Cardiac perforation is a feared complication of transvenous. Original Article

Transcription

1 Original Article Cardiac Perforation From Implantable Cardioverter- Defibrillator Lead Placement Insights From the National Cardiovascular Data Registry Jonathan C. Hsu, MD, MAS; Paul D. Varosy, MD; Haikun Bao, PhD; Thomas A. Dewland, MD; Jeptha P. Curtis, MD; Gregory M. Marcus, MD, MAS Background Cardiac perforation is a feared complication of implantable cardioverter-defibrillator (ICD) lead implantation because of the potential for significant morbidity and mortality. Predictors of perforation and the severity of associated adverse events have not been well studied. We sought to identify predictors of cardiac perforation from ICD lead implantation and subsequent outcomes. Methods and Results We studied first-time ICD recipients in the ICD Registry implanted between January 2006 and September Using hierarchical multivariable logistic regression adjusting for patient, implanting physician, and hospital characteristics, we examined the predictors of perforation and the association of perforation with other major complications, length of stay, and in-hospital mortality. Cardiac perforation occurred in 625 patients (0.14%). After multivariable adjustment, older age, female sex, left bundle branch block, worsened heart failure class, higher left ventricular ejection fraction, and non single-chamber ICD implant were associated with a greater odds of perforation. Conversely, atrial fibrillation, diabetes mellitus, previous cardiac bypass surgery, and higher implanter procedural volume were associated with a lower odds of perforation (all P<0.05). After adjustment, ICD recipients with perforation had greater odds of other associated major complications (odds ratio, 27.5; 95% confidence interval, ; P<0.0001), postprocedural hospital stays >3 days (odds ratio, 16.3; 95% confidence interval, ; P<0.0001), and in-hospital death (odds ratio, 17.7; 95% confidence interval, ; P<0.0001). Conclusions In a large population of ICD recipients, specific patient and implanter characteristics predicted cardiac perforation risk. Cardiac perforation was associated with a substantially increased risk of other major complications, prolonged hospital stays, and death. (Circ Cardiovasc Qual Outcomes. 2013;6: ) Key Words: adverse events complication cardiac perforation complication defibrillators, implantable mortality national registry Cardiac perforation is a feared complication of transvenous pacemaker and implantable cardioverter-defibrillator (ICD) lead implantation because of the potential for significant morbidity and mortality. 1 Small studies and registries with older lead technologies have estimated the prevalence of cardiac perforation to be 0.1% to 0.8% for pacemakers and 0.6% to 5.2% for ICDs, 2 5 but no large-scale study has evaluated the prevalence of cardiac perforation during implantation of modern-day ICDs. Importantly, predictors of cardiac perforation from ICD implantation and the severity of adverse events that result from the complication have not been well studied. We analyzed data from the ICD Registry of the National Cardiovascular Data Registry (NCDR), a national registry of ICD implantations that captures detailed clinical patient, physician, and hospital information as well as in-hospital outcomes. By assessing a large population of ICD recipients, we sought to examine predictors of cardiac perforation from ICD implantation and the magnitude of associated adverse events. Methods Data Source The NCDR ICD Registry was created in 2006 to meet the requirements of the Center for Medicare & Medicaid Services Coverage with Evidence Development decision. 6 The Heart Rhythm Society and American College of Cardiology collaborated to establish a national registry of ICD implantations funded by a combination of hospital fees and grants from payers and device companies. Hospitals are mandated to provide data on Medicare beneficiaries receiving an ICD for primary prevention of sudden cardiac death; however, 71.5% Received April 13, 2013; accepted August 13, From the Section of Cardiac Electrophysiology, Division of Cardiology, Department of Medicine, University of California, San Diego, CA (J.C.H.); Section of Cardiac Electrophysiology, Division of Cardiology, Department of Medicine, University of California, San Francisco, CA (T.A.D., G.M.M.); VA Eastern Colorado Health Care System, University of Colorado, Denver, CO (P.D.V.); the Colorado Cardiovascular Outcomes Research Group, Denver, CO (P.D.V.); and the Section of Cardiovascular Medicine, Department of Internal Medicine, Yale University School of Medicine, New Haven, CT (H.B., J.P.C.). This manuscript was handled independently by Sana Al-Khatib, MD, MHS, as a Guest Editor. The Editors had no role in the evaluation of this manuscript or in the decision about its acceptance. Correspondence to Jonathan C. Hsu, MD, MAS, University of California, San Diego, 9444 Medical Center Dr, MC7411, La Jolla, CA Jonathan.Hsu@ucsd.edu 2013 American Heart Association, Inc. Circ Cardiovasc Qual Outcomes is available at DOI: /CIRCOUTCOMES

