Author + information
- Received August 6, 2008
- Revision received October 7, 2009
- Accepted November 2, 2009
- Published online March 2, 2010.
- Jon A. Grammes, DO⁎,
- Christopher M. Schulze, DO⁎,
- Mohammad Al-Bataineh, MD⁎,
- George A. Yesenosky, MD⁎,
- Christine S. Saari, MSN, CRNP⁎,
- Michelle J. Vrabel, MSN, CRNP⁎,
- Jay Horrow, MD, MS†,
- Mashiul Chowdhury, MD‡,
- John M. Fontaine, MD⁎ and
- Steven P. Kutalek, MD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Steven P. Kutalek, Drexel University College of Medicine, 245 North 15th Street, Mail Stop 470, Philadelphia, Pennsylvania 19102
Objectives We describe the feasibility, safety, and clinical outcomes of percutaneous lead extraction in patients at a tertiary care center who had intracardiac vegetations identified by transesophageal echocardiogram.
Background Infection in the presence of intracardiac devices is a problem of considerable morbidity and mortality. Patients with intracardiac vegetations are at high risk for complications related to extraction and protracted clinical courses. Historically, lead extraction in this cohort has been managed by surgical thoracotomy.
Methods We analyzed percutaneous lead extractions performed from January 1991 to September 2007 in infected patients with echocardiographic evidence of intracardiac vegetations, followed by a descriptive and statistical analysis.
Results A total of 984 patients underwent extraction of 1,838 leads; local or systemic infection occurred in 480 patients. One hundred patients had intracardiac vegetations identified by transesophageal echocardiogram, and all underwent percutaneous lead extraction (215 leads). Mean age was 67 years. Median extraction time was 3 min per lead; median implant duration was 34 months. During the index hospitalization, a new device was implanted in 54 patients at a median of 7 days after extraction. Post-operative 30-day mortality was 10%; no deaths were related directly to the extraction procedure.
Conclusions Patients with intracardiac vegetations identified on transesophageal echocardiogram can safely undergo complete device extraction using standard percutaneous lead extraction techniques. Permanent devices can safely be reimplanted provided blood cultures remain sterile. The presence of intracardiac vegetations identifies a subset of patients at increased risk for complications and early mortality from systemic infection despite device extraction and appropriate antimicrobial therapy.
- percutaneous lead extraction
- intravascular infection
- cardiac device infection
Landmark trials have expanded the indications for cardiac rhythm management (CRM) devices (1–4). The consequences of device-related complications are well documented (5–8). Infection of CRM devices has been reported in 0.8% to 19.9% of patients (9). Mortality rates for untreated patients are as high as 66% compared with 18% for patients treated with antibiotics and device removal (5,7,8,10–12). Eradication of device-related infection requires complete system removal, given the morbidity and mortality associated with antibiotics alone or partial device removal (5,7,11,13–16).
Patients with infected devices and intracardiac vegetations represent a high-risk population, the incidence of which is difficult to determine from available studies. Arber et al. (17) documented the incidence of endocarditis with device infection to be 9.4%. Patients with vegetations >1 cm have historically been managed with extraction through open thoracotomy due to potential for septic embolization and hemodynamic compromise (6,10,14,18–20). A review of the literature reveals limited data describing the efficacy and safety of percutaneous lead extraction in this subset (6,14,19–21).
In this study, we document the incidence of intracardiac vegetations in patients with CRM devices referred to an extraction center and evaluate the safety and efficacy of lead extraction in this high-risk population. Further, we examine the safety and clinical outcomes associated with reimplantation of a new CRM device.
Our institution's institutional review board approved the study. A comprehensive database is maintained for all patients undergoing lead extraction. Patients who had objective evidence of device-related infection and echocardiographic evidence of intracardiac vegetations and underwent device extraction at our institution from January 1991 to September 2007 were included in this study. Office charts and hospital records were reviewed to gather pertinent clinical information.
All patients who underwent lead extraction in 2004 without evidence of intracardiac vegetations served as a reference population. We compared implant duration and extraction times between the 2004 cohort and our study population. To standardize comparisons, only data from the first lead extracted in each patient were used.
Device-related infection was defined by signs of local pocket inflammation, generator or lead erosion, or positive blood cultures. Device-related endocarditis was defined according to the modified Duke criteria (22). Patients with a diagnosis of definite or possible endocarditis were included. Intracardiac vegetation was defined as a discrete, echogenic, oscillating mass found on a valve, lead, or endocardial surface and confirmed in multiple views by echocardiography. Echocardiograms were reviewed by an echocardiographer and electrophysiologist to determine vegetation size and location and to distinguish true vegetation from false echodensities. In 23 cases, echocardiogram images were not available for review and information was obtained from study reports.
