Navigation of lead extraction—is it possible? Impact of preprocedural electrocardiogram-triggered computed tomography on navigation of lead extraction†

Abstract

OBJECTIVES

As the number of transvenous lead extractions continues to increase, preprocedural protocols for this procedure must be assessed. The objective of this study was to determine whether an electrocardiogram (ECG)-triggered computed tomography (Et-CT) with three-dimensional (3D) reconstructions could aid lead extractors in choosing the optimal tools to improve procedural success and avoid complications.

METHODS

In this study, 31 patients scheduled for transvenous lead extraction underwent a preprocedural Et-CT between January 2016 and May 2017. Both 3D-reconstructions and the two-dimensional files were reviewed for possible lead adhesions, calcifications, migrations or perforations.

RESULTS

Mean age was 46.7 ± 14.0 years. Seventy-one percent of patients were men, and 29.0% had undergone prior cardiac surgery. Indications for extraction included infection (n = 18, 58.1%), lead dysfunction (n = 8, 25.8%), upgrade (n = 3, 9.7%), severe tricuspid regurgitation (n = 1, 3.2%) and superior vena cava occlusion (n = 1, 3.2%). Eighteen patients had an implantable cardioverter defibrillator (58.1%). Sixty-eight of 70 targeted leads were extracted with a mean of 2.2 leads per patient and an average lead age of 109.3 ± 58.7 months. Et-CT files supported transvenous lead extraction by revealing possible adhesions in 16 patients, 5 perforations and 2 venous occlusions. Lead extraction was performed using the excimer laser, mechanical tools and femoral snares. Complete procedural success was achieved in 93.5% (n = 29) of cases. Clinical success was 100%, and intraoperative mortality was 0%.

CONCLUSIONS

A preprocedural Et-CT with 3D reconstructions can help to visualize lead alignment and identify abnormalities that may foreshadow procedural difficulties. A preprocedural Et-CT may therefore aid lead extractors in choosing the optimal extraction tool and strategy.

INTRODUCTION

The use of cardiac implantable electronic devices has increased significantly over the past several decades [1–3]. This increased number of implanted devices has paralleled a growing need for transvenous lead extraction (TLE), mainly due to lead failures and device infections. While advancement of TLE techniques has allowed for safer and more successful extractions, the procedure is still associated with morbidity and mortality [4–7], particularly in elderly, multimorbid patients with several, chronically implanted leads. The most feared complications in these complex cases are vascular tears of the superior vena cava (SVC) and cardiac avulsion. To address these potential complications, different TLE tools have been developed such as simple manual traction stylets, telescoping sheaths, femoral snares and laser-powered sheaths. However, despite years of experience and advancements in TLE, there is still no recommendation on the optimal extraction approach, tool and technique based on each patient’s unique risk. Moreover, no guideline exists on when an operator should crossover from one extraction tool to another during a procedure [1, 8].

The objective of this study was to investigate whether preprocedural electrocardiogram (ECG)-triggered computed tomography (Et-CT) imaging followed by three-dimensional (3D) reconstruction using a specialized software could guide TLE operators both in visualizing anatomical and/or lead abnormalities that pose potential risks and in choosing the optimal tool to achieve procedural success and avoid complications.

MATERIALS AND METHODS

Patient selection

Between January 2016 and May 2017, we performed a preoperative Et-CT in 31 high-risk cases undergoing TLE at the University Heart Center Hamburg, University Hospital Eppendorf, Hamburg, Germany. The total number of TLEs performed during the study period was 80. Indications for lead extraction, clinical outcomes and complications were classified based on the 2009 Heart Rhythm Society (HRS) Expert Consensus document [8]. Complete procedural success was defined as removal of all target leads and all lead material from the vascular space, with the absence of any permanently disabling complication or procedure-related death. Clinical success was defined as removal of all targeted leads and lead material from the vascular space or retention of a small portion of the lead that does not negatively impact the outcome goals of the procedure.

