Adults with congenital heart disease are at risk of lethal ventricular arrhythmias and are candidates for implantable cardiac defibrillator (ICD) therapy, yet implant risks, long-term outcomes, and rates of appropriate and inappropriate ICD therapies are not well characterized. We reviewed clinical, implantation, and follow-up data on all transvenous ICDs in adults with congenital heart disease at the Mayo Clinic from 1991 through 2008. Seventy-three adults with congenital heart disease received 85 ICDs. Implantation diagnoses included tetralogy of Fallot (44%) and congenitally corrected transposition of the great arteries (17%). Implantation indication was occurrence of sustained ventricular arrhythmias (secondary prevention) in 36% and prophylactic (primary prevention) in the remainder. There were no major implant-related complications. During follow-up (2.2 ± 2.8 years, range 0 to 15) 11 patients died and 4 patients received heart or heart/lung transplants. An appropriate shock for a ventricular arrhythmia was observed in 19% of patients and an inappropriate shock was observed in 15% of patients. Likelihood of an appropriate shock was associated with increased subpulmonic ventricular pressure. In conclusion, implantation of transvenous ICDs in adults with congenital heart disease is associated with a low risk of implant complications. In this high-risk adult population the rate of inappropriate ICD shocks is low, whereas the likelihood of appropriate therapy for potentially lethal ventricular arrhythmias is high. These data suggest overall benefit of ICD therapy in adults with congenital heart disease.
It is estimated that approximately 800,000 adults living in the United States have congenital heart disease. Sudden cardiac death is the mechanism of death in 26% of these patients, which is 25 to 100 times higher than in an age-matched control population. Use of implantable cardiac defibrillators (ICDs) in this patient population has increased dramatically in the previous 15 years yet little is known about device implant-related complications and long-term outcomes including rates of appropriate or inappropriate ICD shocks.
Methods
All adult patients referred to the Mayo Clinic with congenital heart disease who underwent successful implantation of a transvenous ICD from February 1991 through February 2008 were included. Patients who underwent device implantation with a thoracotomy or nontransvenous approach, those in whom transvenous device implantation was unsuccessful, or patients with hypertrophic cardiomyopathy, arrhythmogenic right ventricular cardiomyopathy, long QT syndrome, and isolated left ventricular noncompaction were excluded. This study was reviewed and approved by the Mayo Clinic institutional review board.
Patient data including clinical notes, operative reports, and outside correspondence were prospectively entered into a centralized database that was retrospectively queried. Demographic data (age, gender, body mass index), underlying congenital diagnosis, radiographic images, and electrocardiographic, echocardiographic, and laboratory data were reviewed. Left ventricular ejection fraction (EF) was calculated using the parasternal long-axis view, as previously described. Systemic right ventricular size and EF were visually estimated by cardiologists with training and expertise in congenital heart disease using all available views. Subpulmonic pressures were estimated using right atrioventricular valve regurgitation signal added to an estimated right atrial pressure based on inferior vena cava size and response to respirations. Secondary prevention indication was defined as ICD implantation because of documented spontaneous ventricular arrhythmia causing hemodynamic instability or requiring cardioversion. Primary prevention indication was defined as the presence of clinical characteristics compatible with a high risk of occurrence of a ventricular arrhythmia.
Implant data including age at implantation, device and lead models, defibrillation threshold, fluoroscopic times, use of intravenous contrast, any periprocedural complications, and duration of hospital stay were abstracted from the electronic medical record.
All ICDs were implanted by transvenous placement of a ventricular lead for defibrillation and pacing using standard techniques under fluoroscopic guidance. In patients with a history of atrial arrhythmias or where sinus node dysfunction or conduction system disease was present, an atrial lead was also placed. The decision to implant a coronary sinus lead for cardiac resynchronization therapy was based on extrapolation of adult cardiology data and guidelines. Specific choice of generator and leads was at the discretion of the operator. Defibrillation threshold testing was performed at the time of device implantation except in the presence of a contraindication such as intracardiac thrombus or concern for hemodynamic instability. An adequate defibrillation threshold testing was defined as a threshold ≥10 J below the maximum output of the device.
All devices were programmed at time of implantation for a ventricular fibrillation zone with a single burst of antitachycardia pacing before shock or during charging, depending on the device model implanted. Devices were interrogated 1 day after implantation to determine appropriate device function. Subsequent device interrogation occurred 1 month after device implantation and every 3 months thereafter or during other scheduled clinical visits if required to exclude ventricular arrhythmia or device complication. Information regarding patients who did not return for device follow-up was obtained from clinical notes and correspondence.