2 Hsu et al Cardiac Perforation and ICD Adverse Events 583 WHAT IS KNOWN It is known that cardiac perforation is a complication that may occur during implantable cardioverterdefibrillator implantation. It is known that cardiac perforation may be associated with increased morbidity and mortality, but the magnitude of this association has never been studied. WHAT THE STUDY ADDS The prevalence of cardiac perforation during modern-day implantable cardioverter-defibrillator implantation is 0.14%. Specific patient and implanter characteristics predict cardiac perforation risk; older age, female sex, left bundle branch block, worsened heart failure class, higher left ventricular ejection fraction, and non single-chamber implantable cardioverter-defibrillator implant are associated with a greater odds of perforation, whereas atrial fibrillation, diabetes mellitus, previous cardiac bypass surgery, and higher implanter procedural volume are associated with a lower odds of perforation. Cardiac perforation from implantable cardioverterdefibrillator implantation is associated with a substantially increased risk of other major in-hospital complications, prolonged hospitalization, and death. of the 1375 participating hospitals are providing data on all patients undergoing ICD implantation, and these hospitals submit 88.4% of all ICD implants included in the registry. 6 This study included all eligible patients enrolled in the ICD Registry after January 2006, which included Versions 1.0 and 2.0 data. ICD Registry data have undergone data quality standard testing, including an auditing program to confirm completeness and verify accuracy, as previously detailed. 6,7 Study Population All patients with implant data submitted to the registry between January 1, 2006, and September 30, 2011, were considered for analysis (n= ). Patients implanted with a previous ICD or cardiac resynchronization therapy with defibrillator (n= ), those with a previous pacemaker (n=53 970), and those with an epicardial lead placed during the index procedure (n=6843) were excluded from the analysis, leaving patients. Definition of Cardiac Perforation Cardiac perforation was defined in data coding as puncture or migration of a lead through the myocardium that may or may not have been self-sealing or symptomatic and could have occurred between the start of the procedure until hospital discharge. To account for potential variations in data coding, patients who experienced cardiac tamponade (fluid in the pericardial space compromising cardiac filling and requiring intervention at any occurrence between the start of the procedure until discharge) were also considered to have experienced a cardiac perforation. Adverse Outcomes The first adverse outcome of interest was the occurrence of any associated major in-hospital complication during or after ICD implantation. An associated major complication was defined as a plausibly related complication to cardiac perforation and included cardiac arrest, myocardial infarction, or infection requiring antibiotics. The second outcome was duration of hospitalization from device implant to discharge. For analytic purposes, length of hospital stay was dichotomized to >3 days or 3 days based on the distribution of the cohort. The third outcome was the occurrence of in-hospital death during or after ICD implantation. Statistical Analysis Normally distributed continuous variables are expressed as means and SDs, whereas continuous variables with skewed distributions are expressed as medians and interquartile ranges. Baseline patient, implanting physician, and hospital characteristics were compared between patients with and without cardiac perforation using the χ 2 test for categorical variables and t tests or Wilcoxon rank-sum test for continuous variables, as appropriate. Independent associations of various characteristics with the outcome of cardiac perforation were identified using a hierarchical logistic regression model to account for clustering of patients within hospitals. Covariates selected for the multivariate analyses were chosen based on the plausibility that they could be associated with cardiac perforation. In the next stage of the analysis, hierarchical logistic regression models were developed to assess the independent association of cardiac perforation with adverse outcomes, again by accounting for clustering of patients within hospitals in multivariable models. 