Patients were classified as high risk and considered for an Et-CT study if they fulfilled any of the following characteristics: low body mass index, female patients, low left ventricular ejection fraction, implantable cardioverter defibrillator (ICD) leads, dual-coil ICD leads, multiple indwelling leads (≥4) or leads at least >4 years [9]. In addition to the above-mentioned criteria, we selected patients for an Et-CT study if the preoperative venography revealed aggressive calcified adhesions or venous occlusions or if the chest X-ray was suspicious for perforation or lead migration.

Computed tomography imaging protocol

Our Et-CT protocol had been established in a previous study investigating the feasibility of a preprocedural Et-CT in 89 patients with cardiac implantable electronic devices undergoing transcatheter aortic valve replacement. In the protocol, the two-dimensional (2D) and 3D Et-CT reconstructions were investigated for lead adhesions, calcifications and perforations. Only 58 of the 89 Et-CTs were suitable for 2D and 3D analysis due to either the CT scan window (different window of interest in transcatheter aortic valve replacement patients) or the timing between image acquisition and the administration of intravenous dye solution. To address these issues in our current study, we adapted the CT window of interest and the timing of intravenous dye solution administration for our TLE Et-CT protocol. Additionally, we developed an algorithm for 3D CT analysis prior to TLE.

All Et-CT examinations were performed on a 256-multidetector computed tomography scanner (Philips iCT, Philips Healthcare, Netherlands). A prospective ECG-triggered acquisition was performed in a single breath-hold with image acquisition at the 70% interval. Et-CT scans started from the lower neck through the heart apex for evaluation of central veins and heart chambers. Image acquisition was initiated 70 s after the intravenous administration of 120 ml non-ionic contrast medium containing 400 mg I/ml (Imeron 400, Bracco Imaging, Italy) injected at 4 ml/s. Scans were only performed at a heart rate of <80 bpm. Patients with a heart rate >80 bpm received preprocedural medication of 50–100 mg atenolol.

Image analysis

One radiologist, who was not involved in the planning or initial clinical interpretation of the scheduled patients, independently evaluated the 2D imaging files. He was unaware of any suspected lead abnormalities. Our lead extraction team consisting of a cardiac surgeon and a cardiologist performed the 3D reconstruction by using a specialized software (3mensio Structural Heart™, Pie Medical Imaging, Maastricht, Netherlands). The 2D and the 3D reconstruction files were investigated for the following predefined abnormalities: lead adhesions (interlead and vascular adhesions), lead migrations, perforations, calcifications, venous occlusions or stenosis. The term ‘lead migration’ describes transmural extravascular migration of leads or parts of a lead at the vascular sites.

Extraction technique

All cases were performed in a hybrid suite under general anaesthesia by the lead extraction team with the aid of fluoroscopy, transoesophageal echocardiography and intra-arterial blood pressure monitoring. Patients were prepared for emergent sternotomy with cardiopulmonary bypass on standby. We started every procedure by placing 1 femoral arterial sheath and 2 venous femoral sheaths for safety reasons (temporary pacing, occlusion balloon or cardiopulmonary bypass) and to introduce a 5-Fr pigtail catheter into the right internal jugular vein. For extraction via the subclavian route, we used 14-Fr or 16-Fr 80-Hz excimer laser sheaths (GlideLight™, Spectranetics Corporation, Colorado Springs, CO, USA), mostly as a single-sheath technique without using outer sheaths as previously described by our group [10]. In the event of failure, a selective venography via the pigtail catheter was performed to visualize the size and shape of the adhesions. In these cases, we switched to 11-Fr or 13-Fr mechanical tools (TightRail™; Spectranetics or Evolution^® RL Mechanical Dilator Sheath; Cook Medical Inc., Bloomington, IN, USA). In cases involving fractured or damaged leads following the use of powered tools, we crossed over to a femoral approach, utilizing different snares as a last step of the TLE procedure.

