Author + information
- Received September 12, 2009
- Revision received November 9, 2009
- Accepted November 18, 2009
- Published online February 23, 2010.
- Douglas S. Lee, MD, PhD*,†,* (, )
- Andrew D. Krahn, MD¶,
- Jeffrey S. Healey, MD, MSc∥,
- David Birnie, MBChB#,
- Eugene Crystal, MD‡,
- Paul Dorian, MD§,
- Christopher S. Simpson, MD**,
- Yaariv Khaykin, MD††,
- Douglas Cameron, MD†,
- Amir Janmohamed, MD‡‡,
- Raymond Yee, MD¶,
- Peter C. Austin, PhD*,
- Zhongliang Chen, MD, MSc*,
- Judy Hardy, RN*,
- Jack V. Tu, MD, PhD*,‡,
- Investigators of the Ontario ICD Database
- ↵*Reprint requests and correspondence:
Dr. Douglas S. Lee, Institute for Clinical Evaluative Sciences, Room G-106, 2075 Bayview Avenue, Toronto, Ontario M4N 3M5, Canada
Objectives This study examined the predictors of early complications after defibrillator implantation.
Background Although implantable cardioverter-defibrillators are widely used, predictors of procedural complications and the consequences of these events have not been determined.
Methods In a prospective, multicenter, population-based clinical outcomes registry of all newly implanted defibrillator patients at 18 centers in Ontario, Canada, we examined 45-day complications and all-cause mortality from February 2007 to May 2009. Complications were determined longitudinally and were categorized as direct implant-related or indirect events.
Results Among 3,340 patients (mean age 63.8 ± 12.5 years, 78.5% men), major complications occurred in 4.1% of de novo procedures. Compared with those undergoing a single-chamber device, implantation of a cardiac resynchronization defibrillator (adjusted hazard ratio [HR]: 2.17, 95% confidence interval [CI]: 1.38 to 3.43, p < 0.001) or dual-chamber device (adjusted HR: 1.82, 95% CI: 1.19 to 2.79, p = 0.006) was associated with increased risk of major complications. Major complications were increased in women (adjusted HR: 1.49, 95% CI: 1.02 to 2.16, p = 0.037) and when left ventricular end-systolic dimension exceeded 45 mm (adjusted HR: 1.54, 95% CI: 1.08 to 2.20, p = 0.018). Major complications (excluding death) occurring early after defibrillator implantation were associated with increased adjusted risk of subsequent death up to 180 days after defibrillator implant (adjusted HR: 3.70, 95% CI: 1.64 to 8.33, p = 0.002). Direct implant-related complications were associated with increased risk of early death (adjusted HR: 24.89, p = 0.01), whereas indirect clinical complications conferred increased risk of near-term death (adjusted HR: 12.35, p < 0.001) after defibrillator implantation.
Conclusions Complications after de novo defibrillator implantation were strongly associated with device type. Major complications were associated with increased risk of mortality.
Sudden cardiac events due to ventricular tachyarrhythmia are important causes of premature death. Implantable cardioverter-defibrillators (ICDs) have been found to reduce the risk of arrhythmic death and total mortality in patients with left ventricular systolic dysfunction and selected patients with prior cardiac arrest or sustained ventricular tachycardia (1–4). In response to the broadening indication for these devices, the use of ICDs has increased to include patients who are asymptomatic but could be at increased risk of sudden cardiac death.
Implantation of a cardioverter-defibrillator entails a surgical procedure with the attendant potential for complications. Little is known about the rates and predictors of complications after ICD implantation. Indeed, patients who undergo defibrillator implantation have significant cardiac conditions and noncardiac comorbidities and may therefore be at higher risk of developing complications from the procedure than has been demonstrated in randomized trials (5). Furthermore, the types of devices being used currently also differ from those used in the clinical trials, with greater use of dual-chamber devices or cardiac resynchronization defibrillators. Complications of ICD implantation are not only important from the standpoint of patient care, but they also add significantly to length of hospital stay, increase health care costs, and influence the defibrillator's cost-effectiveness (6).