8 Covariates in the multivariable models assessing adverse outcomes included patient demographics (age, sex, race, insurance payer), comorbidities (congestive heart failure, New York Heart Association class, syncope, ventricular tachycardia, cardiac arrest, atrial fibrillation, nonischemic cardiomyopathy, ischemic heart disease, previous myocardial infarction, previous coronary artery bypass graft surgery, previous percutaneous coronary intervention, cerebrovascular disease, chronic lung disease, diabetes mellitus, hypertension, end-stage renal disease), diagnostic information (indication for implantation, device type, left ventricular ejection fraction [LVEF], QRS duration, blood urea nitrogen level), implanter characteristics (implanter specialty training, implanter volume), and hospital characteristics (profit type, region, size, teaching status). We performed an exploratory subset analysis in patients with only Version 2.0 (after April 1, 2010) data that included right ventricular endocardial defibrillation lead information to assess a potential association of right ventricular lead characteristics (largest lead diameter and active versus passive fixation) and the risk of cardiac perforation. In this subset analysis, lead characteristics were added to the covariates included in the main analytic multivariable models to assess for an independent association between lead characteristics and cardiac perforation. We also performed a prespecified subgroup adjusted analysis to determine specific groups at higher or lower risk of any associated complication or in-hospital death from cardiac perforation, including tests for interactions. Statistical tests were 2-sided and considered significant if they yielded a P<0.05. Analyses were performed using the SAS Statistical Package Version 9.2 (SAS Institute, Cary, NC). The Yale University Human Investigation Committee approved analysis of data from the ICD Registry. Results Cardiac perforation occurred in 625 patients (0.14%). Baseline characteristics of ICD recipients with and without cardiac perforation are presented in Table 1. 9 In unadjusted analysis, variables associated with cardiac perforation included older age, female sex, worsened New York Heart Association heart failure class, and nonischemic cardiomyopathy. Cardiac perforation occurred less often in ICD recipients with a history of atrial fibrillation, ischemic heart disease, previous myocardial infarction, previous coronary artery bypass surgery, and diabetes mellitus. Recipients of cardiac resynchronization therapy with defibrillator were more likely to experience cardiac perforation compared with single- or dual-chamber ICD recipients. Clinical characteristics associated with cardiac perforation included a left bundle-branch block. Among implanting physician and hospital characteristics, lower physician

3 584 Circ Cardiovasc Qual Outcomes September 2013 Table 1. Unadjusted Baseline Characteristics of Patients With and Without Cardiac Perforation From ICD Implantation Characteristic Cardiac Perforation (n=625) No Cardiac Perforation (n= ) P Value Patient demographic characteristics Age, y 70.1± ±13.2 < Female sex 54.2% 27.6% < Race White 81.1% 80.7% Black 13.0% 14.4% Asian 1.4% 1.2% Other 4.5% 3.7% Hispanic ethnicity 5.6% 5.6% Congestive heart failure 80.6% 76.9% NYHA class < I 10.6% 13.3% II 28.3% 36.6% III 56.3% 46.1% IV 4.6% 3.8% Atrial fibrillation/atrial flutter 22.9% 28.5% Nonischemic dilated cardiomyopathy 48.3% 33.4% < Ischemic heart disease 48.0% 63.0% < Previous myocardial infarction 37.3% 51.8% < Previous CABG 2.6% 31.6% < Hypertension 74.2% 76.9% Diabetes mellitus 28.2% 37.7% < Cerebrovascular disease 12.5% 14.3% End-stage renal disease 3.8% 3.9% Patient diagnostic data LVEF, % 28.3± ± LBBB 46.7% 25.9% < BUN level, mg/dl 23.5± ± Device characteristics ICD type < Single-chamber ICD 13.9% 25.0% Dual-chamber ICD 41.3% 43.1% CRT-D 44.3% 31.8% Implanter characteristics Implanting physician training Board-certified EP 54.7% 58.2% Board-eligible EP 9.1% 7.8% HRS guidelines 9.0% 8.2% Surgery boards 1.1% 1.4% None of the above 8.0% 6.9% Unknown 18.1% 17.4% Implanting physician volume 91.3± ± Hospital characteristics Hospital type Public 10.1% 9.5% Not-for-profit 72.8% 74.1% Private 14.2% 13.2% (continued)

4 Hsu et al Cardiac Perforation and ICD Adverse Events 585 Table 1. Continued. Characteristic Cardiac Perforation (n=625) No Cardiac Perforation (n= ) P Value Unknown 2.9% 3.2% Teaching hospital 58.4% 57.6% Categorical variables are reported as proportions (%); continuous variables are reported as mean±sd. Implanting physician volume reflects average number of cases reported to the ICD Registry by implanters during the time period studied. Boardeligible EP physicians are physicians who have completed a formal cardiac EP training program before passing the American Board of Internal Medicine certification examination in cardiac EP. Heart Rhythm Society (HRS) guidelines represent an alternative training pathway to EP board eligibility/certification for implantation of ICDs. 9 BUN indicates blood urea nitrogen; CABG, coronary artery bypass grafting; CRT-D, cardiac resynchronization therapy with defibrillator; EP, electrophysiology; ICD, implantable cardioverter-defibrillator; LBBB, left bundle branch block; LVEF, left ventricular ejection fraction; and NYHA, New York Heart Association. volume was associated with cardiac perforation, whereas physician training, hospital profit type, and teaching hospitals were not associated with cardiac perforation. After multivariable adjustment, older age, female sex, worsened New York Heart Association heart failure class, non