Statistical analysis

Data were prospectively collected via our institution’s ongoing TLE registry. Continuous variables are expressed as mean ± standard deviation for normal distributions and as median and interquartile ranges for other distributions. Categorical variables are summarized as counts and percentages. Statistical analysis tested whether the additional use of mechanical tools was related to any of the following variables: age, gender, chronic renal insufficiency, total number of leads and remarkable findings on Et-CT. The Fisher’s exact test was applied for small sample sizes (expected cell sizes <5), and the χ² test was applied for all other samples. A 2-tailed P-value of 0.05 was considered statistically significant. Due to the small cohort, results are of an exploratory and descriptive character. An independent statistician performed all analyses using a statistical software package (IBM SPSS version 24, SPSS Inc. an IBM Company, Chicago, IL, USA).

RESULTS

Patients and leads

For the 31 patients, the mean age was 46.7 ± 14.0 years. Seventy-one percent of patients were men. Twelve patients had a left ventricular ejection fraction <30% (38.7%) and 9 (29.0%) patients had undergone prior cardiac surgery. Renal insufficiency was present in 11 (36.5%) patients, of which 1 patient was on regular haemodialysis. A pacemaker (PM) was implanted in 13 patients, and an ICD was implanted in the remaining 18 patients (including 12 cardiac resynchronization therapy defibrillator patients). The majority of the devices were left sided (n = 23). Seven devices were on the right side, and 1 patient had leads implanted on both sides. Thirteen patients were PM dependent, of which 3 patients required a temporary transcutaneous pacing system. Patient baseline characteristics are presented in Table 1.

A total of 80 leads were present in our cohort, including 23 atrial leads, 20 right ventricular pacing leads, 26 right ventricular ICD leads (19 dual-coil leads) and 11 coronary sinus leads. The mean dwell time of all leads was 113.1 ± 54.2 months, and the mean dwell time of all extracted leads was 109.3 ± 58.7 months. Active lead tip fixation was present in 26 (83.9%) patients. The remaining 5 patients had leads with active and passive fixation mechanisms. Sixty-eight of the 70 targeted leads were successfully removed (mean of 2.2 leads per patient). We were unable to extract 2 abandoned ICD leads in a patient scheduled for a cardiac resynchronization therapy defibrillator upgrade because of a damaged lumen. However, this did not negatively impact the clinical success of the procedure.

Electrocardiogram-triggered computed tomography analysis

The mean area–dose product of the Et-CT was 533.2 ± 327.2 cGy ⋅ cm². Due to poor imaging quality, 1 CT scan was excluded from further analysis. Analysis of 2D-CT and 3D reconstructions using 3mensio Structural Heart (Pie Medical Imaging) (Table 2) was unremarkable in 7 (23.3%) patients. Adhesions between the leads (interlead adhesions) were the most common finding (n = 16, 53.3%) (Fig. 1). Interlead adhesions were observed in 16 patients when using a combination of 2D and 3D CT analysis and in 13 patients when using 2D CT files only. Et-CT with 3D imaging was suspicious for vascular adhesion at the SVC in 1 patient (in 2 patients with 2D CT analysis). In 1 patient (6 patients with 2D CT analysis), Et-CT with 3D imaging depicted both interlead adhesions and vascular adhesions. We could not determine the grade of adhesions due to the lack of standardization in differentiating between metallic artefacts and true adhesions. An independent radiologist performed analysis of the 2D-CT files for calcification around the leads in patients where additional mechanical tools had been used for TLE.

Et-CT was suspicious for myocardial perforation in 5 (16.6%) patients (Fig. 2), with extrusion of the lead tip beyond the myocardial border. None of these patients had abnormal pacing impedances or capture thresholds. We did not apply a 5-point scale to grade the 2D-CT files for possible perforation as proposed by Balabanoff et al. [11]. Venous occlusion was seen in 2 (6.7%) patients (Table 3). A combination of 2D- and 3D-reconstruction analysis showed no clear signs of lead migration, whereas 2D-CT analysis was consistent with migration in 1 patient.