Implant duration was the time between initial lead implantation to the time it was extracted. Lead extraction time was defined as the duration from the time when the head of the lead was cut off until its complete removal. Extraction time for leads removed with traction only was the time from removal of the anchoring sleeve to complete removal.
Because of potential for septic embolization, pre-operative transesophageal echocardiography (TEE) is used to risk stratify patients. After extraction, TEE is performed in all patients before device reimplantation. In addition, intraoperative intracardiac echochardiography (ICE) may be used to assess lead fibrosis and vegetation stability and to monitor for complications. All patients undergo serial blood cultures and intravenous antibiotic therapy according to sensitivities of isolated organisms and the recommendations of an infectious disease consultation. Temporary transvenous pacing wires are placed when indicated despite presence of vegetations until a permanent system can safely be implanted.
Standard methods for percutaneous lead extraction are utilized at the discretion of the extracting physician (20,23–28). An electrosurgical dissection sheath (Cook Corp., Vandergrift, Pennsylvania) or Excimer laser (Spectranetics Corp., Colorado Springs, Colorado) is used to extract heavily fibrotic leads. Extensive surgical debridement of the pocket is performed using electrocautery and blunt dissection. Timing of device reimplantation is based primarily on sterility of blood cultures, but also on resolution of vegetations. Reimplantation is usually performed on the contralateral side. In select cases, an ipsilateral lateral subpectoral location may be used depending on the extent of pocket infection.
Patients are followed up in the outpatient office after extraction. In cases when patients were referred back to their primary institution, the primary cardiologist or internist was contacted for information regarding the patient's medical progress.
Mean and SD and/or median and interquartile ranges (IQRs) summarize continuous demographic data. The Wilcoxon test compares implant durations and extraction times in the study population to those of the 2004 reference population. Kaplan-Meier curves and product-limit life-table estimates analyzed the time from lead extraction to reimplantation of a new lead, stratified into groups by microorganism type. Both log-rank and Wilcoxon tests were used to compare those groups with respect to time to reimplantation. All analyses used a significance level of 0.05.
Figure 1represents the patients referred for lead extraction from January 1991 through September 2007. Nine hundred eighty-four patients underwent extraction, and 1,838 leads were removed. Four hundred eighty patients (49%) required device extraction of 1,000 leads for systemic or localized pocket infection. One hundred patients had echocardiographic evidence of intracardiac vegetations (Figs. 2Aand 2B), and they represent 10% of our total extraction population (Table 1).Vegetations were found on leads in 56 cases and valves in 35 cases; in 9 cases, this information was not available. Vegetations ranged in size from 0.2 to 4.0 cm in largest longitudinal diameter; the mean diameter was 1.6 cm. All patients underwent complete device extraction in our electrophysiology laboratory; 215 leads were removed using standard percutaneous techniques. No patients required surgical extraction. Seventy-one patients were men. Mean age was 67 ± 15 years (age range 22 to 92 years).
Figure 3depicts the infectious pathogens isolated from cultures. The most common infectious organisms were methicillin-resistant Staphylococcus aureus(MRSA) and methicillin-sensitive Staphylococcus aureus(MSSA), isolated from 34 and 25 cultures, respectively. Other organisms included coagulase-negative Staphylococcusspecies (CNSS) (n = 14), Enterococcus faecalis(n = 4), Streptococcusspecies (n = 4), vancomycin-resistant Enterococcus(n = 1), Citrobacter(n = 1), and Candidaspecies (n = 1). Sixteen patients had negative blood cultures despite having evidence of endocarditis. Five of these patients had pocket tissue cultures that were positive, including MRSA in 1 patient, MSSA in 2 patients, and CNSS in 2 patients.
Implant duration and extraction times
Implant duration in the study population ranged from 1 to 300 months (mean 50.9 ± 52.8 months, median 32.5 months, IQR 61 months). Extraction times ranged from 1 to 187 min (mean 11.2 ± 21.9 min, median 3 min, IQR 13 min). Implant duration for the first lead extracted from each patient in the study population, 54.4 ± 56.8 months (median 34 months, IQR 68.5 months, n = 99), did not differ from that of the 2004 reference population, 40.2 ± 45.3 months (median 24 months, IQR 36.5 months, n = 66; p = 0.095 by Wilcoxon test). Extraction time for the study population, 10.8 ± 16.8 min (median 4 min, IQR 15 min, n = 99), did not differ from that of the reference population, 5.3 ± 6.3 min (median 1 min, IQR 6 min, n = 66; p = 0.067 by Wilcoxon test).