In this study, we examined the outcomes of patients undergoing defibrillator implantation in the Ontario ICD Database, a population-based, prospective registry of patients undergoing device implantation. In de novo defibrillator recipients undergoing the procedure for primary or secondary prevention, we examined the frequency of these early complications and their predictors. We hypothesized that the burden of complications after defibrillator implantation would be predicted by both patient and device characteristics.
We examined patients undergoing de novo implantation of a cardioverter defibrillator (e.g., for primary or secondary prevention indication) in the Ontario ICD Database between February 2007 and May 2009. The design and methodology of the Ontario ICD Database have been published elsewhere (7). Briefly, this project is an ongoing prospective clinical registry of all patients (age ≥16 years) undergoing defibrillator implantation in Ontario, Canada, which has been mandated by the Ontario Ministry of Health and Long-Term Care. Patient data are collected upon initial evaluation, at the time of defibrillator implant, and at each follow-up visit in the device clinic. As a prescribed entity, we were able to collect data on all patients in this registry without informed consent, and therefore all patients had their data entered into the database without participation bias (8).
Details of the primary data collection process for the Ontario ICD Database have been described previously (7). Briefly, data in the ICD database were collected by the electrophysiologist and a trained research coordinator at defibrillator implantation centers, who entered the data into a secure, firewall- and password-protected Web-based registry at the Institute for Clinical Evaluative Sciences. We collected data on patient characteristics, indication for the defibrillator, comorbidities, left ventricular ejection fraction, and implant-related data. We also collected clinical and device-related events occurring in follow-up, with data entered at the implanting hospital or peripheral ambulatory device clinics. Data quality was continually assessed by: 1) regular review and correspondence with study sites to ensure accuracy of coded data; 2) automated range checks, notification to study sites of uncoded data elements followed by recoding; 3) logic checks of key data elements including dates of all events; and 4) ongoing random site audits for reliability and data collection mechanisms.
Deaths were identified by clinical follow-up at routine defibrillator clinic visits and via the Registered Persons Database linked using the patient's unique encrypted health card number. Vital status data were available for all study patients. Ethics approval was obtained from the Sunnybrook Health Sciences Centre before study initiation. A complete list of participating defibrillator implant and follow-up sites is shown in Online Appendix 1.
At defibrillator clinic visits after device implantation, patients were routinely evaluated for any device-related complications that occurred within 45 days of the procedure using a standard checklist. We selected 45 days for determination of complications for several reasons: 1) it was temporally close to the device implant date; 2) all sites conducted an initial post-implant clinic visit within 45 days; and 3) most early complications occurred during this period, thus maximizing the potential capture of events (9). Because the date of the initial post-implant clinic visit could vary, we cross-referenced the recorded dates of complications from subsequent clinic visits and counted those events occurring temporally within 45 days following implantation. These complications were categorized as major or minor (Table 1),based on the consensus of electrophysiologists from all participating hospitals, and was guided by a previous report (10). In general, major complications were distinguished from minor events by the need for an operation to revise the lead or device or for therapeutic relief. Clinical complications were considered to be major if hospitalization or substantive parenteral therapy were required. Major complications were subclassified as direct (implant-related) or indirect (clinical disease) events. Major complications included death unless otherwise specified. Patients were deemed to have had “any complication” if they experienced a major or a minor complication.
Factors associated with complications
The variables considered in the univariate analysis were determined after review of the literature and included age, sex, prior ventricular tachycardia, and primary disease etiology (e.g., ischemic, nonischemic cardiomyopathy, or other conditions such as congenital heart disease or hypertrophic cardiomyopathy). Cardiac factors included prior myocardial infarction, revascularization procedure, heart failure, New York Heart Association functional class, pre-existing pacemaker system, and atrial fibrillation. Noncardiac factors included general laboratory measures (e.g., serum creatinine concentration, estimated glomerular filtration rate, serum sodium, hemoglobin) and comorbidities. Medications were also examined, including angiotensin-converting enzyme inhibitor or angiotensin-receptor blocker, beta-adrenoreceptor antagonists, loop diuretic, antiplatelet or anticoagulant drug, amiodarone, and class IC or other class III antiarrhythmic drugs. Diagnostic investigations included QRS interval duration (>140 ms vs. ≤140 ms), left ventricular ejection fraction (≤20%, 21% to 35%, or >35%), left ventricular dimension in end-systole or end-diastole, and left atrial size. We also examined the type of ICD implanted (e.g., dual-chamber, cardiac resynchronization therapy defibrillator device [CRT-D], or single-chamber device), active versus passive fixation, and single versus dual coil right ventricular lead.
We performed descriptive analyses of patients undergoing ICD implant for primary or secondary prevention indications. We then performed univariate Cox regression analysis to determine predictors of time to complication after ICD implant. Potential predictors with p < 0.25 on univariate analysis were included in the multiple regression model using a stepwise approach, and predictors with p < 0.05 in the multivariable model were retained. The final model selection was based on clinical and statistical significance. All models included age irrespective of statistical significance, because of the potential of this variable to account for complications.
We determined the association of major complications excluding death (e.g., major nondeath complications) with mortality employing complication status as a time-varying covariate in Cox regression analysis. In the time-varying model, patients were considered complication-free until a major complication occurred, at which time they were reclassified. Complications occurring more than 45 days after implantation were not included in the determination of exposure status. We determined the association of major complications with death occurring up to 180 days after defibrillator implant adjusted for age, sex, and ICD type. The preceding analyses were also repeated for minor complications. Continuous data are presented as mean ± SD and compared using Student ttest. Categorical variables were compared using the chi-square test, Fisher exact test, or the Mantel-Haenszel test for trend. Exact 95% confidence intervals (CIs) were determined using published methods. All analyses were performed using SAS software, version 9.1 (SAS Institute, Cary, North Carolina).
Among a total of 4,217 referrals for de novo ICD from February 2007 to May 15, 2009, patients were excluded because they refused the procedure (n = 183), underwent the procedure within 45 days of the last follow-up date (n = 273), or did not meet criteria for the device (n = 421). The final study cohort consisted of 3,340 patients, and their characteristics are shown in Table 2according to primary or secondary prevention status. Follow-up data for mortality and complications were available for all study patients. The majority of patients had ischemic heart disease, and nearly two-thirds of device implants were for primary prevention indications.
A total of 172 primary (7.4%) and 76 secondary (7.6%) prevention patients experienced any complication. There was no significant difference in overall complication rates by primary or secondary prevention indications. Major complications occurred in 90 patients (3.8%) undergoing primary prevention ICD and 48 (4.8%) who underwent secondary prevention device implantation. Minor complications occurred in 125 (5.3%) and 45 (4.5%) of those undergoing primary or secondary prevention defibrillator implantation. Only a small proportion of complications occurred in-hospital, at rates of 0.5% for major, 0.5% for minor, and 0.8% for any complication.
Univariate predictors of complications
Univariate predictors of major and minor complications are shown in Online Appendixes 2 and 3, respectively. The most significant factor associated with complications was the type of device implanted, with high risk in those undergoing dual-chamber device and the highest risk for CRT-D (Fig. 1).The number of new leads implanted was also a significant predictor of major complications, with hazard ratios of 1.78 (95% CI: 1.21 to 2.63) for 2 leads and 2.17 (95% CI: 1.38 to 3.40) for 3 leads, relative to 1 lead (p = 0.004 and p < 0.001, respectively). Hazard ratios (HRs) for minor complications were 1.51 (95% CI: 1.06 to 2.15, p = 0.023) and 2.42 (95% CI: 1.64 to 3.56, p < 0.001) for 2 and 3 leads, respectively. The risk of major complications was also increased in women (HR: 1.57, 95% CI: 1.09 to 2.26) and with New York Heart Association functional class III to IV (HR: 1.47, 95% CI: 1.04 to 2.08).
Minor complications were more likely to occur in women (HR: 1.40, 95% CI: 1.00 to 1.96), those with New York Heart Association functional class III to IV (HR: 1.73, 95% CI: 1.27 to 2.35), or those with nonischemic dilated cardiomyopathy (HR: 2.21, 95% CI: 1.35 to 3.61, when compared with ischemic). Univariate analyses for primary prevention device were similar to those for de novo implants; however, anticoagulant drug use was a significant predictor of complications in the former group. Among those taking oral anticoagulants, aspirin or clopidogrel, the HRs were 2.15 (95% CI: 1.08 to 4.30, p = 0.030) for major and 1.85 (95% CI: 1.06 to 3.24, p = 0.031) for minor complications.
Multivariable predictors of complications
Multivariable predictors of complications among de novo ICD recipients are shown in Table 3.Age was not significantly associated with major or minor complications and therefore, these hazard ratios are not shown. Dual-chamber and CRT-D defibrillators were the most significant predictors of major complications, with >1.5- and 2-fold increase in risk, respectively, compared with those receiving a single-chamber device. Increased left ventricular end-diastolic dimension ≥45 mm was associated with a 1.5-fold increase in risk. Women were more likely to develop major complications than men (Table 3). Minor complications were also increased in those receiving dual-chamber and CRT-D devices, and in those with nonischemic cardiac disease etiologies (Table 3).
In the subset of patients undergoing primary prevention ICD implantation, device type had an even greater effect on major complications with a 2.6- and 3.2-fold risk for dual-chamber and CRT-D, respectively (Table 4).Prior use of antiplatelet agents or anticoagulants increased the risk of major complications, as did the use of class IC or class III antiarrhythmic drugs other than amiodarone (Table 4).
Types of complications
The complications occurring with highest frequency are shown in Table 5.In some cases, multiple complications occurred in the same patient. The most frequent major complications were lead-related, followed by pocket infection, electrical storm, and pulmonary edema requiring hospitalization. The majority of lead replacements (76 of 94 = 81%) and lead extractions (21 of 23 = 91%) were not preceded by a lead recall. Incisional infections and pocket hematomas that did not require an operative procedure (medically managed) were the most common minor complications.
Association of complications with death
Patients who experienced a major complication (excluding death) had a significantly higher rate of subsequent mortality than those without a major complication within the first 45 days after ICD implantation. Crude mortality rates were 2.2% versus 0.5% at 45 days (p = 0.036), 2.9% versus 1.2% at 90 days (p = 0.09), and 5.1% versus 1.8% at 180 days (p = 0.018) for those with versus those without a major nonfatal complication within the first 45 days. Adjusted mortality was significantly increased in those who experienced a major complication within 45 days, but not in those with minor complications (Table 6).
In further analyses, we found that mortality rates were higher among those who experienced any major complication within the 45-day perioperative period compared with those without complications (Fig. 2).The HR for direct implant-related complications adjusted for age, sex, and ICD type was 24.89 (95% CI: 2.11 to 294.26) in the first 45 days following defibrillator implantation (p = 0.01). Patients with indirect clinical complications (excluding death) occurring within 45 days of ICD implant were associated with substantially higher mortality rates when compared with those patients without early complications (Fig. 2). The adjusted HR for indirect clinical complications was 12.35 (95% CI: 4.74 to 32.18) in the 180-day follow-up period after defibrillator implantation (p < 0.001).
In this population-based registry, we found that perioperative complications were not infrequent, with 4.1% experiencing major complications after de novo defibrillator implantation. The predominant factor associated with both major and minor complications was the type of device implanted, with increased risk in those receiving dual-chamber devices and the highest risk after cardiac resynchronization defibrillator implantation. Women and those with an enlarged left ventricular chamber size were at increased risk of complications after de novo defibrillator implantation. Use of anticoagulants or antiplatelet agents and antiarrhythmic drugs also predicted increased risk of complications among those undergoing a primary prevention defibrillator. The occurrence of a major, early complication was associated with an increased risk of death with a 3- to 4-fold risk up to 6 months after device implantation. Stratified by the type of major complication, we found that both direct implant-related and indirect clinical complications were associated with significantly increased risk of death.
Previous studies have reported on selected complications after defibrillator implantation. A German ICD registry reported rates of specific complications and found that pocket hematoma, chronic pain, and lead and device dislodgements leading to operative revisions were the most common events (11). However, participation in this registry was voluntary with considerable losses to follow-up (11). A single-center study spanning a decade of defibrillator implantations reported that 10% of patients experienced a complication within 30 days and that biventricular and dual-chamber devices were associated with more lead-related complications (12). However, this study was small (440 patients) and was limited in the ability to define independent predictors of major complications (12). A larger multicenter study reported that lead- and pocket-related complications occurred in 2.1% and 1.8% of cases, respectively, but predictors of these events were not determined (13). The National Cardiovascular Data Registry ICD Registry reported in-hospital complication rates of 1.3% for major and 3.6% for any complication (14); that registry differed from our study because reporting was not mandatory from all hospitals (15,16), in-hospital complications were the primary events counted, and lead complications were not included (14).
Previous studies evaluating defibrillator complications using large administrative databases have reported different event rates and predictors. Using Medicare administrative databases, 90-day complications were reported to have occurred in 14.2% of patients (17). Many of the major complications in this study overlapped with ours; however, the predictors differed and were predominantly noncardiac comorbidities (17). Our study differed because we considered an early post-procedural time window, which reflects complications related to device implantation, and furthermore, we used clinically obtained data on predictors and events. Importantly, another study of Medicare beneficiaries found that in-hospital complications occurred in 10.8% of ICD recipients and led to significant increases in length of hospital stay and costs compared with an uncomplicated procedure (18).
Our study is novel because the data were collected in a prospective, population-based registry, with detailed longitudinal follow-up. We also examined outcomes bimodally, using a combination of both clinical device and passive administrative data-based follow-up, allowing for capture of important outcomes including death. Our study adds to the current literature in several ways. First, we examined complications that occurred in the early post-defibrillator implant period including events following discharge (19). Second, this registry was mandated by the administrator of health care services in Ontario, and participation from all defibrillator implanting centers was required. Thus, our study was not subject to volunteer bias, and we were able to study all patients irrespective of age. Finally, a wide array of potential clinical predictors and major and minor complications, including death, were systematically evaluated.
The potential mechanisms of the mortality association with major complications may be a direct or indirect result of complications. The early mortality effects of major complications may be the direct consequence of a mechanical procedure-related event (e.g., myocardial perforation). These mechanical complications were potentially related to increasing procedural complexity with greater number of leads implanted, and anticoagulation or antiplatelet agents in the subgroup of patients undergoing primary prevention defibrillators. The latter is particularly relevant given the tendency of some to implant defibrillators without discontinuation of oral anticoagulants combined with bridge therapy (20,21). The mortality effects of indirect clinical complications began early and were maintained up to 180 days following implantation, likely reflecting progressive disease of the underlying cardiovascular disease substrate that predisposed to a higher a priori risk of death. The association of enlarged left ventricular chamber size with complications may reflect altered biventricular and atrial geometry, which confer both an increased propensity to both mechanical complications and increased clinical risk. We have previously shown that heart failure events contribute significantly to reduced survival (22), and progressive heart failure substantially increases the risk of death among defibrillator recipients who have experienced an appropriate therapy (23), signifying the importance of this condition on outcomes after device implantation. High mortality rates have been reported in those with electrical storm (24,25), but the data are conflicting (26), and prior studies have not examined this in the context of complications arising in the perioperative phase. The use of antiarrhythmic drugs reflects underlying abnormalities in the electrical substrate, which may have conferred an increase in clinical complications because of more severe structural disease.
Few studies have systematically examined the predictors of major and minor complications according to indication for the ICD using multivariable analysis. Our study suggests that early complications of ICD implantation are not infrequent and suggest an increased risk of death if major complications occurred. These data also suggest that complications are driven in part by the underlying sickness of the patient, and these factors may be attenuated by optimization of the underlying cardiac disease before defibrillator implantation. This may be most relevant for patients who are referred for CRT-D, who are those at greatest risk of complications driven by heart failure and their underlying cardiac status. Although near-term death was infrequent in defibrillator recipients, the association should also be examined in other cohorts because of the importance of this outcome. Minor complications remain clinically important because these events may influence quality of life, but further data are needed to determine the degree of this association. Finally, it should carefully be considered whether patients require more complex, multilead devices as opposed to simpler devices with fewer leads.
Although we identified a number of predictors of complications, we could not determine the causal pathway by which these factors exerted their effects. It is possible that these effects may be complex, exerting their effects on subcomplications by a multitude of mechanisms. We did not examine intraoperative factors such as procedure duration; however, it would likely have been longer with implantation of more complex devices. We did not examine those undergoing replacement devices, for whom the predictors and consequences of complications may differ substantially from these findings. Lastly, our data suggested an association of major complications with mortality, but could not determine whether the complication was causally related to death.
A significant proportion of patients undergoing de novo implantation of cardioverter defibrillators experienced complications. The complexity of device implanted, female sex, and enlarged left ventricular dimension were significant predictors of major complications. Because direct implant-related complications and indirect disease-related major perioperative events were associated with an increased risk of death, further evaluations of early defibrillator complications, their potential effects on mortality, and preventative strategies are needed.
For a list of participating hospitals and investigators in the Ontario ICD Database and the univariate predictors of major and minor complications, please see the online version of this article.
The Institute for Clinical Evaluative Sciences (ICES) is supported in part by a grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). The opinions, results and conclusions are those of the authors and no endorsement by the MOHLTC or by the ICES is intended or should be inferred. This research was supported by a grant from the MOHLTC, a Canadian Institutes of Health Research (CIHR) clinician-scientist award (to Dr. Lee), a Career Investigator award from the Heart and Stroke Foundation of Ontario (to Dr. Austin), and a Canada Research Chair in Health Services Research (to Dr. Tu). Dr. Krahn has received research funding and is a consultant to Boston Scientific. Dr. Healey has received research funding from Boston Scientific. Dr. Birnie is a consultant to Medtronic and participates in clinical trials sponsored by Medtronic, Boston Scientific, and St. Jude Medical. Dr. Crystal has received educational grants from Medtronic, Boston Scientific, and St. Jude. Dr. Simpson has received speaker's fees from Medtronic. Dr. Khaykin has received honoraria and speaker's fees for St. Jude Medical and Medtronic. Dr. Yee is a consultant for and on the Speakers' Bureau of Medtronic.
- Abbreviations and Acronyms
- confidence interval
- hazard ratio
- cardiac resynchronization therapy defibrillator
- implantable cardioverter-defibrillator
- Received September 12, 2009.
- Revision received November 9, 2009.
- Accepted November 18, 2009.
- American College of Cardiology Foundation
- Connolly S.J.,
- Gent M.,
- Roberts R.S.,
- et al.
- Lee D.S.,
- Tu J.V.,
- Austin P.C.,
- et al.
- Peterson P.N.,
- Daugherty S.L.,
- Wang Y.,
- et al.
- Anderson K.P.
- Gibbs J.L.,
- Monro J.L.,
- Cunningham D.,
- Rickards A.
- Al-Khatib S.M.,
- Greiner M.A.,
- Peterson E.D.,
- Hernandez A.F.,
- Schulman K.A.,
- Curtis L.H.
- Reynolds M.R.,
- Cohen D.J.,
- Kugelmass A.D.,
- et al.
- Tolosana J.M.,
- Berne P.,
- Mont L.,
- et al.
- Exner D.V.,
- Pinski S.L.,
- Wyse D.G.,
- et al.,
- AVID Investigators
- Brigadeau F.,
- Kouakam C.,
- Klug D.,
- et al.