Author + information
- Received October 15, 2015
- Revision received January 12, 2016
- Accepted January 14, 2016
- Published online June 1, 2016.
- Yuhning L. Hu, MDa,
- Vigneshwar Kasirajan, MDb,
- Daniel G. Tang, MDb,
- Keyur B. Shah, MDa,
- Jayanthi N. Koneru, MDa,
- John D. Grizzard, MDc,
- Kenneth A. Ellenbogen, MDa and
- Jordana Kron, MDa,∗ ()
- aDivision of Cardiology, Pauley Heart Center, Virginia Commonwealth University, Richmond, Virginia
- bDivision of Cardiothoracic Surgery, Pauley Heart Center, Virginia Commonwealth University, Richmond, Virginia
- cDivision of Radiology, Pauley Heart Center, Virginia Commonwealth University, Richmond, Virginia
- ↵∗Reprint requests and correspondence:
Dr. Jordana Kron, Pauley Heart Center, Virginia Commonwealth University, 1200 East Marshall Street, 3rd Floor, PO Box 980053, Richmond, Virginia 23298-0053.
Objectives This study investigated the mechanism of lead malfunction by monitoring lead parameters throughout left ventricular assist device (LVAD) implantation.
Background Implantable cardioverter-defibrillator (ICD) lead malfunction can occur after LVAD implantation.
Methods ICD lead data were prospectively evaluated during and after LVAD implantation and at 12 pre-specified intraoperative time points.
Results We prospectively evaluated 32 patients with ICDs who underwent LVAD implantation, of whom 20 patients underwent serial testing at 12 intraoperative steps. Post-operative right ventricle (RV) sensing had decreased by >50% from baseline in 7 patients (22%), with RV sensing improving at 1 to 7 weeks in 2 patients (28.6%). Nine patients (28.1%) had >10-ohm (Ω) high-voltage (HV) impedance changes from baseline to final impedance. In all 5 patients with >50% decrease in RV sensing and all 7 patients with a >10-Ω HV impedance change who underwent intraoperative testing, changes were not detected until after weaning from cardiopulmonary bypass. Patients with decreased RV lead sensing >50% (n = 7) had lower glomerular filtration rates (48.7 ± 21.9 ml/min/1.73 m2 vs. 68.4 ± 22.5 ml/min/1.73 m2, respectively, p = 0.0489), were more likely to have undergone concomitant RVAD placement (42.9% vs. 0%, respectively, p = 0.0071), concomitant tricuspid valve surgery (57.1% vs. 16%, p = 0.0469), or to have had cardiac tamponade or unplanned return to the operating room (57.1% vs. 12%, p = 0.0258).
Conclusions ICD lead malfunction can occur following LVAD implantation but may improve over time. Intraoperative RV sensing and HV impedance changes were not detected until after weaning from cardiopulmonary bypass, suggesting the mechanism of RV lead malfunction may be related to LV unloading and concomitant leftward septal shift. A conservative approach is warranted in many patients with ICD parameter changes after LVAD implantation because parameter abnormalities may improve over time. (Implantable Cardioverter Defibrillator (ICD) Function During Ventricular Assist Device (VAD) Implantation; NCT01576562)
Ventricular assist devices (VAD) have become an important therapeutic option for patients with advanced heart failure. Studies have shown that implantable cardioverter-defibrillator (ICD) lead parameters can change following LVAD, but the mechanism is unknown. One retrospective study showed significant changes in all right ventricle (RV) lead parameters before and after LVAD implantation, with 13% of patients requiring lead revisions and 20% requiring ICD testing (1). Additional studies have found significant alterations in ICD lead parameters following LVAD implantation and adverse events related to ICD malfunction requiring intervention (2–4). We sought to elucidate the mechanism of lead malfunction by monitoring lead parameters throughout LVAD surgery. Several plausible mechanisms could explain ICD lead parameter changes after LVAD implantation, including mechanical disruption of the leads themselves from insertion of the LVAD; inflammatory and metabolic changes during cardiopulmonary bypass; and alteration in ventricular geometry due to LV decompression and septal shift (2). The purpose of the present study was to prospectively evaluate ICD lead function at 12 time points during LVAD implantation to provide insight into the mechanism of malfunction.
