Gender Differences in Procedure-Related Adverse Events in Patients Receiving Implantable Cardioverter-Defibrillator Therapy
Background— Women are at higher risk than men for adverse events with certain invasive cardiac procedures. Our objective was to compare rates of in-hospital adverse events in men and women receiving implantable cardioverter- defibrillator (ICD) therapy in community practice.
Methods and Results— Using the National Cardiovascular Data Registry ICD Registry, we identified patients undergoing first-time ICD implantation between January 2006 and December 2007. Outcomes included in-hospital adverse events after ICD implantation. Multivariable analysis assessed the association between gender and in-hospital adverse events, with adjustment for demographic, clinical, procedural, physician, and hospital characteristics. Of 161 470 patients, 73% were male, and 27% were female. Women were more likely to have a history of heart failure (81% versus 77%, P<0.01), worse New York Heart Association functional status (57% versus 50% in class III and IV, P<0.01), and nonischemic cardiomyopathy (44% versus 27%, P<0.01) and were more likely to receive biventricular ICDs (39% versus 34%, P<0.01). In unadjusted analyses, women were more likely to experience any adverse event (4.4% versus 3.3%, P<0.001) and major adverse events (2.0% versus 1.1%, P<0.001). In multivariable models, women had a significantly higher risk of any adverse event (OR 1.32, 95% CI 1.24 to 1.39) and major adverse events (OR 1.71, 95% CI 1.57 to 1.86).
Conclusions— Women are more likely than men to have in-hospital adverse events related to ICD implantation. Efforts are needed to understand the reasons for higher ICD implantation–related adverse event rates in women and to develop strategies to reduce the risk of these events.
Received May 20, 2008; accepted December 8, 2008.
Randomized clinical trials have demonstrated the efficacy of implantable cardioverter-defibrillators (ICDs) in reducing mortality in patients at risk for life-threatening cardiac arrhythmias1–5; however, ICD implantation may be associated with adverse events that may attenuate the benefits derived from therapy. With the dramatic increase in the number of ICD implantations,6 it is important to understand complications related to implantation. Data regarding implantation-related adverse events from rhythm-management devices, both pacemakers7–11 and ICDs,12–15 are limited because they are derived from clinical trials, single-center reports, or studies using administrative data. Clinical trials of procedures generally include experienced centers and highly selected patient populations.16–18 Single-center reports also may not reflect the patient populations, practice patterns, or outcomes in wider community practice. Finally, administrative data are limited with respect to important potential confounders and do not permit standardized definitions of adverse events.
Editorial p 1069
Clinical Perspective p 1084
Studies of some invasive cardiovascular procedures indicate that women are at particularly high risk for procedural complications compared with men. For example, women have higher rates of bleeding, stroke, and in-hospital mortality after revascularization procedures.19 Proposed explanations are that women are older and have a higher prevalence of comorbidities and risk factors. However, remarkably little is known about gender differences in complications after ICD implantation. Although 1 study has reported that women may be more likely to experience ICD implantation–related complications,14 that study was confined to the Medicare population and was limited to administrative data to characterize the study population and to identify complications.
Accordingly, we hypothesized that women would be more likely than men to experience an implantation-related adverse event after ICD implantation. To address this question, we used data from a national registry that contained detailed clinical data and used standardized definitions of adverse events for patients undergoing ICD implantation in community practice. We studied ICD implantation–related adverse event rates among men and women and evaluated whether complication rates varied according to important patient and device factors.
Analyses used data from the National Cardiovascular Data Registry ICD Registry. In January 2005, in conjunction with the expansion of coverage for ICDs used for primary prevention, the Center for Medicare and Medicaid Services mandated that hospitals participate in a national ICD registry. Because many questions about the safety and effectiveness of primary prevention ICDs remain, the Center for Medicare and Medicaid Services used the expansion of coverage to support the development of an ICD registry to gather additional practical evidence to help clinicians and patients make informed decisions.20 Although the Center for Medicare and Medicaid Services mandate requires only that Medicare patients receiving primary prevention ICDs be entered into the database, secondary prevention implantations are also represented, because the majority of hospitals voluntarily enter information on all ICD implantations regardless of indication or insurance status.21 Clinical, demographic, and procedural information is collected in addition to information about adverse events until the time of discharge by use of standardized data elements and definitions. Data are submitted by participating hospitals using American College of Cardiology–certified software, and data quality is examined with the Data Quality Reporting process.21 Implanting centers must enter complete and accurate data to receive Medicare reimbursement.