single-chamber ICD, higher LVEF, and left bundle branch block were independently associated with a greater odds of cardiac perforation (Table 2). Conversely, a history of atrial fibrillation/ atrial flutter, diabetes mellitus, previous coronary artery bypass surgery, and higher implanting physician procedural volume were associated with a lower odds of cardiac perforation. In a subset analysis restricted to the cohort of ICD recipients implanted after April 2010 with available right ventricular lead characteristics (n= total patients with 144 cardiac perforations [0.14%]), a smaller lead diameter was associated with cardiac perforation, but there was no significant difference in cardiac perforation based on the lead fixation type (Table 3). In multivariable models that included the right ventricular lead characteristics along with variables plausibly associated with cardiac perforation, a smaller lead diameter remained associated with an increased cardiac perforation risk. When right ventricular lead diameter was instead categorized by incremental French size, the greater adjusted odds of cardiac perforation in both <7F (odds ratio, 1.51; 95% confidence interval, ; P=0.5185) and 7F to 8F (odds ratio, 1.55; 95% confidence interval, ; P=0.1840) leads were not statistically different than >8F leads. Association of Cardiac Perforation With Other Procedural Complications As shown in Table 4, the crude risk of any other associated major procedural complication (defined as cardiac arrest, myocardial infarction, or infection requiring antibiotics) was higher in those with cardiac perforation (7.4%) than in those without (0.3%). Most of the other complications listed on the ICD Registry forms, including pneumothorax, hemothorax, Table 2. Statistically Significant Predictors Associated With Cardiac Perforation From ICD Implantation After Multivariable Adjustment Variable Adjusted Odds Ratio 95% Confidence Interval P Value Age (per 10 y increase) < Female sex < NYHA class I Reference II III IV ICD type Single-chamber ICD Reference Dual-chamber ICD CRT-D LVEF (per 5% increase) LBBB < Atrial fibrillation/atrial flutter Diabetes mellitus Previous CABG < Implanting physician volume The adjusted odds ratio reported for implanting physician volume is per 10 ICD implant procedures performed during the time period studied. CABG indicates coronary artery bypass grafting; CRT-D, cardiac resynchronization therapy with defibrillator; ICD, implantable cardioverter-defibrillator; LBBB, left bundle branch block; LVEF, left ventricular ejection fraction; and NYHA, New York Heart Association.

5 586 Circ Cardiovasc Qual Outcomes September 2013 Table 3. Subset Analysis of Right Ventricular Defibrillation Lead Characteristics and Association With Cardiac Perforation From ICD Implantation Right Ventricular Lead Characteristic Cardiac Perforation (n=144) No Cardiac Perforation (n=99 328) P Value Adjusted Odds Ratio for Cardiac Perforation (95% CI) P Value Largest diameter of lead, 8.1± ± ( ) French size (per 1F size unit decrease) Type of lead fixation Passive 7.6% 5.4% Reference Active 92.4% 94.6% 0.95 ( ) Categorical variables are reported as proportions (%); continuous variables are reported as mean±sd. The adjusted odds ratio reported for largest diameter of lead is per 1F unit size decrease. CI indicates confidence interval; and ICD, implantable-cardioverter defibrillator. transient ischemic attack or stroke, peripheral embolus, conduction block, and valve injury, were also significantly more common in those with cardiac perforation (Table 4). Those with cardiac perforation had a significantly greater odds of any other associated major procedural complication in both unadjusted and multivariable adjusted analyses (Figure 1). We did not find evidence that the association between cardiac perforation and any associated major procedural complication was significantly different for any prespecified subgroup of patients (P>0.05 for all interaction terms; Figure 2A). Association of Cardiac Perforation With Length of Hospital Stay The median length of hospital stay for the entire cohort was 1 day (interquartile range, ), and the total length of hospital stay was 4 days in 90% of patients. The median length of stay was significantly longer in ICD recipients with cardiac perforation than those without (Table 4). The largest upper bound of the interquartile range observed in those with cardiac perforation approached 7 days compared with 2 days in those without cardiac perforation. In both unadjusted and multivariable adjusted analyses, ICD recipients with cardiac perforation had a greater odds of hospital stay >3 days compared with those without cardiac perforation (Figure 1). Association of Cardiac Perforation With In-Hospital Death The crude risk of in-hospital death was significantly higher in those with cardiac perforation (5.6%) compared with those without cardiac perforation (0.4%; Table 4). ICD recipients with cardiac perforation experienced a greater odds of inhospital death in unadjusted and multivariable adjusted analyses (Figure 1). We did not find evidence that the association between cardiac perforation and in-hospital death was significantly different for any prespecified subgroup of patients (P>0.05 for all interaction terms; Figure 2B). Table 4. Adverse Events and Specific Complications Associated With Cardiac Perforation From ICD Implantation Adverse Event Cardiac Perforation (n=625) No Cardiac Perforation (n= ) P Value Any associated major complication* 7.4% 0.3% < Length of stay (implant to discharge), d 4.0 ( ) 1.0 ( ) < In-hospital death 5.6% 0.4% < Specific associated complication Cardiac arrest 6.88% 0.28% < Myocardial infarction 0.32% 0.03% < Infection requiring antibiotics 0.32% 0.04% Pneumothorax 2.40% 0.41% < Hematoma 1.76% 0.65% Hemothorax 1.60% 0.06% < Transient ischemic attack or stroke 0.96% 0.06% < Peripheral embolus 0.32% 0.02% < Conduction block 0.32% 0.03% Valve injury 0.16% 0.00% < Drug reaction 0.00% 0.08% Peripheral nerve injury 0.00% 0.00% Categorical variables are reported as proportions (%); continuous variables are reported as median (interquartile range). ICD indicates implantable-cardioverter defibrillator. *Any associated major complication is defined as cardiac arrest, myocardial infarction, or infection requiring antibiotics.

6 Hsu et al Cardiac Perforation and ICD Adverse Events 587 Figure 1. Unadjusted (white boxes) and multivariable adjusted (black boxes) odds ratios (ORs) of any associated major complication, length of hospital stay >3 days, and in-hospital mortality among implantable-cardioverter defibrillator (ICD) recipients with cardiac perforation (defined as cardiac perforation or pericardial tamponade). The reference group for each analysis is ICD recipients without cardiac perforation. The horizontal error bars denote 95% confidence intervals (CIs). *Adjusted for patient demographics (age, sex, race, insurance payer), comorbidities (congestive heart failure, New York Heart Association class, syncope, ventricular tachycardia, cardiac arrest, atrial fibrillation, nonischemic cardiomyopathy, ischemic heart disease, previous myocardial infarction, previous coronary artery bypass graft surgery, previous percutaneous coronary intervention, cerebrovascular disease, chronic lung disease, diabetes mellitus, hypertension, end-stage renal disease), diagnostic information (indication for implantation, device type, left ventricular ejection fraction, QRS duration, blood urea nitrogen level), implanter characteristics (implanter specialty training, implanter volume), and hospital characteristics (profit type, region, size, teaching status). Discussion In a large, national sample of first-time ICD recipients, we identified several patient and implanter characteristics that predicted cardiac perforation risk. Patients who experienced cardiac perforation had a substantially increased risk of adverse events both before and after adjustment for potential confounders, including a >26-fold increased odds of any other associated major complication, 16-fold increased odds of prolonged hospital stay >3 days, and 15-fold increased odds of in-hospital death. Previous studies examining the incidence and predictors of cardiac perforation from lead implantation were small, studied older leads and included pacemaker lead implantation. 4,5,10 In a study by Mahapatra et al, 11 active fixation leads, steroid drug use, and temporary pacing wire use were associated with postimplant pericardial effusion in patients undergoing permanent pacemaker lead insertion. A large-scale study evaluating several potential predictors of cardiac perforation in ICD recipients has not been reported. Although cardiac perforation is recognized as a serious complication from ICD lead implantation, this is the first study to quantify the magnitude of the association between cardiac perforation and in-hospital adverse events. We found the prevalence of cardiac perforation in modern-day ICD implants to be relatively infrequent at 0.14%. However, with >600 cardiac perforation events, this is the largest study of its kind, enabling comprehensive analyses of predictors and outcomes of this important complication. In multivariable adjusted analysis, several patient and implanter characteristics were associated with cardiac perforation risk. Although our study was not equipped to identify the underlying mechanisms of these associations, several plausible explanations exist. Both older age and female sex have anecdotally been associated with an increased perforation risk. 12 In both populations, a thinner myocardial wall more susceptible to puncture may be responsible for an increased perforation risk. The presence of left bundle branch block was also associated with an increased risk of cardiac perforation and may be related to force on the lead tip from early, dyssynchronous right ventricular wall activation with subsequent late left ventricular activation. Worsened New York Heart Association heart failure class was independently associated with an increased cardiac perforation risk that may be related to overall frailty that affords susceptibility to complications. Apparently incongruous with these results was our finding that an increased LVEF was associated with a higher risk of perforation. Although speculative, it is possible that a smaller right ventricular volume or some as yet unknown factor associated with a normal LVEF (such as a more fragile right ventricle in arrhythmogenic right ventricular cardiomyopathy) may be responsible for an increased perforation risk. Alternatively, a higher LVEF may result in more forceful myocardial contractions against a lead tip, predisposing to cardiac perforation. Consistent with previous studies, devices with more leads were associated with more risk of cardiac perforation. 