In this prospective single-center study at Virginia Commonwealth University (Richmond, Virginia), we evaluated 32 patients with pre-existing ICDs who underwent LVAD implantation from May 1, 2011 to January 30, 2014, of whom a subgroup of 20 patients underwent serial intraoperative testing. Inclusion criteria included patients ≥18 years of age; ICD wireless technology, so that interrogation throughout surgery would not interfere with the surgical field; and willingness of patient or next of kin to provide informed consent. Patients were excluded if they were pacemaker-dependent or unwilling to provide informed consent. This study was approved by the institutional review board of the Office of Research at the Virginia Commonwealth University and is registered at NCT01576562.
ICD. Patients with single- or dual-chamber or biventricular ICDs were included. Device manufacturers included Boston Scientific (Natick, Massachusetts), Medtronic (Minneapolis, Minnesota), and St. Jude Medical (Sunnyvale, California).
Thirty-two patients underwent continuous LVAD implantation performed by two cardiothoracic surgeons (V.K., D.T.). Thirty-one patients received axial flow LVAD implants (HeartMateII, Thoratec Corp., Pleasanton, California), and 1 patient received a centrifugal-flow LVAD implant (HeartWare, Inc., Framingham, Massachusetts). Twelve-point intraoperative data were collected from a subgroup of 20 patients undergoing LVAD implantation with the axial flow device. Implantation of the axial flow LVAD proceeded as follows: a median sternotomy was performed; a VAD pocket was created in the peritoneal space, and a tunnel to the right rectus was created for the device driveline; cardiopulmonary bypass (CPB) was performed after cannulation of the aorta and right atrium (RA); an LV outflow graft was sewn to the ascending aorta, the LV apex was cored, and the LVAD inflow cannula was attached; CPB was weaned, the pericardium was reconstructed, and the chest was closed. One patient (#27) underwent LVAD implantation with the centrifugal flow LVAD, and standard surgical fashion proceeded as follows: a median sternotomy performed, aorta and RA were cannulated, CPB was implemented, a sewing ring was attached to the LV apex, the apex was cored, the LVAD pump was implanted and secured in the pericardial space, the LV outflow graft was sewn to the ascending aorta, a tunnel to left rectus was created for the driveline, CPB was weaned, the pericardium was reconstructed, and the chest was closed.
Data collection and endpoints
Sensing, impedance, and thresholds were tested at baseline for each lead. In 20 patients, intraoperative sensing and impedances were collected serially at 11 additional time points during LVAD implantation, including induction of anesthesia, median sternotomy, creation of VAD pocket, initiation of cardiopulmonary bypass, sewing of LVAD outflow graft to ascending aorta, coring of LV apex, attachment of inflow cannula, weaning of cardiopulmonary bypass, reconstruction of pericardium, closing of sternum, and final parameters. Pacing threshold was not tested intraoperatively, to prevent any potential hemodynamic instability related to ventricular pacing. Lead measurements were made after each operative step was completed. Post-operative data recorded at 1 week and at 3 to 12 months included complete ICD interrogation, occurrence of arrhythmia, antiarrhythmic medications, lead revisions, ICD programming changes, and generator changes. Patients were followed until heart transplantation, death, or the end of follow-up period (October 2015).
ICD lead malfunction
Lead malfunction for RA, RV, and LV leads was defined as a decrease in sensing >50% from baseline or an increase in threshold >50% from baseline. Sensing that decreased by >50% from baseline but subsequently increased to ≥50% of the baseline value was considered improved. A pacing threshold that increased by >50% from baseline was considered improved if it returned to ≤150% of the baseline value. RV pacing lead impedance was considered significantly changed if it increased or decreased by >100 Ω from the baseline value to the final value at the end of LVAD implantation. HV impedance was considered significantly changed if it increased or decreased by >10 ohm (Ω) from the baseline vale to the final value at the end of LVAD implantation. These definitions were used to help determine the mechanism of lead malfunction and were not intended to reflect a change that was necessarily clinically significant.
The intrinsic QRS interval in sinus rhythm was measured using electrocardiography (ECG) before and after LVAD implantation. QRS measurements were made by computer and confirmed manually.
Chest radiography analysis
An independent radiologist blinded to the clinical and device outcomes of the patients reviewed chest radiographs (CXR) obtained immediately prior to and immediately following LVAD implantation. Lead positions on the CXRs were compared for any change in conformation or gross dislodgement. “Bowing” was defined as a mild upward curvature on the septum, with no acute angulation or evidence of fracture.
