardial tamponade approximately 30 min after completion of the procedure. This was discovered after removal of the IABP and the transseptal and epicardial access sheaths, at which time protamine had already been administered for reversal of systemic anticoagulation. In this latter case, there was no source of bleeding identified during surgical exploration. Both patients underwent successful sternotomies and continue to do well >3 months after the procedures. There were no vascular access site injuries or groin hematomas in any of the patients that required either intervention, holding of anticoagulation therapy, or transfusion of blood products.
This is the largest study on the effects of mechanical support for hemodynamic maintenance during mapping and ablation of unstable VT and the first to evaluate the use of noninvasive neuromonitoring modalities such as cerebral oximetry in this setting as an indicator of end-organ perfusion during VT ablation.
The pLVAD group was maintained in VT nearly 2.5 times longer than the non-pLVAD group and required less premature terminations of ongoing VT. Importantly, the significantly greater duration of VT was possible without any comparative evidence of tissue hypoperfusion or end-organ damage.
The retrospective nature of this study precludes robust conclusions on the effectiveness of pLVAD use during VT ablation. However, there were some very interesting observations in this study that suggest benefit. First, despite a significantly greater duration of time spent during VT in the pLVAD group (66.7 min vs. 27.5 min; p = 0.03), there were no differences in the post-procedural measurable outcomes such as renal function or LAP. Second, there was a significantly diminished need for early VT termination of sustained VT for hemodynamic instability in the pLVAD group (0.9 vs. 3.9 terminations/procedure; p ≤ 0.001). This was likely a reflection of the hemodynamic maintenance-capabilities afforded by the pLVAD. Finally, the majority of the patients in the pLVAD group (90%) had at least 1 VT that was terminable by radiofrequency energy delivery during ongoing VT. This finding compared favorably to the 38% value observed in the non-pLVAD group. In toto, these observations provide the scientific basis for a prospective randomized controlled trial of the use of pLVAD during catheter ablation of scar-related VT.
From a feasibility perspective, interference between the electroanatomic mapping system and the pLVAD was not prohibitive during the majority of the cases. Not surprisingly, the EMI was greatest during mapping in the ventricular outflow tracts because of the proximity to the magnetic motor of the pLVAD device. It is also important to note that all of the left ventricular mapping in this study was performed using a transseptal approach. In other procedures we have performed using a retrograde mapping approach since this patient cohort, we have noted additional electrical noise on the mapping electrogram channels.
Safety of pLVAD
The major potential safety issues one might expect with the pLVAD are avulsion of the aortic valve, vascular injury at the point of insertion into the femoral artery, hematoma/pseudoaneurysm/retroperitoneal bleeding after device removal, and stroke/systemic embolism (12). The absence of these device-related complications in this series is reassuring. Use of intracardiac echocardiography during pLVAD insertion, the suture-based technique to pre-close the arteriotomy site, and close attention to anticoagulation likely contributed to the good safety profile observed in this series.
Hemodynamics and cerebral oximetry
Pulse oximetry monitoring, which requires pulsatile flow and adequate peripheral perfusion, is a relatively late warning sign of deoxygenation (13). However, the utility of SctO2 (which does not require pulsatile flow) for the early detection of impaired tissue oxygenation and perfusion has been demonstrated in a variety of clinical settings (14,15). By using the brain as an index organ, interventions to improve cerebral oxygenation had systemic benefits. Similarly, we propose that during VT ablation, SctO2 may serve as a better marker of hemodynamic stability than blood pressure. Limitations of this technique at this point are that safe lower limits have not been defined, particularly in the VT ablation setting. In our series, although a strict protocol by which to prematurely terminate VT was not used, we generally consider early termination of VT if the SctO2 value trends below 55% (16). Although established hemodynamic monitoring modalities such as arterial blood pressure and pulse oximetry should still be included in the decision-making process, advances in neuromonitoring provide valuable information of adequate tissue perfusion during unstable VT.
By virtue of the retrospective nature of this study, we were unable to quantify the true difference between the 2 groups with respect to duration of entrainment mapping/activation, only the length of time the patients were in VT. Furthermore, the nonrandomized assignment of patients to 1 particular class of support was based on physician preference, and in 4 cases no support was used secondary to peripheral arterial disease. The relatively small number of patients and abbreviated follow-up period precludes a relevant comparative assessment of clinical outcome. Regarding the neuromonitoring with cerebral oximetry and BIS, this retrospective study does not attempt to establish safe limits that can be seen as lower thresholds for VT termination. A fundamental question raised by this research is whether a combined approach of entrainment mapping followed by substrate modification (the strategy facilitated by the pLVAD) is truly superior to a purely substrate modification strategy. In addition, it is important to recognize both the additional time required to implant and remove the pLVAD (approximately 10 to 30 min, depending on operator experience), and the additional cost it adds to the procedure (approximately $25,000). Whether these additional factors are overall in favor or against use of the pLVAD for VT ablation can only be definitively evaluated by prospective randomized trials.
In patients with scar-related VT, percutaneous LVAD support is able to safely maintain end-organ perfusion despite extended periods of hemodynamically unstable VT to allow for detailed activation and entrainment mapping. As identified by termination by ablation during ongoing VT, this translated to a high rate of identification of the critical VT isthmus. This provides the scientific basis for future prospective randomized clinical outcome studies.
Drs. Reddy and d'Avila have received consulting fees from Biosense Webster, the manufacturer of the electroanatomic mapping system used in this series. Dr. Reddy has received grant support (albeit not for this study) from ABIOMED, Inc., the manufacturer of the percutaneous left ventricular assist device. Drs. Miller, d'Avila, and Reddy have received speaker honoraria from ABIOMED, Inc.
All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- activated clotting time
- electromagnetic interference
- intra-aortic balloon pump
- implantable cardioverter-defibrillator
- left atrial pressure
- monomorphic ventricular tachycardia
- percutaneous left ventricular assist device
- cerebral tissue oxygen saturation
- ventricular tachycardia
- Received March 22, 2011.
- Revision received June 13, 2011.
- Accepted June 20, 2011.
- American College of Cardiology Foundation
- Wijnmaalen A.P.,
- Schalij M.J.,
- von der Thusen J.H.,
- Klautz R.J.,
- Zeppenfeld K.
- Fischer G.W.
- Dixon S.R.,
- Henriques J.P.,
- Mauri L.,
- et al.
- Soejima K.,
- Couper G.,
- Cooper J.M.,
- Sapp J.L.,
- Epstein L.M.,
- Stevenson W.G.
- Sjauw K.D.,
- Konorza T.,
- Erbel R.,
- et al.
- Tobias J.D.
- Murkin J.M.,
- Arango M.