13,14 Of interest, although the unadjusted analyses suggested an incrementally increased risk of perforation with each additional lead (more for cardiac resynchronization therapy with defibrillators than dual-chamber ICDs), the multivariate analysis suggests that perforation risk increases primarily with non single-chamber devices as a group. Several characteristics were protective against cardiac perforation. Atrial fibrillation/atrial flutter was associated with a decreased cardiac perforation risk, which may be because of atrial enlargement from the arrhythmia, more atrial fibrosis, or less forceful contraction of the fibrillating atrium against the lead tip. A history of diabetes mellitus was also protective against cardiac perforation and may afford protection from myocardial wall puncture because of myocardial fibrosis seen in this patient population. 15 Previous coronary artery bypass graft surgery dramatically reduced the risk of cardiac perforation, likely from surgically induced pericardial fibrosis decreasing clinically meaningful or detectable cardiac perforation. Although teaching hospital status and implanting physician training were not associated with perforation, greater

7 588 Circ Cardiovasc Qual Outcomes September 2013 Figure 2. The adjusted odds ratios for any associated major complication (A) and in-hospital death (B) are shown for various prespecified subgroups among those with cardiac perforation compared with those without cardiac perforation. All odds ratios are adjusted for the same covariates listed in Figure 1. The dashed vertical line in each panel represents the adjusted odds ratio of the adverse outcome for the entire cohort, and the horizontal error bars denote 95% confidence intervals. All interaction P values exceeded CRT-D indicates cardiac resynchronization therapy with defibrillator; ICD, implantable cardioverter-defibrillator; and LVEF, left ventricular ejection fraction. implanter experience was protective against cardiac perforation risk, the latter being consistent with previous studies evaluating complications during ICD implantation. 16 Because the ICD Registry began collecting lead characteristics only after April 1, 2010, we were limited in our investigation of the association between right ventricular defibrillation lead characteristics and cardiac perforation. In multivariable analysis, implantation with a smaller diameter right ventricular lead was associated with an increased cardiac perforation risk. It is plausible that a smaller diameter lead tip results in more force per unit area of myocardium, predisposing to myocardial perforation. Interestingly, no differences by fixation type (active versus passive) were observed. These findings suggest that the design and composition of the right

8 Hsu et al Cardiac Perforation and ICD Adverse Events 589 ventricular lead may be an independent determinant of cardiac perforation risk, apparently driven primarily by the lead diameter. The absolute difference in major associated complications and mortality between ICD recipients with cardiac perforation versus those without perforation was profound (7.4% versus 0.3% and 5.6% versus 0.4%, respectively), suggesting that many acute problems associated with new ICD implantation may ultimately be related to cardiac perforation. Consistent with these findings, the median length of stay was markedly longer in those with perforation (4 days versus 1 day). The considerably greater odds of adverse events in those with cardiac perforation was consistent across all 3 outcomes studied, persisted despite adjustment for potential confounders, and was not different across prespecified subgroups. These increased risks have important ramifications relevant to patients and practicing physicians, particularly as such major complications and longer hospital stays likely adversely affect patient quality of life and translate into increased healthcare utilization and costs. If indeed cardiac perforation is at the core of the majority of acute complications that occur in new ICD implants and the lead characteristics influence that risk, these data may inform lead manufacturers to be even more vigilant in ensuring the safety of lead implant procedures. In addition, knowledge that cardiac perforation is associated with a considerably increased risk of adverse events may allow for more aggressive