Clinically relevant lead malfunction was defined as revision of a lead or turning off of a lead due to noncapture or undersensing. The time to death or orthotopic heart transplantation was recorded in October 2015.
Fisher exact test was used for categorical variables to compare the rates of occurrence between the groups. Student t test was used for continuous variables that were normally distributed. Wilcoxon rank test was used for continuous variables that were not normally distributed. A p value of ≤0.05 was considered significant. Kaplan-Meier curves were created based on survival time to death or cardiac transplantation. Undergoing cardiac transplantation was counted as a censored observation.
Data were collected prospectively in 32 patients who underwent LVAD implantation 50.0 ± 46.8 months after ICD implantation. Of the 32 patients, 22 were male (68.8%), 19 had coronary artery disease (59.3%), 25 had LVAD implantation as destination therapy (78.1%), and mean age was 55.8 ± 13.9 years (Table 1).
Long-term lead outcomes
RV lead sensing
RV leads were tested in 32 patients. RV sensing decreased by more than 50% from baseline in 7 patients (22%) immediately post-LVAD, and RV sensing improved over time in 2 patients (28.6%). RV sensing had improved after 1 week in 1 patient and after 7 weeks in 1 patient (Table 2, Figure 1).
RV lead pacing thresholds
In 2 patients (6.3%), who also had decreased sensing, RV pacing threshold was >50% elevated from baseline immediately post-operatively and was persistently elevated at 1 week (Online Table 1). One RV pacing threshold returned to baseline at 3 month (50%).
RV lead impedances
Average RV pacing impedance was 429.3 ± 46.5 Ω at baseline, 447.6 ± 94.2 Ω at the end of LV implantation (p = 0.59), 418.0 ± 49.5 Ω at 1 week (p = 0.85), and 449.7 ± 85.9 Ω at 3 to 12 months follow-up (p = 0.63). p Values for baseline impedance values are shown for comparison. Average RV HV impedance was 48.9 ± 13.8 Ω at baseline, 41.1 ± 11.2 Ω at the end of LV implantation (p < 0.0001), 38.9 ± 9.2 Ω at 1 week (p < 0.0001), and 44.4 ± 9.6 Ω at 3 to 12 months follow-up (p < 0.0001). None of the 32 leads tested had RV pacing impedances that were out of normal range for the lead before, during, or after LVAD implantation. None of the 32 leads tested had HV impedances that were out of the expected range for the given lead at any point.
Twenty-two RA leads were tested. RA sensing was decreased by more than 50% in 4 patients (18.2%) immediately post-LVAD with RA sensing improved at 1 week in 3 patients (75%). None of the patients with a decrease of more than 50% in atrial sensing had significant changes in RA pacing threshold, defined as an increase of more than 50% in RA sensing. Sensing during atrial fibrillation or flutter was not compared with sensing during sinus rhythm. Three patients (13.6%) had an increase >50% in RA pacing threshold, 2 of whom (66.7%) had atrial undersensing at 1 week follow-up.
LV lead sensing was performed in 10 devices, and threshold testing was performed in 15 leads (Online Table 2). Six patients (40%) had increases in threshold, 2 (33.3%) of whom improved over time. One LV lead had both a decrease of >50% in LV sensing threshold and an increase of >50% in LV pacing threshold immediately post-operatively, and leads were turned off at 1 week for noncapture at maximum output. Both the sensing and capture thresholds were noted to have improved at the third and 12th month follow-up interrogations in this patient. One LV lead was turned off at 1 week due to noncapture at maximum output and then found to have normal sensing and capture threshold at day 121 follow-up.
Intraoperative lead testing
Twenty patients underwent testing at 12 serial intraoperative time points. Intraoperative measurements were recorded at the completion of each operative step. Of the 5 patients with significant decreases in RV sensing, in whom lead function was tested at serial time points during VAD implantation, a persistent RV sensing decrease >50% was detected in all cases after weaning from CPB (immediately after weaning from CPB in 1 patient, after closing the chest in 3 patients, and on final testing in the operating room in 1 patient) (Table 2, see ∗). At 1 week, 3 of the 5 patients (60%) had persistently low R waves; in 1 patients (20%), R waves returned to baseline; and 1 patient had died (20%).