Between January 2006 and December 2007, 1224 hospitals submitted data to the National Cardiovascular Data Registry ICD Registry that passed the Data Quality Reporting thresholds. Data included 204 700 unique patients undergoing device implantation. We identified a final cohort of 161 470 patients undergoing first-time ICD implantation, excluding patients who had undergone a prior ICD implantation (n=43 230).
The occurrence of any adverse event during the hospitalization for ICD implantation was the primary end point, and the occurrence of major adverse events was the secondary end point. Events that occur during or after the implantation procedure until the time of hospital discharge are reported by implanting centers using standard definitions as outlined in the Appendix (Data Supplement). Major adverse events included cardiac arrest, cardiac perforation, cardiac valve injury, coronary venous dissection, hemothorax, pneumothorax, deep phlebitis, transient ischemic attack, stroke, myocardial infarction, pericardial tamponade, and arteriovenous fistula. Any adverse event included both major adverse events as described above and drug reaction, conduction block, hematoma, lead dislodgment, peripheral embolus, superficial phlebitis, peripheral nerve injury, and infection related to the device. In secondary analyses, hospital length of stay (in days) was assessed.
Predictor Variable and Confounding Variables
The primary predictor variable of interest was patient gender. The other confounding variables considered were patient demographics (age, race, insurance payer); reason for hospitalization; patient comorbidities and risk factors (history of congestive heart failure, atrial fibrillation or flutter, cardiac transplantation, cause of cardiomyopathy, myocardial infarction, coronary artery bypass grafting, percutaneous coronary intervention, valvular surgery, cerebrovascular disease, pacemaker insertion, chronic lung disease, diabetes mellitus, hypertension, renal failure [hemodialysis], and admission New York Heart Association classification); diagnostic information (ejection fraction, whether an electrophysiology study was performed, electrophysiology study results, QRS duration, PR interval, presence of atrial or ventricular conduction delays, creatinine, serum urea nitrogen, sodium level, brain natriuretic peptide level, and systolic blood pressure); procedural information (ICD type [single-chamber, dual-chamber, or biventricular ICD] and lead implantation method); and physician and hospital characteristics (whether the implanting physician is certified in electrophysiology per self-report, for-profit or not-for-profit hospital, rural or nonrural hospital, hospital teaching status, number of beds, and annual number of ICD implantations at each hospital).
The rate of missing data was extremely low for all variables (<0.5%, with the exception of left ventricular ejection fraction, which was missing in 1.5% of cases). To avoid casewise deletion of those cases with missing data points, the missing values were imputed. For categorical variables, the missing variables were imputed as the most common value among those with the data present. For example, for the categorical variable of New York Heart Association class, which was missing in 0.15% of cases, missing values were imputed as class I. For continuous variables, the missing values were imputed as the median among those with the data present. Dummy variables were constructed to indicate when the variable was missing. In the multivariable models, both the imputed values and the dummy variable were included.
Baseline demographic and clinical factors were compared between men and women with the χ2 test for categorical variables and t tests for continuous variables. The rates of in-hospital adverse events were compared between men and women with the χ2 test.
The independent association of gender with device implantation–related adverse events was assessed with hierarchical logistic regression with PROC GLIMMIX, with accounting for clustering of patients within hospitals. The demographic, clinical, procedural, and physician characteristics in Table 1⇓ were all included in the models. The single physician-level characteristic was treated as a patient-level factor. No variable selection procedures were performed.
The relationship between gender and adverse events was also explored in prespecified subgroups. Because older people are more prone to adverse events from invasive procedures, and because biventricular ICD implantation is a more extensive and time-consuming procedure, analyses were stratified by age and device type. Additionally, patients were stratified by payer and indication (primary versus secondary prevention). Multivariable hierarchical logistic regression models were constructed by use of the same approach as the full risk models, stratified by each of the prespecified subgroups. The statistical significance of differences among strata was tested with 2-way interaction terms in full models.