treatment in those who experience the complication (such as intensive care unit monitoring) to avert further morbidity and mortality. By distinguishing easily recognizable predictors of cardiac perforation identified in this study, implanters may be better able to counsel patients about their risks, better able to quantify that risk, and may alter their approach (such as placing fewer total leads) in those most vulnerable. Further research into the mechanisms of novel predictors found to be associated with cardiac perforation in our study is warranted, particularly on the association of left bundle branch block and increased LVEF with a higher cardiac perforation risk, as well as the association of atrial fibrillation and diabetes mellitus with a lower cardiac perforation risk. Specifically, now that these covariates have been identified, other device registries as well as prospective studies might consider special efforts to make sure these covariates are included on case report forms. Although some of these effects may be specific to ICDs, it is likely that the mechanisms may extend to pacemakers (and therefore studies of pacemakers) as well. Study Limitations Our study has several limitations. First, our analyses were limited to adverse events during the index hospitalization and may not be generalizable to patients who experience late cardiac perforations after leaving the hospital. However, previous studies have shown that the majority of complications from ICD implantation are recognized before discharge. 17 Second, cardiac perforation occurred in a small proportion (0.14%) of all ICD recipients but included >600 patients, providing ample power to perform the analyses described. In addition, a lack of power should not result in spurious false-positive associations. Third, we do not have intraprocedural or postprocedural information about the type/implant location of the lead that caused the cardiac perforation. That dual-chamber ICD and cardiac resynchronization therapy with defibrillator systems were associated with a higher perforation risk than single-chamber ICD systems suggests that atrial and coronary sinus leads were responsible for at least part of the cardiac perforation risk. Furthermore, the nature of the ICD Registry, with limited data fields, did not allow further investigation into the means of cardiac perforation detection or the clinical presentation of the complication. Fourth, also because of the limitations inherent to the ICD Registry, we were not able to assess the impact of several covariates previously associated with cardiac perforation during pacemaker implantation, including echocardiagraphic parameters, specific procedural characteristics such as fluoroscopy times, and use of a temporary pacemaker. 11 Fifth, although body mass index may theoretically be important, this covariate was not available in the majority of the cohort, and our previous analysis focusing on body mass index failed to detect a statistically significant association with cardiac perforation. 18 Sixth, we cannot rule out reporting bias as an explanation for the association found between cardiac perforation and other associated complications; for example, hospitals more likely to report cardiac perforation as a complication may also be more apt to report other complications. Finally, as with any observational study, we cannot exclude the possibility that residual confounding explains our results. However, our extensive multivariable adjustment did not meaningfully change any of our results, and it seems unlikely that an unmeasured confounder could explain the magnitude of the adverse events associated with cardiac perforation. Conclusions In a large, national registry of first-time ICD recipients, specific patient and implanter characteristics predicted an increased cardiac perforation risk. Cardiac perforation from implantation was associated with a substantially increased risk of other major in-hospital complications, prolonged hospitalization, and death. These findings help to inform implanting physicians about readily identifiable baseline characteristics that may predispose to cardiac perforation and therefore warrant cautionary measures. These results also highlight and quantify the magnitude of specific adverse events associated with cardiac perforation. Sources of Funding This research was supported by the American College of Cardiology Foundation s National Cardiovascular Data Registry. The views expressed in this article represent those of the authors and do not necessarily represent the official views of the NCDR or its associated professional societies identified at The ICD Registry is an initiative of the American College of Cardiology Foundation and the Heart Rhythm Society. Disclosures Dr Dewland has received educational travel grants from Medtronic and Boston Scientific. Dr Curtis has modest ownership in Medtronic and receives salary support from the American College of Cardiology NCDR. Dr Marcus receives research support from Medtronic. The other authors report no conflicts.