Procedural data

Mean procedural duration was 123.1 ± 58.0 min. The 80-Hz excimer laser (14- and 16-Fr sheaths) was used in 18 (58.1%) patients, and the Cook Evolution RL (13-Fr sheath) was used in 2 (6.5%) patients. In the remaining 11 patients, TLE was performed using a combination of the excimer laser, powered mechanical tools [Cook Evolution RL (13 Fr) or Spectranetics Tight rail (11 Fr)], and/or femoral snares (Table 4). The additional use of mechanical tools (in addition to the excimer laser) was not related to age (P = 0.832, U-test), gender (P > 0.999, Fisher’s exact test), renal insufficiency (P = 0.248, Fisher’s exact test), total number of implanted leads (P = 0.172, χ² test) or the existence of interlead adhesions in the Et-CT (P = 0.538, Fisher’s exact test).

Clinical success was achieved in all patients, and complete procedural success was achieved in 29 (93.5%) patients. In 1 of the 2 patients in which procedural success was not achieved, we were unable to stabilize 2 abandoned ICD leads with lead locking devices (LLD™; Spectranectics, Colorado Springs, USA). We did not attempt further extraction manoeuvres because the indication for TLE was an upgrade, which was successfully performed by extracting the newest (active) ICD lead. The other case involved TLE for atrial and ventricular lead dysfunction. After extracting the ICD lead and implanting a new ICD lead, we were unable to implant a new atrial lead from the left side because of aggressive adhesions and totally occluded subclavian–innominate veins. Therefore, we had to implant a new atrial lead from the right side.

Two minor complications (1 pocket haematoma and 1 pneumothorax) and 1 major complication occurred after TLE. The major complication was unrelated to TLE as it involved resuscitation for an unknown reason 8 h post-TLE after removing a femoral compression device. No intraprocedural complications were observed. Two deaths unrelated to TLE occurred during the hospital stay due to septic shock in 1 patient and hypoxic encephalopathy after resuscitation in the other patient.

DISCUSSION

Our data showed that an Et-CT prior to TLE was able to visualize lead alignment (Fig. 3) from the subclavian vein to the heart as well as detect lead abnormalities such as vascular adhesions, interlead adhesions, myocardial perforations, venous occlusions and calcification. These findings have been described to some extent in a case report and in a study by Lewis et al. [12, 13] in 30 patients who received a preprocedural ECG-gated CT. However, while Lewis et al. focused on identifying significant perforations and lead adhesions to central venous structures to assess the risk for complications during TLE, they did not consider the potential role of Et-CT in navigating or planning TLE. Abnormalities seen on Et-CT may not only identify patients at a higher risk for complications during TLE but also help to optimize and refine the intraoperative TLE setting. In high-risk patients with severe interlead adhesions and/or adhesions at the SVC predisposing patients for SVC tears, the use of femoral sheaths and a pigtail catheter in the internal jugular vein should be routinely applied to safely manage a tear with rescue devices. Another option in these high-risk TLE patients is the prophylactic placement of an endovascular occlusion balloon in the vena cava [14]. Due to signs of perforation and adhesion seen on ET-CT, we utilized an endovascular balloon in addition to our routine intraoperative setting as a backup in some patients with very old leads. Therefore, an Et-CT can increase the safety of TLE, especially in patients with infections where TLE is mandatory. In contrast to other groups [13], we did not reject TLE in case of a possible perforation on Et-CT, but we were more vigilant about perioperative pericardial effusion resulting in pericardiocentesis.