In the secondary multivariable analyses of length of stay, the outcome was log-transformed to accommodate the skewed distribution of this variable. After regression, the adjusted length of stay was exp-transformed back to its natural unit. All analyses were performed with the SAS statistical package version 9.1 (SAS Institute, Cary, NC).
The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
Of the 161 470 patients, 73% were men, and 27% were women. Women were more likely to have a history of heart failure, worse New York Heart Association functional status, and nonischemic cardiomyopathy. Women were also more likely to receive biventricular ICDs (Table 1⇑). Given the large number of patients in the present study cohort, differences in baseline characteristics may be statistically significant but have little clinical significance.
The overall rate of any adverse event in the entire cohort was 3.6%, with a higher rate in women than in men (4.4% versus 3.3%, P<0.001; Figure 1). The rate of major adverse events in women was nearly double that of men (2.0% versus 1.1%, P<0.001), with an overall rate of 1.3% (Figure 1). When adverse events were considered individually, hematoma and lead dislodgment were the most common. Women were more likely to have a drug reaction, cardiac perforation, conduction block, coronary venous dissection, lead dislodgment, hemothorax, pneumothorax, deep phlebitis, and pericardial tamponade (Table 2). The rates of cardiac arrest, valve injury, hematoma, peripheral nerve injury, peripheral embolus, superficial phlebitis, transient ischemic attack or stroke, myocardial infarction, arteriovenous fistula, and infection did not differ significantly between men and women (Table 2). In-hospital mortality, which was not considered among adverse events, was not different between men and women (Table 2).
Hospital length of stay was longer among all patients with a reported complication (7.11 days) than among those without a complication (3.83 days, P<0.001). The difference in length of stay was significant across strata according to the reason for admission, including those who were admitted specifically for the ICD implantation (n=99 250). In this group, length of stay was significantly higher in patients with a complication (3.91 days) than in those without a complication (1.50 days, P<0.0001). In adjusted analyses, length of stay was significantly higher in patients with complications, by 11.3 days (P<0.0001).
After adjustment for demographic, admission, clinical, and physician factors, women had significantly higher odds of any adverse event (OR 1.32, 95% CI 1.24 to 1.39) and major adverse events (OR 1.71, 95% CI 1.57 to 1.86) than men. Additionally, in stratified analyses, the gender difference in rate of any adverse event was greater among patients ≥65 years (OR 1.38, 95% CI 1.29 to 1.48) than in those <65 years (OR 1.18, 95% CI 1.07 to 1.31; P for age-gender interaction=0.023). However, the association between gender and any adverse event did not vary by device type or indication for implantation (Figure 2). Similarly, the association between gender and major adverse events did not vary within any of the predefined subgroups.
Among patients receiving ICDs, women had a higher rate of any adverse event and almost double the rate of major adverse events compared with men. After accounting for demographic, clinical, and procedural differences between men and women, women had 32% higher odds of experiencing an adverse event and 71% higher odds of experiencing a major adverse event than men.
The present study expands the current knowledge of ICD implantation–related adverse events by evaluating a broad range of adverse events in a large cohort of patients receiving ICDs in real-world clinical practice. The observed adverse event rate of 3.6% is consistent with clinical trial data in which adverse event rates range from 1.3% to 5% for ICD implantation.1,4,22,23 Of note, clinical trials have not consistently defined or reported adverse events, which renders comparisons among these studies difficult. The present findings are also consistent with the short-term adverse event rate of 3.2% found among a nationwide sample of >900 hospitals for device implantations from 1997 to 2004.6 The adverse event rate we observed is lower than the 10.8% rate previously reported among Medicare beneficiaries receiving ICDs; however, that study was limited by the use of administrative data to identify adverse events, inclusion of both initial and replacement implantations, evaluation of Medicare beneficiaries only, lack of contemporary data, and low hospital implantation volume.14 Finally, the present findings are consistent with the perioperative complication rate of 3.7% observed in the Ontario ICD Database.24 This database is comparable to the US National Cardiovascular Data Registry ICD Registry in that it is a prospective registry of patients undergoing ICD implantation with mandated participation by the Ministry of Health and Long-Term Care in Ontario. Trained data collectors enter data into the Ontario ICD Database, and quality controls include random site audits with reabstraction of randomly selected charts by central study personnel.