We identified 6 patients (18.8%) who had a pacing impedance change >100 Ω from baseline to final impedance during LVAD implantation, 3 of whom also had >50% decrease in RV lead sensing. Three of these patients had impedance levels tested at serial time points during the surgery. There was no consistent pattern of change in pacing impedance (Online Table 3).
We identified 9 patients (28.1%) who had >10-Ω HV impedance change from baseline to final impedance during LVAD implantation, 2 of whom also had >50% decrease in RV lead sensing. Seven of these patients had HV impedances tested at serial time points during the surgery. All seven patients first had >10-Ω impedance change detected after weaning from CPB. (Table 3, see ∗).
The intrinsic QRS duration measured from the surface ECG decreased after LVAD implantation (123.9 ± 31.9 ms vs. 98.8 ± 32.0 ms, respectively; mean difference: 25.1 ms, p < 0.05; 95% CI: 18.4 to 31.8) (Figure 2).
Chest radiography analysis
CXRs read immediately before and after LVAD implantation by an independent, blinded radiologist revealed that, of 32 RV defibrillator leads, no leads were grossly dislodged (Figure 3). Ten RV leads (31.3%) showed upward bowing of the lead following LVAD implantation, 2 of which also showed kinking of the lead. Three of 7 leads (42.9%) that had >50% decrease in RV sensing had bowing on CXR compared with bowing in 7 of 25 leads (28%) that did not have a decrease of >50% in RV sensing (p = 0.6479). One of 2 leads with kinking had a decrease of > 50% in RV sensing. Of the 2 leads that had an elevated RV pacing threshold, 1 (50%) had upward bowing on CXR.
No RA lead (n = 22) showed any change in CXR location pre- and post-LVAD implantation. No LV lead (n = 15) showed any radiographic changes pre- and post-LVAD implantation.
Clinical outcomes related to lead dysfunction
One RV lead was turned off (Patient #8), and 1 was revised (Patient #22) due to noncapture at maximal output. One patient (#25) with a decrease of >50% in RV sensing threshold was found to have 3-cm chronic ICD lead thrombus intraoperatively and underwent debridement and removal of ICD lead thrombus.
One RA lead was turned off due to undersensing and noncapture. One patient (#8) had a known dislodged RA lead from a right heart catheterization prior to LVAD implantation.
One LV lead (Patient #22) was turned off intraoperatively due to noncapture and 2 LV leads (Patients #13 and #32) were turned off at 1 week due to noncapture. During a 121-day device check for Patient #32, the LV lead was found to be capturing and functioning appropriately.
Predictors of RV lead dysfunction
Clinical, surgical, echocardiographic, and hemodynamic parameters of patients with >50% decrease in RV lead sensing (n = 7) were compared with those of patients without >50% decrease in RV lead sensing (n = 25) (Table 4). Mean time from lead implant to LVAD was 87.8 months (range: 1.5 to 94 months) in the patients with decreased RV lead sensing (n = 7) versus 30.4 months (range: 0.5 to 171) in patients with no decreased RV sensing (n = 25). Patients with decreased RV lead sensing >50% had lower GFR (48.7 ± 21.9 ml/min/1.73 m2 vs. 68.4 ± 22.5 ml/min/1.73 m2, respectively, p = 0.0489) and were more likely to have undergone concomitant RVAD placement (42.9% vs. 0%, respectively, p = 0.0071), concomitant tricuspid valve (TV) surgery (57.1% vs. 16%, respectively, p = 0.0469), or to have had cardiac tamponade or unplanned return to the operating room (57.1% vs. 12%, respectively, p = 0.0258). There were no significant differences in true bipolar sensing versus integrated bipolar sensing between leads that had >50% decrease in RV sensing and those that did not.
Stroke occurred in 3 patients (9%). Pump thrombosis occurred in 2 patients, 1 of whom underwent LVAD replacement within 30 days of initial implantation. A massive pulmonary embolism was found in one patient while undergoing an emergent mediastinal reexploration.
Mortality and transplantation
Of 32 patients, 13 (40.6%) died during the follow-up period (through October 2015). Nine patients (28.1%) underwent orthotopic heart transplant, 1 of whom died 3 days later. Four of 7 patients (57.1) with >50% decrease in RV lead sensing died compared with 9 of 25 patients (36.0%) patients who did not have a >50% decrease in RV lead sensing (Figure 4). Three patients (8.1%) died within 30 days post-LVAD.
