Author + information
- Received August 15, 2003
- Revision received December 17, 2003
- Accepted December 23, 2003
- Published online May 19, 2004.
- Fiona Walker, MD*,†,
- Samuel C Siu, MD, SM, FACC*,†,
- Shane Woods*,
- Douglas A Cameron, MD*,†,
- Gary D Webb, MD, FACC*,† and
- Louise Harris, MB, ChB, FACC*,*,† ()
- ↵*Reprint requests and correspondence:
Dr. Louise Harris, Peter Munk Cardiac Centre 3-562, Toronto General Hospital, 150 Gerrard Street West, Toronto, ON M5G 2C4, Canada.
Objectives The purpose of this retrospective study was to define long-term outcomes after pacemaker therapy in adults with congenital heart disease (CHD).
Background Adults with CHD represent a unique and expanding population. Many will require pacemaker or implantable defibrillator therapy, with a lifelong need for re-intervention and follow-up. They pose technical and management challenges not encountered in other groups receiving pacing, and the complication and re-intervention rates specific to this population are not well-defined.
Methods We reviewed outcomes of 168 adults with CHD, 89 females, mean age 40 years, in whom a pacemaker or anti-tachycardia device was implanted.
Results Mean age at implant was 28 years with mean pacing duration 11 years at follow-up (range, 0.5 to 38.0). Seventy-two (42%) received initial dual-chamber devices and remained in this mode, while 23 (14%) went from ventricular to dual-chamber pacing in follow-up. Initial mode of pacing did not have a significant effect on subsequent atrial arrhythmia. Patients receiving an initial epicardial system were younger than those paced endocardially (17 ± 12 years vs. 35 ± 16 years, p < 0.001) and more likely to undergo re-intervention (p = 0.019). Difficulty with vascular access was encountered in 25 patients (15%), while 45 (27%) experienced lead-related complications. No significant predictors of lead complications were identified.
Conclusions Lead complications were not significantly different for epicardial versus endocardial, nor physiologic versus ventricular pacing, but a trend toward improved lead survival in patients receiving endocardial leads at first implant was observed. Adults with CHD remain at risk for atrial arrhythmias regardless of pacing mode.
Adults with congenital heart disease (CHD) are part of a unique and expanding patient population. Some of these patients will require permanent pacemaker (PPM) or defibrillator therapy, with the attendant need for re-intervention and follow-up over a lifetime. Although the transvenous approach for implantation is preferred in adults, this may not always be technically feasible, not only because of abnormal anatomy but because reparative or palliative surgery may limit vascular access. Epicardial pacing circumvents the need for vascular access but often at the expense of long-term lead durability with an increased requirement for surgical re-intervention (1). Adults with CHD therefore pose technical and management challenges not encountered in other paced groups.
Complication and re-intervention rates specific to this population are unknown. Pacing experience in adults with acquired heart disease is not necessarily representative of this group, and it may not be appropriate to extrapolate from experience in the pediatric paced population. The purpose of this retrospective study was to define long-term outcomes after PPM in adults with CHD.
This retrospective study represents our experience with PPM, or anti-tachycardia device therapy (ATD), in consecutive adults with CHD (age ≥18 years as of July 2000) at the Toronto Congenital Cardiac Centre for Adults (TCCCA). Patients were identified from the pacemaker implantation registry (Toronto General Hospital) and the database of the TCCCA.
Patients with pacemakers implanted for isolated congenital complete atrio-ventricular (AV) block or isolated long QT syndrome were excluded. The University Health Network Research Ethics Board approved this study.
The following data were collected from patient charts: congenital anatomic diagnosis; type and number of cardiac surgeries; age at cardiac surgery(s) and at pacemaker implant(s); electrophysiologic (EP) indication for PPM; type of device and leads implanted/explanted; method of implant (endocardial/epicardial); number, indication, and age at re-operations/re-interventions; pacing mode and mode change over time; complications; nature and frequency of arrhythmias. Atrial arrhythmias (AT) were detected by: 1) chart review and electrocardiographic (ECG) documentation of all visits to the TCCCA; and/or 2) atrial electrogram and ECG documentation of six monthly visits to the pacemaker clinic; and, more recently, 3) stored logs and electrograms from newer generation dual-chamber pacemakers. Routine transthoracic echocardiograms are performed annually for patients of the TCCCA. Transesophageal echocardiography and venography to confirm thrombosis were performed only when clinically indicated. We defined lead complications as any of: lead fracture, lead recall, thrombus or vegetation on pacing lead, or wound infection/erosion. Major clinical complications include venous thrombosis, stroke, or cardiac death. Follow-up data were updated to December 2001.