The most important contribution of the present study is the comparison of gender-specific ICD implantation–related adverse event rates. The association between female gender and adverse outcomes found in the present study is consistent with previous studies that examined this relationship for percutaneous coronary intervention and coronary artery bypass grafting.19 Postulated reasons for the higher complication rates observed with these revascularization procedures are that women with cardiovascular disease are older and have a higher prevalence of comorbidities and risk factors such as diabetes mellitus, peripheral vascular disease, and hypertension.19 In the large cohort of patients we studied, men and women did not differ with respect to age, and although women did have a higher prevalence of comorbidities and risk factors, after adjustment for these factors, the association between sex and adverse events persisted. Thus, the factors thought to contribute to higher complication rates in women after other procedures do not appear to entirely explain the higher adverse event rate observed with ICD implantation.
Smaller body size is another proposed explanation for the differences in outcomes seen with revascularization procedures and may explain the higher adverse event rate observed in women with ICD device implantation. Indeed, many of the adverse events that were more common in women were mechanical (cardiac perforation, coronary venous dissection, lead dislodgment, hemothorax, pneumothorax, and pericardial tamponade). This suggests that women may have anatomy that poses greater technical challenges, such as a thinner right ventricular wall and smaller blood vessel diameter, which may predispose them to adverse events. Because neither body mass index or body surface area was available in the National Cardiovascular Data Registry data set, the extent to which variations in body size attenuate gender differences in complication rates could not be ascertained. The present findings are consistent with the few existing studies of pacemaker implantation–associated adverse events that found mechanical complications including lead dislodgment and pneumothorax to be most common.7–11 However, gender has not been identified as a risk factor for pacemaker-related adverse events. Studies of adverse events associated with pacemaker implantation are limited because they were single-center reports, used small cohorts, or were performed in clinical trial cohorts.
The reasons for higher adverse event rates in women with ICD implantation should be investigated and, where possible, eliminated. Improvements have occurred over time in the outcomes of women undergoing revascularization with percutaneous coronary intervention and coronary artery bypass grafting.25–27 Improvements in outcomes may be due to improved patient selection, advances in technology and procedural techniques, greater operator experience, and refinement of perioperative management and adjunctive medical therapy. Future studies are needed to investigate the reasons for differences in ICD implantation–related adverse event rates between men and women so that measures can be taken to reduce the increased risk in women.
The findings of the present study should not be interpreted to suggest that practitioners should use different thresholds to implant ICDs in women than in men. The decision to pursue this therapy should be based primarily on estimates of the mortality benefits of therapy and patient preferences. However, it is plausible that nonfatal adverse events may influence both cost and patient-centered outcomes, such as quality of life after device implantation. Thus, studies to understand the impact of ICD complications on these other outcomes are warranted in parallel with efforts to understand the causes and reduce the incidence of complications.
The results of the present study highlight the important role of registries in the evaluation of adverse events outside the context of clinical trials. The assessment of harm is often inadequate in randomized trials for several reasons. First, randomized trials are generally underpowered to evaluate adverse events. Second, selection criteria in clinical trials may result in populations with different adverse event profiles.17,18,28 Third, the sites performing procedures in clinical trials are often limited in number and have clinicians with a greater level of training and experience performing the procedure.16,17 Thus, adverse event rates from clinical trials may not be generalizable. Similarly, single-center reports have limited generalizability. For this reason, the Center for Medicare and Medicaid Services established the National Cardiovascular Data Registry ICD Registry to determine whether outcomes of ICD implantation differ in general practice compared with clinical trials.