A systemically analysed Et-CT prior to TLE may help in choosing the optimal extraction tool and approach for TLE. Our group [15] and other groups [7, 16] have described promising results of TLE with the 80-Hz excimer laser as the sole extraction tool in patients with significantly shorter lead dwell times than in our present study (50.3 ± 18.4 months vs 109.3 ± 58.7 months). However, the excimer laser is sometimes unsuccessful as an exclusive tool in patients with older leads and calcified interlead or vascular adhesions. In these cases, complete TLE success can only be achieved with the additional use of powered mechanical tools with rotating threaded tip sheaths. Our study is the first to systematically evaluate preprocedural 2D CT files and 3D reconstruction files of an Et-CT using 3mensio Structural Heart for these adhesions and other predefined abnormalities. Identification of strong and severely calcified adhesions between the leads or adjacent to the venous wall before the actual extraction may guide the clinician in planning the optimal strategy and extraction tool for the procedure. This is especially helpful when deciding to switch from one tool to another (from the laser sheath to a mechanical tool). However, identification of calcifications is still the most challenging task of preprocedural analysis as it requires differentiation between chalk and metallic streak artefacts. Due to the lack of an exact definition and classification of interlead adhesions, vascular adhesions and calcifications on Et-CT, further improvement of image analysis using CT raw data and the development of standards are necessary. In our first small series, we were, therefore, unable to detect a relationship between the additional use of mechanical tools and interlead adhesions in patients where the excimer laser as an exclusive tool failed. Large multicentre studies and further refinement of image analysis are now necessary to evaluate whether it is possible to choose the optimal extraction tool depending on the type of interlead and vascular adhesions seen on preprocedural Et-CT.

The development of image integration systems for the hybrid operating suite may also allow integration of Et-CTs into fluoroscopy systems as has already been established in electrophysiology labs that integrate CTs and magnetic resonance imaging into mapping systems. Integration of Et-CTs with intraoperatively performed venography, live fluoroscopy and 3D-transoesophageal echocardiography could enable us to perform real-time TLE navigation in the future.

As with any X-ray-based imaging modality, there is a concern about radiation exposure. However, given the potential fatal complications of TLE, we believe the benefits of an Et-CT outweigh the risks of the additional radiation exposure in high-risk TLE patients. In low-risk patients with a short lead dwell time, a preprocedural Et-CT is probably not required.

Limitations

The major limitation of this study is that it is a small, non-randomized, retrospective, single-centre study. The only way to evaluate the potential value of an Et-CT would be a prospective clinical trial comparing procedural data and outcome of patients with an Et-CT prior to TLE with a similar group of patients without an Et-CT. Therefore, a definite statement considering the value of an Et-CT cannot be made based on this preliminary work.

Furthermore, only patients deemed to be at high risk received an Et-CT. Due to limited data on risk factors of patients undergoing TLE [17], a standard definition of high-risk TLE patients has not yet been established. Thus, the definition of high risk is centre specific. Additionally, one of the major limitations in radiological analysis was the exact differentiation of true calcifications and interlead adhesions from metallic streak artefacts as well as the classification of the severity of interlead adhesions. Although this limitation may be overcome in future, an Et-CT is unable to distinguish leads that are adjacent to the SVC lateral wall from the ones that are embedded in the wall. Other intraoperative imaging techniques such as 3D transoesophageal echocardiography offer real-time, dynamic monitoring and may provide such information.

CONCLUSION

A preprocedural Et-CT followed by 2D and 3D reconstruction using a specialized software may offer the potential to identify lead abnormalities that may foreshadow complications during TLE. Integrating a special Et-CT into the preprocedural workflow of patients undergoing TLE can help both to reduce complications by optimizing the intraoperative setting and to achieve complete procedural success by choosing the optimal extraction tool for each patient. Large, multicentre studies are warranted to prospectively evaluate whether an Et-CT has an impact on outcome and complications of TLE.

Conflict of interest: Samer Hakmi is a consultant/training proctor of Spectranetics Corp. All other authors declared no conflict of interest.

REFERENCES

Author notes