Stored electrograms obtained after interrogation of the device were reviewed to determine appropriateness of each therapy and for classification of ventricular arrhythmias. Each ventricular arrhythmia episode was classified as sustained (requiring antitachycardia pacing or shock therapy to terminate) or nonsustained (spontaneous termination). Ventricular arrhythmias with a cycle length <220 ms were classified as ventricular fibrillation. ICD therapies were defined as appropriate if they were for ventricular arrhythmias and inappropriate if they were for supraventricular rhythms such as atrial fibrillation or sinus tachycardia or because of device malfunction. When stored electrograms were unavailable, assessment was based on the treating clinicians’ interpretation as documented in the clinical notes.
Vascular injury, clinically significant hematoma, cardiac perforation requiring pericardiocentesis, pneumothorax, lead failures requiring revision, or device infection requiring explantation were considered implant-related complications, which were classified as early (occurring ≤28 days after implantation) or late (>28 days after implantation).
Unless otherwise noted, continuous variables were expressed as mean ± SD and categorical variables as absolute number and percentage. Cumulative Kaplan–Meier survival curves were constructed for rates of appropriate and inappropriate ICD shocks. These survival curves were used to calculate rates of patients receiving their first shock episode. Hazard ratios for appropriate and inappropriate ICD shocks were determined using a Cox regression model and calculated for the following categorical variables: age <30.5 years (lowest quartile), age >50 years (highest quartile), gender, presence of atrial lead, secondary prevention indication, PR interval >200 ms, QRS duration >180 ms, body mass index >30 kg/m 2 , systemic EF ≤30%, more than mild subpulmonic ventricular enlargement, subpulmonic ventricular systolic pressure >55 mm Hg (50% of mean systemic systolic blood pressure), hemoglobin <11.4 g/dl (lowest quartile), hemoglobin >14.2 g/dl (highest quartile), and creatinine >1.2 mg/dl (highest quartile). Statistics were performed with JMP 8.0 (SAS Institute, Cary, North Carolina).
Results
In total 73 patients in whom 85 ICDs were implanted were included. Patient characteristics are presented in Figure 1 and Table 1 . All patients underwent preimplantation echocardiography, which demonstrated a systemic EF ≤30% in 25 patients (34%). Subpulmonic systolic ventricular pressure was >55 mm Hg in 20 patients (33% of those documented). In 17 patients (24% of those assessed) the subpulmonic ventricle was severely enlarged.
| Men | 50 (68%) |
| Age (years) | 41 ± 14 |
| Body mass index (kg/m 2 ) | 28 ± 5 |
| Systemic ejection fraction (%) | 44 ± 18 |
| Hemoglobin (g/dl) | 12.9 ± 1.8 |
| Creatinine (mg/dl) | 1.2 ± 0.6 |
| Subpulmonary ventricular pressure (mm Hg) | 51 ± 24 |
| QRS duration (ms) | 167 ± 35 |
| Secondary prevention | 26 (36%) |
| History of nonfatal cardiac arrest | 4 (5%) |
| Sustained ventricular tachycardia, treated with shock | 18 (25%) |
| Sustained ventricular tachycardia, pharmacologic treatment | 2 (3%) |
| Hypotension, concurrent nonsustained ventricular tachycardia | 2 (3%) |
| Primary prevention | 47 (64%) |
| Inducible ventricular tachycardia | 14 (19%) |
| Decreased ejection fraction | 20 (27%) |
| Symptoms, nonsustained ventricular tachycardia on Holter monitor | 8 (11%) |
| Other | 5 (7%) |
Indication for device implantation was for secondary prevention in 26 patients (36%), of whom 4 patients (5% of entire cohort) had a history of sudden cardiac arrest episode. Forty-seven patients (64%) underwent ICD implantation for primary prevention of sudden cardiac arrest. The most common reason was systemic EF ≤40% (27% of entire cohort). The systemic ventricle was a morphologic left ventricle (EF 29 ± 4%, range 20 to 33) in 10 patients and a morphologic right ventricle (EF 23 ± 6%, range 15 to 35) in 10 patients. Other indications for primary prevention included prolonged QRS interval in repaired tetralogy of Fallot (n = 2,206 and 208 ms, respectively) and severe subpulmonic right ventricular dysfunction in a patient with Ebstein anomaly (n = 1). Two other patients underwent device implantation specifically for treatment of atrial arrhythmias; these devices, which were implanted in 2002 with atrial arrhythmia-terminating capabilities, also had ventricular defibrillation capabilities and were therefore included in this analysis.