In this prospective study, we evaluated lead parameters at serial time points intraoperatively during LVAD implantation to try to elucidate the mechanism of ICD lead malfunction following LVAD. We found that ICD lead parameters may change during and after LVAD implantation, and these acute changes may improve over weeks to months. Intraoperative RV sensing and HV impedance changes were not detected until after weaning of cardiopulmonary bypass, suggesting the mechanism of RV lead malfunction may be related to LV unloading and concomitant leftward septal shift.
ICD lead parameter changes
Prior studies have evaluated ICD performance post-LVAD and found that sensing and pacing parameters can change post-operatively. Three retrospective studies showed statistically significant deterioration of RV sensing thresholds post LVAD implantation (1–4). Median RV lead sensing threshold decreased by 21% after continuous flow LVAD with 1 in 5 patients having sensing amplitudes of 5 mV or less (3). In our study, 22% of patients had a decrease of more than 50% in R-wave sensing threshold. Decreased RV lead sensing is of concern because it may lead to underdetection of tachyarrhythmia and delayed appropriate therapy. Significant increases in RV pacing threshold have not been consistently observed in previous studies. One study showed a significant increase in RV pacing threshold 6 months after LVAD implantation (2). In 6% of our patients, RV pacing threshold increased by more than 50%.
Possible mechanisms of ICD lead parameter change
Although proposed in previous reports, mechanical disruption of the leads during surgery did not explain our findings. We found no radiographic changes in positions of the leads before and after VAD implantation (Figure 3). There was no gross radiographic dislodgement observed in any RA, RV, or LV leads by a blinded radiologist comparing pre- and post-LVAD chest radiographs. Although bowing was observed in 31% of RV leads, this finding did not occur significantly more often in leads that had >50% reduction in R wave sensing. If frank lead dislodgement were the mechanism of dysfunction, newly placed leads would potentially be more at risk for dysfunction than long-term leads. In fact, RV leads with a decrease of >50% in sensing threshold were implanted for a mean 88 months, longer than the leads with no significant decrease in sensing threshold, mean 30 months.
By prospectively recording ICD lead parameters frequently throughout and following VAD implantation, we localized the time of RV lead sensing changes to after weaning the patients from CPB. Interestingly, all patients who had a change of >10-Ω HV impedance also had the change first detected after weaning from CPB. We hypothesize that ICD lead parameter changes following LVAD are due to unloading of the LV and concomitant leftward shifting of the septum, altering the ICD lead/myocardium conformation. However, due to the small number of patients in the study, the mechanism cannot be determined definitively. Myocardial edema and ischemia associated with CPB could contribute to the altered lead properties. CPB induces a well-described systemic inflammatory response that can lead to multiorgan system failure (5). The inflammatory response is due to surgical trauma, activation of blood components in the extracorporeal circuit, ischemia and reperfusion injury, and endotoxin release (6). Inflammatory markers are increased as early as 1 hour after initiation of CPB (7). ICD lead parameter changes related to systemic inflammation have not been described in published reports.
We identified a higher prevalence of reduced GFR, TV repair, and RVAD placement in those with reduced RV lead sensing. Patients with renal dysfunction may have greater fluid shifts during the perioperative period. Patients requiring RVAD support or concomitant TV surgery may have more baseline RV dysfunction and greater post-procedure changes in RV hemodynamics. When TV repair was performed, the repair and RVAD implantation were done after LVAD implantation, prior to weaning from CPB. ICD lead malfunction in some patients could be due directly to the TV repair or RVAD implantation. Cardiac tamponade and return to the operating room are variables that may identify a sicker subset of patients with prolonged cardiopulmonary bypass time, who may have required more resuscitative fluids and may have had more perioperative hemodynamic changes.
The decrease in LV size and septal shift that occur after LVAD implantation are demonstrated on pre- and post-CT scans in one patient (Figure 5A) and transesophageal echocardiography images on another patient (Figure 5B), neither of whom had significant changes in ICD parameters. The impact of these multiple changes may be unmasked during weaning from CPB, when venous return diverts back to the heart. For bipolar ICD leads, sensing occurs between the distal and proximal electrodes (true bipolar) or the distal electrode and the distal defibrillator coil (integrated bipolar). A sensing change of >50% occurred similarly in true bipolar and integrated bipolar sensing, suggesting that the distal electrode likely plays a role in the sensing changes (Table 4). We speculate that the shifting of the interventricular septum toward the decompressed LV changes the orientation of the interface between the distal electrode and the septal myocardium. As ventricular remodeling occurs post-LVAD, the distal electrode and the septal myocardium undergo further changes in orientation, which may lead to further changes in RV sensing. Further research is needed to evaluate the RV and LV changes that occur after LVAD implantation.