Certain factors should be considered in the interpretation of the present study. First, adverse event rates are self-reported by hospitals and may be underreported, which could result in spuriously low adverse event rates; however, because it is unlikely that patterns of reporting would vary by patient gender, any underreporting would result in attenuation of the observed relative risk of complications between sexes.29 Second, the present analysis was limited to short-term in-hospital adverse events. Adverse events such as infections, lead malfunctions, or shocks, which do not manifest until after hospital discharge, were not captured. Additionally, the adverse events captured were limited to those available as data elements in the data collection tool. Events such as incessant ventricular fibrillation at implantation, worsening heart failure, and failed biventricular ICD implantation were not included as adverse events, as they were in some clinical trials1,30; however, a broad range of implantation-related adverse events important to patients were included in the present analysis. Third, some adverse events are defined by the registry by their most severe consequences (eg, pneumothorax was defined as air in the thorax sufficient to require a chest tube), and therefore, a number of minor complications may not have been captured; however, standard definitions were used for all adverse events independent of patient gender. Finally, we did not have any data on factors such as height, weight, or body mass index and thus were unable to adjust for the potential confounder of body size. Similarly, we did not have information on the lead types used, which may confound the observed difference in the rate of lead dislodgment between men and women.
In summary, in-hospital adverse events occurred in almost 1 in 25 patients undergoing first-time ICD implantation. Gender was a strong and independent factor associated with adverse events in a large cohort of patients receiving an ICD in clinical practice. Although these results should not preclude women from receiving ICDs, a better perspective of the consequences of adverse events on outcomes such as cost and quality of life would be useful to inform decision making concerning ICD implantation. Furthermore, these data support efforts to understand the reasons for higher ICD implantation–related adverse event rates in women and to develop strategies to reduce this risk.
Sources of Funding
This research was supported by the National Cardiovascular Data Registry. Dr Peterson is supported by the American Heart Association Pharmaceutical Roundtable Award #0670017N.
Dr Heidenreich serves as a consultant for Boston Scientific. Dr Masoudi has served on advisory boards for Takeda NA, Amgen, and United Healthcare; he has received grant funding from Amgen; he has contracts with the American College of Cardiology and the Oklahoma Foundation for Medical Quality; and he serves as an Associate Editor for Journal Watch Cardiology, Massachusetts Medical Society, and for Circulation: Quality of Care and Outcomes. The other authors report no conflicts.
Bardy GH, Lee KL, Mark DB, Poole JE, Packer DL, Boineau R, Domanski M, Troutman C, Anderson J, Johnson G, McNulty SE, Clapp-Channing N, Davidson-Ray LD, Fraulo ES, Fishbein DP, Luceri RM, Ip JH; the Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT) Investigators. Amiodarone or an implantable cardioverter-defibrillator for congestive heart failure. N Engl J Med. 2005; 352: 225–237.
Moss AJ, Hall WJ, Cannom DS, Daubert JP, Higgins SL, Klein H, Levine JH, Saksena S, Waldo AL, Wilber D, Brown MW, Heo M; Multicenter Automatic Defibrillator Implantation Trial Investigators. Improved survival with an implanted defibrillator in patients with coronary disease at high risk for ventricular arrhythmia. N Engl J Med. 1996; 335: 1933–1940.
Moss AJ, Zareba W, Hall WJ, Klein H, Wilber DJ, Cannom DS, Daubert JP, Higgins SL, Brown MW, Andrews ML; Multicenter Automatic Defibrillator Implantation Trial II Investigators. Prophylactic implantation of a defibrillator in patients with myocardial infarction and reduced ejection fraction. N Engl J Med. 2002; 346: 877–883.
Link MS, Estes NA III, Griffin JJ, Wang PJ, Maloney JD, Kirchhoffer JB, Mitchell GF, Orav J, Goldman L, Lamas GA; Pacemaker Selection in the Elderly (PASE) Investigators. Complications of dual chamber pacemaker implantation in the elderly. J Interv Card Electrophysiol. 1998; 2: 175–179.
Eberhardt F, Bode F, Bonnemeier H, Boguschewski F, Schlei M, Peters W, Wiegand UK. Long term complications in single and dual chamber pacing are influenced by surgical experience and patient morbidity. Heart. 2005; 91: 500–506.
van Eck JW, van Hemel NM, Zuithof P, van Asseldonk JP, Voskuil TL, Grobbee DE, Moons KG. Incidence and predictors of in-hospital events after first implantation of pacemakers. Europace. 2007; 9: 884–889.