Initial implantation procedure was a single-chamber ICD in 26 patients (36%), dual-chamber ICD in 39 patients (53%), and a cardiac resynchronization therapy device with defibrillator in 8 patients (11%). Reasons for implantation of a dual-chamber ICD were sinus node dysfunction (n = 4), atrial arrhythmia (n = 12), complete heart block (n = 15), conduction system disease noted on electrophysiologic study (n = 3), bifascicular block (n = 1), first-degree atrioventricular block and right bundle branch block (n = 1), or operator preference (n = 3). Two patients underwent ablations of supraventricular arrhythmias immediately before device implantation. No patient had a ventricular tachycardia ablation procedure before implantation. Fifteen patients (21%) required device replacement during the study period, which was due to elective replacement indicators (n = 7), device recall (n = 3), upgrade to a dual-chamber ICD (n = 2), upgrade to a cardiac resynchronization therapy device (n = 1), device infection (n = 1), and need for a high-output generator (n = 1). Device replacement was 4.1 ± 4.1 years after initial implantation. Three of these device replacements occurred at outside facilities.
Fluoroscopic guidance was used in 81 of the 85 procedures (95%); the remaining 4 cases were generator changes and did not require fluoroscopy. Median duration of fluoroscopy was 10.1 minutes (range 0.1 to 132.6). Intravenous contrast was used in 53 (62%) procedures to aid in vascular access.
Defibrillation threshold testing was performed in 79 (93%) implantation procedures. Three patients had increased defibrillation thresholds that required implantation of a subcutaneous array (n = 1) or epicardial patches (n = 2). Defibrillation threshold testing was deferred in 6 cases because of the presence of severe right ventricular dysfunction (n = 1), atrial fibrillation with inadequate anticoagulation (n = 1), pregnancy (n = 1), impending surgery (n = 2), and inability to induce ventricular fibrillation (n = 1).
There were no intraprocedural complications. One patient required repositioning of the atrial lead 24 hours after implantation because of lead dislodgement. One patient required an additional 24 to 48 hours of hospitalization for each of the following indications: need for intravenous fluids, transient creatinine increase, need for diuresis, postprocedure tachycardia, and anemia requiring blood transfusion. In these patients there were no further complications from the device implantation procedure. An additional 25 patients required prolonged hospitalization (>48 hours) after device implantation that was not directly related to the procedure.
Four patients required lead revision 436 ± 432 days after device implantation because of lead dislodgement (n = 2, 1 atrial lead, 1 right ventricular ICD lead) or increased pacing thresholds (n = 2, 1 right ventricular ICD lead, 1 left ventricular coronary sinus lead). Two patients developed device infections and required device and lead explantation. In 1 of these patients, after an appropriate duration of antibiotics, a new device was implanted. A device was not placed in the other patient. In 1 patient the generator eroded from the preperitoneal space in the left lower quadrant into the peritoneal cavity and required repositioning between the internal and external oblique muscles.
Three patients required device replacement because of device recall (2 with Medtronic Marquis device, Minneapolis, Minnesota; 1 with Guidant A-135, now Boston Scientific, Natick, Massachusetts). A Medtronic Sprint Fidelis lead (Minneapolis, Minnesota) was implanted in 7 patients, 2 of whom were pacemaker dependent, but after review it was decided not to perform a lead revision.
During follow-up (2.2 ± 2.8 years, range 0 to 15) 14 patients (19%) received an appropriate shock for a sustained ventricular arrhythmia (time to first shock 1.9 ± 1.7 years, range 32 days to 5.6 years; Figure 2 ). The rate of first appropriate ICD shock was 12% in the first year, with a mean annual rate of 8.6% for the first 5 years. Risk of appropriate ICD shock was increased with pressure overload of the subpulmonic ventricle: 35% of patients with an estimated subpulmonic ventricular pressure >55 mm Hg received a shock versus 5% of patients with lower pressures (hazard ratio 6.1, confidence interval 1.4 to 41.3, p = 0.01; Table 2 ). Dilation or enlargement of the subpulmonic ventricle was also associated with a trend toward an increased risk of appropriate shocks with 26% of patients with more than mild enlargement of the subpulmonic ventricle receiving an appropriate shock compared to 10% of patients with normal or only mildly enlarged subpulmonic ventricles, although this did not reach statistical significance (hazard ratio 3.04, confidence interval 0.9 to 13.6, p = 0.07).