The effect of LVAD support on right and left atrial size and function is not known. One study showed that severity of mitral regurgitation decreased significantly post-LVAD, with 76% of patients having moderate to severe mitral regurgitation pre-operatively compared with 8% with moderate or severe mitral regurgitation at 1 month post-LVAD (p < 0.001) (8). The post-LVAD improvement in mitral regurgitation likely changes left atrial size, ultimately affecting the interatrial septum and RA size and function. The effect of LVAD implantation on atrial size and hemodynamic function requires further investigation.
ICD lead parameter improvement over time
Importantly, lead dysfunction improved over time in some patients. Almost 30% of patients with decreased RV sensing improved over time, and 50% of patients with elevated RV threshold improved over time. Of 4 patients with impaired RA sensing, 75% improved at 1 week. Of 6 LV leads with increased pacing threshold, 33% improved over time. The frequent peri- and post-operative testing performed in this study allowed us to identify fluctuations in lead parameters that have not been previously described. The return of some parameters to baseline over the weeks and months following VAD highlights the need for careful monitoring and cautious decision-making prior to lead interventions. Such improvements over time may be due to geometry changes and further left ventricular remodeling. However, pharmacologic, metabolic, and endocrinologic changes that are rapidly occurring during the post-operative period may also play a role. Changes due to LV apical myocardial mass excision or myocardial scarring and fibrosis would not be expected to lead to fluctuating lead parameters.
In this prospective study, 7 patients (22%) had RV lead malfunctions post-LVAD, 2 of which (6.3%) required interventions (1 RV lead was turned off, and 1 was replaced due to noncapture). In previous studies, 13 to 18% of RV leads post-LVAD implant required revision (1,3). For most patients, VADs are a bridge to transplantation or destination therapy. In either case, systemic infection relating to ICD infection can be catastrophic. The decision to revise an ICD lead should be made only after careful consideration of the potential risks and benefits to the individual patient. In one study of six patients with LVAD associated infections, patients had high rates of recurrent infections and mortality despite ICD electronic device removal (9). Factors such as pacemaker dependence and history of ventricular arrhythmias must be considered when deciding whether to replace or revise a defibrillator lead. Data suggest that mortality, hospitalization rates, and arrhythmia rates were not improved by cardiac resynchronization therapy in patients with an LVAD (10). The risks of revising a nonfunctional LV lead may outweigh the benefits in many patients. Our data suggest that a conservative approach is warranted in many patients with ICD system malfunction after LVAD implantation, because there is a potential that the acute decline in lead parameters might regress over a period of time.
While there was no significant difference in survival between patients who had a significant change in RV lead sensing and patients who did not, there was a trend toward shorter survival among patients who had changes in RV lead sensing. A significant difference may not have been detected because of the small number of patients in each group. Larger studies are needed to determine whether ICD lead parameter changes identify a subset of LVAD patients with worse outcomes.
We found that intrinsic QRS duration decreased after LVAD implantation. We believe the QRS duration decrease post LVAD implantation is due to reverse electrical cardiac remodeling. One prior study prospectively evaluated the effects of LVAD implantation on electrocardiograms in 12 idiopathic dilated cardiomyopathy patients who underwent LVAD (11). The investigators found that LVAD support caused progressive shortening of QRS and QTc intervals, consistent with reverse remodeling of the cardiac electrical properties. In this study, the QRS duration shortened significantly within the first week, but continued to decrease for up to 6 months of LVAD support. These findings need to be reproduced in larger studies.