Reimplantation of a new device
Fifty-four (54%) patients were reimplanted with a new device during hospitalization (Fig. 4).Two of these patients underwent reimplantation of epicardial systems due to persistent vegetations. Median time to reimplantation was 7 days (mean 11 days). Life-table analysis of time to reimplantation disclosed median times of 7 (95% confidence interval [CI]: 6 to 16) days for patients with negative blood cultures, 8 (95% CI: 7 to undefined) days for patients whose culture results were positive for CNSS, 13 (95% CI: 7 to 27) days for MSSA, and 14 (95% CI: 7 to 60) days for MRSA. The undefined upper limit for CNSS was obtained from a final censored observation. Both the log-rank test (p = 0.3096) and the Wilcoxon test (p = 0.5120) demonstrated no impact of culture result on time to reimplantation.
Follow-up and clinical outcomes
Follow-up was available for 71 patients (Fig. 5).Twenty-nine patients were lost to follow-up because of their diverse locations of origin. The average follow-up was 438 days (median 150 days).
Forty-six patients were not reimplanted during the hospitalization. Four patients were transferred back to their primary institution after device extraction. Eighteen patients were not reimplanted because of persistent systemic infection. Two patients refused reimplantation despite a medical indication, and 2 patients were referred to hospice. One patient was not reimplanted because of metastatic cancer. Of the remaining 19 patients, 18 had no compelling indication to reimplant a permanent device. Extensive thrombosis of bilateral subclavian veins limited access in the other patient. Thirty-nine of the 46 patients (85%) were discharged in stable condition. Seven patients died during the index hospitalization, and 1 experienced a complication during lead extraction; these patients are discussed in the following text.
Fifty-four patients underwent subsequent device reimplantation. Fifty-one patients (94%) were discharged in stable condition. During follow-up, no patient had clinical indicators of relapsing infection that required a second extraction. Three patients died during the index hospitalization and 4 others had complications during lead extraction; these patients are discussed in the following text.
Morbidity and mortality
Of the 71 patients who had long-term follow-up available, 19 patients died (27%). Ten patients died during the index hospitalization, and 9 died after being discharged. Eleven patients (59%) died of persistent septicemia, 1 patient died of SCD, and 7 patients died of unknown causes but had no evidence of ongoing sepsis during follow-up. Four patients died after being reimplanted with a new device despite having sterile blood cultures and no vegetations on repeat TEE. Post-operative 30-day mortality was 10% (Table 2).
Five patients had complications during lead extraction. Two patients had embolization of the vegetation. In 1 of those patients, embolization occurred during attempts to snare a large vegetation before lead removal. The patient remained hemodynamically stable but had chest pain after the procedure. High-resolution computed tomography (Fig. 6)revealed a filling defect in the pulmonary artery. The patient underwent uncomplicated device reimplantation 8 days later. A third patient required an additional procedure to snare a lead fragment that embolized to the pulmonary artery during the initial extraction. The lead was 25 years old and was completely removed in pieces because of its brittle condition. The patient experienced no clinical symptoms and underwent reimplantation 7 days later. A fourth patient had severe tricuspid regurgitation due to a flail posterior leaflet that was not present before extraction. A fifth patient had hypotension after extraction that required vasopressor support for 24 h. Subsequent echocardiogram was unrevealing, and the patient had a new device implanted 12 days later. All 5 patients were discharged in stable condition and had uneventful clinical follow-up.
The management of device-related infection is well described, but considerable variability exists in regional management strategy. Complete hardware removal is essential to eradicate infection and prevent relapsing bacteremia. In 9 retrospective studies, the mortality rate was 41% for patients treated with antibiotics alone compared with 19% for patients receiving antibiotics and complete device removal (28).
Patients with device-related infection and intracardiac vegetations >1 cm have historically undergone surgical thoractomy for device removal because of the potential for septic embolization. Literature supporting this management approach is limited (6,14,28,29). Our single-center experience involving this high-risk population suggests that standard endovascular percutaneous extraction of leads is both feasible and safe.
Incidence of intracardiac vegetations
The incidence of intracardiac vegetations comprised 10% of our entire extraction population and 21% of all patients with device-related infection. The incidence is greater than that found in earlier studies (17) but is comparable to recent studies that describe endocarditis in 20% to 25% of device-related infections (30,31). The incidence will vary according to the institution's referral characteristics, how device-related endocarditis is defined, as well as imaging modalities utilized. For the purpose of this study, we included patients with clinically definite or possible endocarditis by modified Duke criteria. The 16 patients who were classified as possible endocarditis had evidence of vegetations but only clinically local infection with negative blood cultures. Without pathologic verification, these masses could represent fibrous casts or thrombus; however, the incidence of culture-negative endocarditis is ≈5% to 7%, and cultures preceded by antibiotics may be negative in as many as 14% of cases (32). Given the frequent ambiguity of initial culture data, we believe it is important to image all infected patients undergoing device extraction to identify high-risk patients and curtail future infection relapses.
All leads were successfully removed using a percutaneous approach. Implant duration and extraction time between the study group and the reference population were not significantly different. In fact, there was a trend toward shorter extraction times in patients with vegetations. This finding likely reflects improvement in extraction techniques and operator experience over time. Although we did not compare variables such as lead fixation mechanism or degree of fibrosis, our data suggest that the presence of intracardiac vegetations does not prolong extraction time.
Our data parallel the 93% to 97% percutaneous extraction success rates reported in other large cohorts (28,33–34). Smaller studies involving patients with endocarditis also document successful percutaneous extraction (14,15,30,35). Meier-Ewert et al. (30) successfully removed leads in 9 patients, and Victor et al. (15) removed leads in 9 of 12 patients; all vegetations were >1 cm. Massoure et al. (35) removed leads with a mean vegetation size of 1.3 cm in 36 of 37 patients.
One patient required a femoral approach using the Byrd workstation to snare a lead fragment that embolized to the pulmonary artery. Using the needle's eye snare or the Dotter Basket Retrieval System (Cook Vascular, Leechburg, Pennsylvania) in combination with countertraction is a well-described alternative (36–39). Published series from Klug et al. (36) and Espinosa et al. (38) have reported success rates of 87.2% and 82%, respectively.
Our data support the idea that a percutaneous lead extraction in patients with intracardiac vegetations has a high success rate and that leads can be removed in their entirety using standard techniques.
Mortality and morbidity
In our study the post-operative 30-day mortality was 10%. Although a standard measure of surgical outcomes, post-operative mortality may not reflect procedural mortality. The 10 patients who died had multiple comorbidities and overwhelming sepsis; their deterioration was secondary to the natural progression of disease. One patient was transferred from an outside hospital on vasopressor support with septic shock. Although the device was extracted without incident, the patient died later that day after unsuccessful resuscitation. All other deaths occurred at least 5 days after extraction and were not a result of embolized vegetations or from the extraction procedure. These unfortunate outcomes occurred in a critically ill subset of patients who often have extensive comorbidities.
The overall mortality for percutaneous lead extraction reported from large databases or single centers with experienced operators is low. Mortality rates range from 0.1% to 0.6% and major complication rates range from 1.4% to 1.9% (20,27,33,34). Data regarding operative mortality for patients with intracardiac vegetations is less robust. Klug et al. (14) reported a mortality rate before discharge of 7.6% and an overall mortality of 26.9% at 20 months. Further, Victor et al. (15) reported deaths of 3 of 14 patients (21%) who underwent percutaneous lead extraction. Most recently, Sohail et al. (31) reported deaths of 9 of 189 patients (5%). Only 44 patients in that study had endocarditis; 5 of those patients died, yielding a mortality rate of 11%. All patients in our study had intracardiac vegetations. We believe that the 10% 30-day operative mortality rate in our series reflects severity of disease rather than the mode of extraction.
Depending on the definitions used, major complication rates reportedly range from 0.4% to 1.4% (33,34). Sohail et al. (31) reported 20 patients (11%) who had complications after percutaneous extraction, but they are not categorized according to definitions published in previous guidelines. Limited data are available regarding complications during percutaneous lead extraction in patients with intracardiac vegetations. In our experience, 5 patients had complications related to the procedure, but none could be classified as major. We believe this is an acceptable complication rate given the high-risk nature of the population.
Potential for pulmonary embolization
A concern for septic embolization in patients with large vegetations >1 cm is documented in the literature (14,15,19,30). Studies involving right-sided endocarditis have documented an increased risk of embolization with larger vegetations but demonstrate an excellent prognosis and no detrimental impact on survival (35,40,41). Klug et al. (14) reported a pre-operative incidence of embolization, confirmed by ventilation-perfusion scans, in as many as 34% of patients. Meier-Ewert et al. (30) found evidence of pulmonary embolism in 5 of 9 patients (55%) by scintigraphic scans. Hospital stay for these patients was not statistically prolonged compared with patients without embolism.