Gras D, Bocker D, Lunati M, Wellens HJ, Calvert M, Freemantle N, Gervais R, Kappenberger L, Tavazzi L, Erdmann E, Cleland JG, Daubert JC; CARE-HF Study Steering Committee and Investigators. Implantation of cardiac resynchronization therapy systems in the CARE-HF trial: procedural success rate and safety. Europace. 2007; 9: 516–522.
Jacobs AK. Coronary revascularization in women in 2003: sex revisited. Circulation. 2003; 107: 375–377.
Kadish A, Dyer A, Daubert JP, Quigg R, Estes NAM, Anderson KP, Calkins H, Hoch D, Goldberger J, Shalaby A, Sanders WE, Schaechter A, Levine JH; the Defibrillators in Non-Ischemic Cardiomyopathy Treatment Evaluation (DEFINITE) Investigators. Prophylactic defibrillator implantation in patients with nonischemic dilated cardiomyopathy. N Engl J Med. 2004; 350: 2151–2158.
Lee DS, Birnie D, Cameron D, Crystal E, Dorian P, Gula LJ, Healey JS, Janmohammed A, Khaykin Y, Krahn AD, LeFeuvre C, Simpson CS, Yee R, Hardy J, Slaughter PM, Chen Z, Alter DA, Laupacis A, Tu JV. Design and implementation of a population-based registry of implantable cardioverter defibrillators (ICDs) in Ontario. Heart Rhythm. 2008; 5: 1250–1256.
Abbott JD, Vlachos HA, Selzer F, Sharaf BL, Holper E, Glaser R, Jacobs AK, Williams DO; National Heart, Lung, and Blood Institute Dynamic Registry. Gender-based outcomes in percutaneous coronary intervention with drug-eluting stents (from the National Heart, Lung, and Blood Institute Dynamic Registry). Am J Cardiol. 2007; 99: 626–631.
Jacobs AK, Johnston JM, Haviland A, Brooks MM, Kelsey SF, Holmes DR Jr, Faxon DP, Williams DO, Detre KM. Improved outcomes for women undergoing contemporary percutaneous coronary intervention: a report from the National Heart, Lung, and Blood Institute Dynamic Registry. J Am Coll Cardiol. 2002; 39: 1608–1614.
Sharma PP, Greenlee RT, Anderson KP, Chyou PH, Osorio HJ, Smith PN, Hayes JH, Vidaillet H. Prevalence and mortality of patients with myocardial infarction and reduced left ventricular ejection fraction in a defined community: relation to the second multicenter automatic defibrillator implantation trial. J Interv Card Electrophysiol. 2007; 19: 157–164.
Kelsey J, Wittemore AS, Evans AS, Thompson WD. Methods in Observational Epidemiology. New York, NY: Oxford University Press; 1996.
Goldenberg I, Moss AJ, Hall WJ, McNitt S, Zareba W, Andrews ML, Cannom DS; for the Multicenter Automatic Defibrillator Implantation Trial (MADIT) II Investigators. Causes and consequences of heart failure after prophylactic implantation of a defibrillator in the Multicenter Automatic Defibrillator Implantation Trial II. Circulation. 2006; 113: 2810–2817.
Women are at higher risk than men for adverse events with some invasive cardiac procedures. Despite the dramatic increase in the number of implantations of implantable cardioverter defibrillators (ICDs), data regarding implantation-related adverse events are limited because they are derived from clinical trials, single-center reports, or studies using administrative data. Using the National Cardiovascular Data Registry ICD Registry, we compared rates of in-hospital adverse events in men and women receiving ICD therapy in community practice between January 2006 and December 2007. Of 161 470 patients, 27% were female. Women had a significantly higher risk of any adverse event (4.4% versus 3.3%, P<0.001; adjusted OR 1.32, 95% CI 1.24 to 1.39) and major adverse events (2.0% versus 1.1%, P<0.001; adjusted OR 1.71, 95% CI 1.57 to 1.86). Many of the adverse events that were more common in women were mechanical (cardiac perforation, coronary venous dissection, lead dislodgment, hemothorax, pneumothorax, and pericardial tamponade). Although these results should not preclude women from receiving ICDs, efforts are needed to understand the reasons for higher ICD implantation–related adverse event rates in women and to develop strategies to reduce the risk of these events.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.108.793463/DC1.