This was a moderately sized single-center study. The small number of patients in the study limits interpretation of the statistical analysis. VAD implantations were performed by 2 experienced surgeons and may not represent the experience at other centers. Pacemaker-dependent patients were excluded from the study and may have different ICD parameter changes following VAD implant. ICDs with wireless technology were required, and older devices may function differently during and after VAD implantation. ICD lead malfunction following LVAD implantation may be explained by more than one etiologic mechanism, and multiple mechanisms may play a role in an individual patient. The small number of patients in the study, especially the small number of patients undergoing concomitant RVAD or TV repair, limited our ability to determine definitively the cause of ICD lead parameter changes, and further research is needed. Pre- and post-LVAD chest computed tomography scans and transesophageal echocardiography were not systematically performed in all patients, and the images shown may not be representative of all LVAD patients. ECGs were compared before and after LVAD implantation, but not at specific time points. LVAD artifact may decrease the accuracy of ECG measurements.
Prospective evaluation of ICD lead function during and following LVAD implantation suggests that lead malfunction can occur, but may improve over time. Intraoperative RV sensing and HV impedance changes were not detected until after weaning of CPB, suggesting the mechanism of RV lead malfunction may be due to unloading of the LV and concomitant leftward septal shift. A conservative approach is warranted in many patients with ICD parameter changes after LVAD implantation because parameter abnormalities may improve over time.
COMPETENCY IN MEDICAL KNOWLEDGE: We evaluated ICD lead function before, during and after LVAD to gain insight into the mechanism of lead malfunction. The present prospective study is the first to evaluate ICD leads at multiple time points throughout and after LVAD implantation. We found that ICD lead parameter changes are common. RV sensing decreased by >50% from baseline in 22% of patients, pacing threshold increased by >50% in 6% of patients, pacing impedance changed by >100 Ω in 19% of patients, and HV impedance changed by >10-Ω in 28% of patients. However, acute parameter changes may improve over weeks to months. The findings of this study should improve the Patient Care and Procedural Skills ACGME competency. The possibility that lead parameters may improve over time should influence clinician decision-making. For the majority of patients, VADs are a bridge to transplant or destination therapy. In either case, systemic infection relating to cardiac implantable device infection can be catastrophic. The decision to revise an ICD lead should be made only after careful consideration of the potential risks and benefits to the individual patient. Our data suggest that a conservative approach is warranted in many patients with ICD system malfunction after LVAD implantation, because there is a potential that the acute decline in lead parameters might regress over a period of time.
TRANSLATIONAL OUTLOOK: ICD lead malfunction can occur following LVAD implantation, however the mechanism is not known. In the present study, the timing of intraoperative ICD lead sensing and HV impedances changes, which occurred after weaning of cardiopulmonary bypass, lack of change in fluoroscopic position on chest radiographs, and improvement of lead parameters over time suggest that the mechanism of lead malfunction is not related to lead dislodgement, but rather may due to unloading of the left ventricle and concomitant leftward shifting of the interventricular septum. However, the further research is needed to support the current findings. As the LVAD unloads the left ventricle, the interventricular septum shifts leftward. The effect of LV unloading on the electrophysiological function of the heart requires more study, including the impact on defibrillator lead function. We found that intrinsic QRS duration decreased after LVAD implantation, consistent with electrical remodeling. Electrophysiologists should be aware that changes in RV, septal, and LV dynamics that occur after LVAD implantation may affect the electrophysiological properties of the heart. Further research should focus on evaluating the relationship between LVAD therapy for the failing heart and the electrophysiological function of the heart.
The authors thank Beth Vetrovec, Thomas Mordica, Shawn Campbell, and Terry Harris for assistance with data collection during this study. The authors are indebted to Luke Wolfe for assistance with statistical analysis.
For supplemental tables, please see the online version of this article.
Dr. Kasirajan is a consultant for Syncardia, Atricure, and Stryker; and has received research grants through Virginia Commonwealth University from Syncardia, Atricurre, and Thoratec. Dr. Shah is a consultant for HeartWare, Inc.; and has received research grants from Thoratec. Dr. Koneru is a consultant for and has received speaker honoraria from Medtech. Dr. Ellenbogen has received research grants from and is a consultant and speaker for Medtronic, Boston Scientific, St. Jude Medical, and Biotronik. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- cardiopulmonary bypass
- chest radiograph
- glomerular filtration rate
- implantable cardioverter-defibrillator
- left ventricle
- left ventricular assist device
- mitral regurgitation
- right atrium
- right ventricle
- tricuspid valve
- ventricular assist device
- Received October 15, 2015.
- Revision received January 12, 2016.
- Accepted January 14, 2016.
- American College of Cardiology Foundation