9 590 Circ Cardiovasc Qual Outcomes September 2013 References 1. García-Bolao I, Teijeira R, Díaz-Dorronsoro I. Late fatal right ventricular perforation as complication of permanent pacing leads. Pacing Clin Electrophysiol. 2001;24: Carlson MD, Freedman RA, Levine PA. Lead perforation: incidence in registries. Pacing Clin Electrophysiol. 2008;31: Hirschl DA, Jain VR, Spindola-Franco H, Gross JN, Haramati LB. Prevalence and characterization of asymptomatic pacemaker and ICD lead perforation on CT. Pacing Clin Electrophysiol. 2007;30: Trigano AJ, Taramasco V, Paganelli F, Gerard R, Lévy S. Incidence of perforation and other mechanical complications during dual active fixation. Pacing Clin Electrophysiol. 1996;19(11 pt 2): Turakhia M, Prasad M, Olgin J, Badhwar N, Tseng ZH, Lee R, Marcus GM, Lee BK. Rates and severity of perforation from implantable cardioverter-defibrillator leads: a 4-year study. J Interv Card Electrophysiol. 2009;24: Hammill SC, Kremers MS, Stevenson LW, Kadish AH, Heidenreich PA, Lindsay BD, Mirro MJ, Radford MJ, Wang Y, Curtis JP, Lang CM, Harder JC, Brindis RG. Review of the Registry s second year, data collected, and plans to add lead and pediatric ICD procedures. Heart Rhythm. 2008;5: Messenger JC, Ho KK, Young CH, Slattery LE, Draoui JC, Curtis JP, Dehmer GJ, Grover FL, Mirro MJ, Reynolds MR, Rokos IC, Spertus JA, Wang TY, Winston SA, Rumsfeld JS, Masoudi FA; NCDR Science and Quality Oversight Committee Data Quality Workgroup. The National Cardiovascular Data Registry (NCDR) Data Quality Brief: the NCDR Data Quality Program in J Am Coll Cardiol. 2012;60: Austin PC, Tu JV, Alter DA. Comparing hierarchical modeling with traditional logistic regression analysis among patients hospitalized with acute myocardial infarction: should we be analyzing cardiovascular outcomes data differently? Am Heart J. 2003;145: Curtis AB, Ellenbogen KA, Hammill SC, Hayes DL, Reynolds DW, Wilber DJ, Cain ME. Clinical competency statement: training pathways for implantation of cardioverter defibrillators and cardiac resynchronization devices. Heart Rhythm. 2004;1: Hill PE. Complications of permanent transvenous cardiac pacing: a 14- year review of all transvenous pacemakers inserted at one community hospital. Pacing Clin Electrophysiol. 1987;10(3 pt 1): Mahapatra S, Bybee KA, Bunch TJ, Espinosa RE, Sinak LJ, McGoon MD, Hayes DL. Incidence and predictors of cardiac perforation after permanent pacemaker placement. Heart Rhythm. 2005;2: Josephson ME. Clinical Cardiac Electrophysiology: Techniques and Interpretations. 4th ed. Philadelphia, PA: Wolters Kluwer Health/ Lippincott Williams & Wilkins; Bhatta L, Luck JC, Wolbrette DL, Naccarelli GV. Complications of biventricular pacing. Curr Opin Cardiol. 2004;19: Dewland TA, Pellegrini CN, Wang Y, Marcus GM, Keung E, Varosy PD. Dual-chamber implantable cardioverter-defibrillator selection is associated with increased complication rates and mortality among patients enrolled in the NCDR implantable cardioverter-defibrillator registry. J Am Coll Cardiol. 2011;58: Asbun J, Villarreal FJ. The pathogenesis of myocardial fibrosis in the setting of diabetic cardiomyopathy. J Am Coll Cardiol. 2006;47: Freeman JV, Wang Y, Curtis JP, Heidenreich PA, Hlatky MA. Physician procedure volume and complications of cardioverter-defibrillator implantation. Circulation. 2012;125: Al-Khatib SM, Greiner MA, Peterson ED, Hernandez AF, Schulman KA, Curtis LH. Patient and implanting physician factors associated with mortality and complications after implantable cardioverter-defibrillator implantation, Circ Arrhythm Electrophysiol. 2008;1: Hsu JC, Varosy PD, Bao H, Wang Y, Curtis JP, Marcus GM. Low body mass index but not obesity is associated with in-hospital adverse events and mortality among implantable cardioverter-defibrillator recipients: insights from the National Cardiovascular Data Registry. J Am Heart Assoc. 2012;1:e

10 Cardiac Perforation From Implantable Cardioverter-Defibrillator Lead Placement: Insights From the National Cardiovascular Data Registry Jonathan C. Hsu, Paul D. Varosy, Haikun Bao, Thomas A. Dewland, Jeptha P. Curtis and Gregory M. Marcus Circ Cardiovasc Qual Outcomes. 2013;6: ; originally published online September 3, 2013; doi: /CIRCOUTCOMES Circulation: Cardiovascular Quality and Outcomes is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX Copyright 2013 American Heart Association, Inc. All rights reserved. Print ISSN: Online ISSN: The online version of this article, along with updated information and services, is located on the World Wide Web at: Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in Circulation: Cardiovascular Quality and Outcomes can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for which permission is being requested is located, click Request Permissions in the middle column of the Web page under Services. Further information about this process is available in the Permissions and Rights Question and Answer document. Reprints: Information about reprints can be found online at: Subscriptions: Information about subscribing to Circulation: Cardiovascular Quality and Outcomes is online at: