Author + information
- Received July 10, 2017
- Revision received September 14, 2017
- Accepted September 21, 2017
- Published online February 19, 2018.
- Thomas A. Boyle, BSa,∗ (, )
- Daniel Z. Uslan, MD, MBAb,
- Jordan M. Prutkin, MD, MHSc,
- Arnold J. Greenspon, MDd,
- Larry M. Baddour, MDe,
- Stephan B. Danik, MDf,
- Jose M. Tolosana, MD, PhDg,
- Katherine Le, MDe,
- Jose M. Miro, MD, PhDg,
- James E. Peacock, MDh,
- Muhammad R. Sohail, MDe,
- Holenarasipur R. Vikram, MDi,
- Roger G. Carrillo, MDa,∗ (, )
- for the MEDIC Investigators
- aDepartment of Surgery, Division of Cardiothoracic Surgery, University of Miami, Miami, Florida
- bDepartment of Medicine, Division of Infectious Diseases, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, California
- cDepartment of Medicine, Division of Cardiology, University of Washington, Seattle, Washington
- dDepartment of Medicine, Division of Cardiology, Cardiac Electrophysiology Laboratory, Thomas Jefferson University Hospital, Philadelphia, Pennsylvania
- eDepartment of Medicine, Divisions of Infectious Diseases and Cardiovascular Diseases, Mayo Clinic College of Medicine, Rochester, Minnesota
- fDepartment of Medicine, Division of Cardiology, Massachusetts General Hospital, Boston, Massachusetts
- gCardiology and Infectious Disease Services, Hospital Clinic, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), University of Barcelona, Barcelona, Spain
- hDepartment of Medicine, Section of Infectious Diseases, Wake Forest School of Medicine, Winston-Salem, North Carolina
- iDepartment of Medicine, Division of Infectious Diseases, Mayo Clinic, Phoenix, Arizona
- ↵∗Address for correspondence:
Dr. Thomas A. Boyle OR Dr. Roger G. Carrillo, Department of Surgery, Division of Cardiothoracic Surgery, University of Miami Hospital, 1295 NW 14th Street, Suite H, Miami, Florida 33125.
Objectives This study sought to evaluate the impact of abandoned cardiovascular implantable electronic device (CIED) leads on the presentation and management of device-related infections.
Background Device infection is a serious consequence of CIEDs and necessitates removal of all hardware for attempted cure. The merits of extracting or retaining presumed sterile but nonfunctioning leads is a subject of ongoing debate.
Methods The MEDIC (Multicenter Electrophysiologic Device Infection Cohort) prospectively enrolled patients with CIED infections at 10 institutions in the United States and abroad between January 1, 2009, and December 31, 2012. Within a propensity-matched cohort, relevant clinical information was compared between patients who had 1 or more abandoned leads at the time of infection and those who had none.
Results Matching produced a cohort of 264 patients, including 176 with no abandoned leads and 88 with abandoned leads. The groups were balanced with respect to Charlson comorbidity index, oldest lead age, device type, sex, and race. At the time of admission, those with abandoned leads were less likely to demonstrate systemic signs of infection, including leukocytosis (p = 0.023) and positive blood cultures (p = 0.005). Conversely, patients with abandoned leads were more likely to demonstrate local signs of infections, including skin erosion (p = 0.031) and positive pocket cultures (p = 0.015). In addition, patients with abandoned leads were more likely to require laser extraction (p = 0.010).
Conclusions The results of a large prospective registry of CIED infections demonstrated that patients with abandoned leads may present with different signs, symptoms, and microbiological findings and require laser extraction more than those without abandoned leads.
The application of cardiovascular implantable electronic devices (CIEDs) is growing rapidly due to improved technology, expanded indications, and an aging population in developed countries (1–4). Moreover, procedures to remove or revise those devices, which include insertable cardioverter-defibrillators (ICDs), permanent pacemakers, and cardiac resynchronization therapy devices, are experiencing a similar rise (5–10).
A major impetus for surgical revision is CIED infection, which can present as a pocket infection, bacteremia, endocarditis, or some combination thereof. Infections have complicated between 0.5% and 7% of all device-related procedures (8,10–12), and their incidence may be increasing faster than the rate of implantation (13). Regardless of how it presents, CIED infection can cause serious morbidity or death related to infection or its management (14,15). In addition to early identification and initiation of antimicrobial therapy, expeditious removal of all CIED hardware is needed to hopefully reduce the infection complication rate (16).
When a device is being revised or replaced for reasons other than infection, old or nonfunctioning leads are often capped and abandoned. The extent of this practice is not fully known, but at 1 high-volume center equipped for lead extraction, 39% of patients who underwent lead revisions for noninfectious indications had leads capped rather than extracted (17). The decision to extract a sterile but nonfunctioning lead requires weighing the small (11) but immediate risks of extraction with the potential but largely uncharacterized risks of abandonment. Published discussions of the long-term risks associated with abandoned leads (ALs) include stenosis, hardware fracture, migration, and pocket pain. To date, the impact of ALs on future CIED infections has received less attention (17–19). A single-center study of lead extractions recently demonstrated more complications and suboptimal outcomes in patients with ALs (20,21). The present study used a prospective, multicenter registry to evaluate the impact of previously ALs on the presentation and management of device infections.
The MEDIC (Multicenter Electrophysiologic Device Infection Cohort) registry was a cooperative international effort that prospectively enrolled subjects with known CIED infections between January 1, 2009, and December 31, 2012. The registry consisted of demographic, clinical, laboratory, management, and outcome data from patients at 10 medical centers. The study protocol was approved at each site by its respective institutional review board.
With few exceptions, multiple blood cultures and pocket cultures were taken from each patient. Patients received a transthoracic echocardiogram to assess for vegetation, and many received a transesophageal echocardiogram as well. In a number of cases, patients received computed tomography studies to evaluate for septic pulmonary emboli. If at all possible, percutaneous or surgical device removal was attempted. Extracted leads were cultured, and patients received antibiotic therapy in accordance with published guidelines (16).
All patients in the registry had a diagnosis of device infection based on, at a minimum, signs of local inflammation at the generator pocket. Patients were further characterized based on results of microbiological and imaging studies. Leukocytosis was defined as leukocyte values of >11.0 × 109 cells/l. Infective endocarditis was assessed according to the modified Duke criteria (22), which include persistent bloodstream infection with positive blood cultures and visible presence of vegetation on echocardiography. Vegetation findings were defined as oscillating masses in contact with device leads or valves and detectable in at least 2 echocardiographic planes. Infective endocarditis may also be diagnosed in subjects with negative blood cultures, if there is evidence of vegetation and sufficient minor criteria are met. Patients without positive cultures were included if they showed local inflammatory markers such as erythema, warmth, purulent drainage, and/or skin erosion at the generator site.
The Charlson comorbidity index was used to evaluate comorbid conditions. The index considers 19 different disease categories and assigns numerical weights from 1 to 6 based on adjusted 1-year mortality. This index has been validated in multiple clinical contexts as a useful predictor of mortality (23).
Matching and statistical analysis
Propensity score matching was performed using R software version 3.3.1 (R Foundation for Statistical Computing, Vienna, Austria) in conjunction with MatchIt version 2.4-21 software (Harvard University, Cambridge, Massachusetts). The logistic regression for estimating propensity scores considered the following variables: Charlson comorbidity index, age of oldest remaining lead, device type, sex, and race. Nearest-neighbor matching without replacement was used to match nonabandoned lead patients with abandoned lead patients in a 2:1 ratio. Nine patients were excluded from the matching procedure due to incomplete baseline data.
Analysis of the matched cohort was performed using JMP Pro 13 (SAS, Cary, North Carolina). Demographics, comorbidities, signs and symptoms, management strategies, and outcomes were characterized using summary statistics. Specifically, mean ± SD or median (interquartile range [IQR]) for continuous variables were used. Categorical variables were presented as frequency distributions. Fisher exact tests were used to compare categorical variables between patients with and without ALs. Continuous variables from both groups were compared using Wilcoxon tests. All tests were 2-sided and a p value of <0.05 was considered statistically significant. A time-to-event analysis was performed for in-hospital mortality. The time of origin was the date of admission, and the outcome was death, and patients who survived to discharge were censored on the day of discharge. Results of the survival analysis were reported using a Kaplan-Meier plot and a Cox proportional hazards model.
A total of 434 patients were enrolled in the MEDIC registry during the study period, of whom 92 (21.3%) had ALs. Four patients with ALs and 5 without were excluded from the matching process due to missing data. The propensity-matched cohort included 264 patients: 176 with no abandoned leads (NAL group) and 88 with abandoned leads (AL group).
Demographics of the 2 groups (summarized in Table 1) were comparable due to the success of the matching procedure. Age, sex, race, device type, lead age, and Charlson comorbidity index were similar between the 2 groups. Not surprisingly, AL group patients averaged more leads overall (3.2 vs. 2.1, respectively; p < 0.001).
Signs, symptoms, and diagnostic findings differed based on the presence of abandoned leads (Table 2). At the time of admission, patients with abandoned leads had more local evidence of infection. Group AL had more positive pocket cultures (70.5% vs. 53.0%, respectively; p = 0.015) and more pocket erosion (31.8% vs. 19.3%, respectively; p = 0.031) but fewer positive blood cultures (35.6.2% vs. 54.4%, respectively; p = 0.005) and less leukocytosis (20.5% vs. 34.1%, respectively; p = 0.023). Although the following differences did not reach statistical significance, AL group trended toward presenting with more erythema, pain, pocket swelling, pocket warmth, and pocket purulence. Similarly, NAL group trended toward more fever and sepsis. A similar proportion of patients in groups NAL and AL met criteria for diagnosis of infective endocarditis (34.7% vs. 34.1%, respectively; p = 0.927) despite the differences in bacteremia rates, because both groups had vegetation in approximately 1 of 3 patients.
Overall, a pathogen was isolated from cultures of blood, generator pockets, and lead tips from 156 patients (88.6%) in NAL group and 75 patients (85.2%) in AL group. As expected, staphylococcal species were the organisms most frequently isolated from both groups (24). A detailed distribution, which shows significantly different profiles for the two groups (p = 0.035), is shown in Table 3. Coagulase-negative staphylococci were the causative organisms in 40.9% of AL group patients but only 23.3% of patients from NAL group. Cultures from the NAL group were more likely to grow enterococci or Staphylococcus aureus.
These groups also diverged with respect to device management (Table 4). Whereas complete device removal was attempted in similar ratios for both groups, more advanced tools were used for patients with abandoned leads. Roughly half of extractions in the NAL group (58.3%) required laser sheaths, but more than three-fourths (75.6%) of those in the AL group required laser sheaths (p = 0.010). The rates of manual traction use were similar in both groups, suggesting that most operators did not simply skip to more advanced tools before attempting manual traction. Although the differences did not reach statistical significance, extractions in the AL group involved nearly double the number of thoracotomies (17.1% vs. 11.5%, respectively; p = 0.238). Eleven patients in the AL group (13.4%) and 10 patients in the NAL group (6.4%) had residual lead material following extraction attempts (p = 0.092).
On average, patients in the AL group accrued more complications by a margin that was not statistically significant (Table 5). In all, 7.2% of patients in the study died before discharge. A Kaplan-Meier model (Figure 1) and Cox proportional hazards model (Table 5) showed no significant differences in survival between the 2 groups, with a risk ratio for abandoned leads of 0.34 (95% confidence interval: 0.05 to 1.22).
In this large, prospective cohort, we demonstrated that patients with CIED infections and previously abandoned leads have distinguishing characteristics in terms of presentation, progression, and management. In particular, they have: 1) ALs that are more likely to show local and/or more indolent signs of infection than those with NALs; 2) a different distribution of pathogens; and 3) need for more specialized tools for extraction.
Clinical presentation of device infection
Previous analyses from this registry have revealed that clinical presentations of device infection differ depending on the time elapsed since the last procedure and on the size of intracardiac vegetation (25–27). Abandoned leads also play a role. According to our analysis, ALs were associated with more positive pocket cultures and pocket inflammation. On the other hand, these patients are less likely to have leukocytosis or positive blood cultures. Combined with the trends toward less fever and sepsis, these data suggest the importance of physical examination findings in infection detection. It is plausible that these patients were more susceptible to pocket infection and erosion due to capped and abandoned hardware in that space. More research is urgently needed to determine the relative susceptibility of these 2 groups to infections, which this study was unable to do.
The overall predominance of staphylococcal species was consistent with that in existing reports. However, this is the first large study to suggest distinct bacteriological findings in patients with ALs. Forty-one percent of group AL patients were infected with coagulase-negative staphylococci versus 23.3% in the NAL group. It remains to be seen if these infections represent a difference in natural history or a superimposed risk for abandoned leads patients.
Lead management during infection
Rapid and complete removal of CIED systems is a proven strategy for treating device infections. The approach is now a class I recommendation in pocket infections as well as lead-associated endocarditis (12). As a result, infection cases offer a unique view into the consequences of retaining abandoned leads. Despite balanced dwell times in our matched cohort, laser sheaths were used more frequently in those with ALs. Most lead removal procedures can be performed percutaneously, but open-heart surgery is sometimes necessary. The latter approach, not surprisingly, carries a higher risk of complications and is usually reserved for patients with large intracardiac vegetation growths or those for whom percutaneous removal has failed. In a larger sample, the trend that we observed toward thoracotomies in patients with ALs could impact ultimate survival. Furthermore, the ALs in this study had a modest average dwell time of 5 years. Leads in the largest study showing higher mortality for ALs patients had an average dwell time of 10 years (20).
The relatively low overall mortality rate of 7.2% serves as a reminder that effective management is possible even in these very complicated patients. As most patients were ultimately censored, the Kaplan-Meier model likely underestimates long-term survival in both groups.
The Cox regression did not detect significant differences in mortality between the 2 groups, but we did observe a trend toward more complications in patients with ALs. As a result of matching, the comorbidities were similar in both study groups, and these findings may therefore appear generalizable. However, mortality must be considered in light of our other findings, that is, contrasting signs and symptoms, bacteriological studies, and management. We ultimately observed that these groups constitute distinct clinical populations.
The question of whether to extract or abandon a presumably sterile lead during device revision requires weighing immediate and long-term risks. With regard to device infections, the question is whether ALs increase the rate of infections or complicate subsequent management. Without reliable, large-scale data for the prevalence of patients with ALs, it is impossible to answer the first question. As to the second question, this study did not find increased mortality in the ALs cohort, but it clearly suggests that abandonment complicates subsequent management, creating additional risks for patients.
Any interpretation of management strategies is limited because all participating institutions are tertiary referral centers. In settings with fewer resources, the lead management paradigm may be different. The use of these centers for enrollment creates the possibility of referral bias. Moreover, the low proportion of female and minority patients limits the application of these findings to other populations.
Our prospective data from a large registry of CIED infections demonstrate that patients with ALs may have a different presentation of infection and require more aggressive management for complete device removal to achieve infection cure.
COMPETENCY IN MEDICAL KNOWLEDGE: Among patients with cardiovascular implantable electronic devices, individuals with abandoned leads may develop more indolent device infections and require more advanced extraction techniques.
TRANSLATIONAL OUTLOOK: Future studies are needed to evaluate the extent of lead capping and its influence on subsequent CIED infection rates.
The authors acknowledge investigators in the hospital clinics: Juan Ambrosioni, Juan M. Pericàs, Adrian Téllez, Marta Hernandez-Meneses, Asunción Moreno; Cristina Garcia de la Mària, Javier Garcia-Gonzalez; Francesc Marco, Manel Almela, Jordi Vila; Eduard Quintana, Elena Sandoval, Juan C. Paré, Carlos Falces, Daniel Pereda, Ramon Cartañá, Salvador Ninot, Manel Azqueta, Marta Sitges, Barbara Vidal, José L. Pomar, Manuel Castella, José M. Tolosana, José Ortiz; Guillermina Fita, Irene Rovira; David Fuster; Jose Ramírez; Mercè Brunet; Dolors Soy; Pedro Castro, and Jaume Llopis.
This study was funded in part by a grant from the American Heart Association. Data from the Hospital Clinic of Barcelona, Spain, were supported in part by a grant from the Ministerio de Sanidad y Consumo, the Instituto de Salud Carlos III, and Spanish Network for Research in Infectious Diseases grant REIPI RD06/0008, Madrid, Spain. Dr. Greenspon has received honoraria from or is consultant for Medtronic, Boston Scientific, and St. Jude Medical. Dr. Baddour has received royalty payments from UpToDate, Inc., for authorship duties. Dr. Peacock owns common stock in Pfizer. Dr. Sohail has received a research grant from Medtronic, Inc.; and is a consultant for Medtronic, Inc. and Boston Scientific. Dr. Carrillo has received research funding from or is a consultant for Spectranetics, Abbott, Medtronic, St. Jude Medical, Sorin group, and Sensormatic. Dr. Miró has received research funding from Institut d’Investigaciones Biomédiques August Pi i Sunyer; and has received research funding from or is a consultant for Angelini, AbbVie, Bristol-Myers Squibb, Cubist, Genentech, Medtronic, Merck, Novartis, Gilead Sciences, and ViiV healthcare. Dr. Tolosana is a consultant for Boston Scientific, Medtronic, St. Jude Medical, and Biotronik. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- cardiac implantable electronic device
- insertable cardioverter-defibrillator
- methicillin-resistant Staphylococcus aureus
- Methicillin-sensitive Staphylococcus aureus
- Received July 10, 2017.
- Revision received September 14, 2017.
- Accepted September 21, 2017.
- 2018 American College of Cardiology Foundation
- Voigt A.,
- Shalaby A.,
- Saba S.
- Klug D.,
- Balde M.,
- Pavin D.,
- et al.
- Poole J.E.,
- Gleva M.J.,
- Mela T.,
- et al.
- Wazni O.,
- Wilkoff B.L.
- Greenspon A.J.,
- Patel J.D.,
- Lau E.,
- et al.
- Baddour L.M.,
- Epstein A.E.,
- Erickson C.C.,
- et al.
- Rijal S.,
- Shah R.U.,
- Saba S.
- Saba S.
- Trucco E.,
- Mont L.
- Hussein A.A.,
- Tarakji K.G.,
- Martin D.O.,
- et al.
- Gillis A.M.
- Nagpal A.,
- Baddour L.M.,
- Sohail M.R.
- Greenspon A.J.,
- Le K.Y.,
- Prutkin J.M.,
- et al.
- Greenspon A.J.,
- Prutkin J.M.,
- Sohail M.R.,
- et al.