In our experience, only 2 patients experienced witnessed embolization of vegetation material, which measured >2 cm before extraction in both cases. A third patient had presumptive embolization of a vegetation measuring 1.2 cm after extraction, and experienced transient hemodynamic changes requiring brief vasopressor support. All 3 patients made a full recovery and were discharged home. It is conceivable that septic embolization may have caused the death of 1 patient who died the day of extraction; however, her clinical deterioration prevented any confirmatory imaging, and she was overwhelmingly septic before the procedure. Routine imaging was not performed in this series, but given the aforementioned studies and the frequent absence of vegetations seen on post-operative imaging, the incidence of silent embolism is likely higher. Although embolization represents a potential complication and is likely underestimated in our experience, our data suggest that the majority of patients with vegetations up to 4 cm in size can safely undergo percutaneous lead extraction without clinical compromise.
Method of extraction
Historically, patients with intracardiac vegetations have preferably been managed by surgical approaches to directly visualize lead removal and limit the risk of lead fragmentation, tricuspid valve damage, and septic embolization that occurs with percutaneous techniques. In these situations, success with conventional median sternotomy with or without cardiopulmonary bypass as well as the limited atriotomy approach is well described (14,28,29,42,43). Unfortunately, there remains significant morbidity and mortality due to the invasive nature and prolonged recovery times, with mortality rates ranging from 12.5% to 40% (20,28,29). No study to date directly compares surgical and percutaneous approaches, and data are limited in patients with intracardiac vegetations. Table 3compares mortality rates for both surgical and percutaneous techniques.
Our experience represents that of a tertiary care referral center. Results at centers without comparable experience with lead extraction may differ. Further, standardization of post-operative care and long-term follow-up is limited as many patients were referred back to their primary institution. The true incidence of intracardiac vegetations may be overestimated because of our referral bias and the method used to define endocarditis. Although the modified Duke Criteria served as a guideline to define endocarditis, histologic confirmation was not performed, and thus thrombus and fibrous casts may have been mistaken for vegetation. Further, some echocardiograms were unavailable for review and thus represent a source of potential false positive vegetations. In addition, our use of TEE to screen patients with localized lead erosion discloses additional patients with intracardiac vegetations who may have been overlooked otherwise.
As indications for device implantation expand, the prevalence of device-related infections will increase. Vigilant recognition and aggressive management should be the pillars of therapy if potentially fatal complications are to be avoided. Patients with infection and intracardiac vegetations represent a high-risk population with multiple comorbidities and significant mortality regardless of management strategy. Our experience suggests that percutaneous lead extraction, when performed by experienced operators, using standard techniques is both safe and feasible in patients with intracardiac vegetations up to 4 cm. Neither the causative microorganism nor the presence of intracardiac vegetations impacts the time required for lead extraction. Additionally, reimplantation of CRM devices is feasible and safe when guided by standard principles of infectious disease. Prospective randomized trials are needed to determine appropriate methods of surveillance for vegetations and timing of device reimplantation so relapse of infection can be minimized.
Dr. Schulze has trained representatives of St. Jude Medical in device implants; is on the Speakers' Bureau of Medtronic; and has received practice development support from Boston Scientific. Dr. Horrow is an employee of AstraZeneca LP. Dr. Kutalek is a consultant for Spectranectics Co.
- Abbreviations and Acronyms
- confidence interval
- coagulase-negative Staphylococcusspecies
- cardiac rhythm management
- intracardiac echocardiography
- interquartile range
- methicillin-resistant Staphylococcus aureus
- methicillin-sensitive Staphylococcus aureus
- sudden cardiac death
- transesophageal echocardiogram
- Received August 6, 2008.
- Revision received October 7, 2009.
- Accepted November 2, 2009.
- American College of Cardiology Foundation
- Camus C.,
- Leport C.,
- Raffi F.,
- Michelet C.,
- Cartier F.,
- Vilde J.L.
- Baddour L.,
- Bettmann M.,
- Bolger A.,
- et al.
- Bluhm G.
- Klug D.,
- Lacroix D.,
- Savoye C.,
- et al.
- Victor F.,
- De Place C.,
- Camus C.,
- et al.
- Chamis A.L.,
- Peterson G.E.,
- Cabell C.H.,
- et al.
- Voet J.,
- Vanderkerckhove Y.,
- Muyldermans L.,
- et al.
- Li J.S.,
- Sexton D.J.,
- Mick N.,
- et al.
- Kantharia B.,
- Pennington J.,
- Kleinman D.,
- et al.
- Wilkoff B.,
- Byrd C.,
- Love C.,
- et al.
- Sohail M.R.,
- Uslan D.Z.,
- Khan A.H.,
- et al.
- Lamas C.,
- Eykyn S.J.
- Sandre R.M.,
- Shafran S.D.
- Byrd C.L.,
- Wilkoff B.L.,
- Love C.J.,
- et al.