Statistical analysis was performed using SPSS for Windows (Version 10.0, SPSS Inc., Chicago, Illinois). Data are presented as a mean ± SD. Comparisons between patients were performed with the chi-square, Fisher exact, or ttest as appropriate. The Fisher exact test was utilized in place of chi-square test when the expected frequency of a cell within a contingency table was <5. Freedom from re-operation, lead complications, and AT was calculated by the Kaplan-Meier method when the data were stratified by the route (endocardial vs. epicardial) and mode of pacing (dual-chamber or atrial vs. ventricular). Cox regression was used to identify predictors of lead complications in follow-up. Univariate predictors of events with a significance <0.10 were entered into a multivariate Cox proportional hazards model using a backward elimination algorithm. Level of statistical significance was set at 0.05 (two-sided).
A total of 168 adult patients with CHD were identified from the databases as having a pacemaker or ATD implanted. Of these, 89 were female and 79 male, mean age 40 years (range, 17 to 85 years), at time of data collection. There was a total of 1,582 paced years for the study population, with a mean duration of pacing 11 ± 9 years; median, 10 years; range, 0.5 to 38.0 years. The first implant for most patients occurred in adulthood, with 113 first implants in patients >18 years. Fifty-five first implants (33%) were performed in childhood (1 to 18 years), and five of these in infants <1 year of age. The mean age at first implant for the study population was 28 years. Re-intervention was required in 63% of patients: 39 had a total of two pacemaker interventions, 31 had 3 pacemaker surgeries, 16 had 4 pacemaker surgeries, 8 had 5 pacemaker surgeries, and 12 had >5 pacemaker surgeries. The indication for a pacemaker re-intervention was generator end-of-life characteristics in 57 patients, lead-related in 34, and both indications in 15. A total of 71% of patients with an epicardial system and 48% with an endocardial system required >1 pacemaker (p = 0.007 by chi-square test).
The congenital anatomic diagnoses were varied with both simple and complex defects represented (Table 1). Patients with atrial, ventricular, and AV septal defects comprised almost 30% of the paced population. Transposition of the great arteries and congenitally corrected transposition accounted for 23%, tetralogy of Fallot 14%, and a substantial proportion (17%) had complex anatomic CHD (including tricuspid atresia, pulmonary atresia, single ventricle, and so on). Details of pacemaker history and outcomes for the four anatomic diagnostic categories with the largest numbers of patients as per Table 1are provided in Table 2.
Indications for pacing are outlined in Table 3. Seventy-six patients (45%) required pacing peri-operatively after cardiac surgical repair or palliation. Of the remaining 92 patients, 87 received pacemakers for late sinus node or AV nodal conduction disease, the majority (78 patients) having undergone surgery years previously. The indication for PPM was AV block in 65% patients, sinus node disease in 29%, and long QT or tachycardia in the remainder.
To simplify data analysis, anatomic congenital diagnoses were grouped into three categories: shunts (including AV septal defects with associated AV valve disease) = group 1 (n = 53), aortic valve disease with or without coarctation of the aorta = group 2 (n = 16), and complex = group 3 (n = 99). The latter included patients with tetralogy of Fallot, Mustard repair, Fontan repair, and double outlet right ventricle. The type of surgical intervention was also subdivided into four categories: none = group 1 (n = 10); shunt closure = group 2 (n = 39); valve repair/replacement/coarctation of the aorta = group 3 (n = 21); complex surgery (Mustard/Fontan, etc.) = group 4 (n = 98). The arrhythmia diagnostic groups were divided into: group 1 = patients with peri-operative sinus or AV nodal disease; group 2 = remote or non-surgical disease; and group 3 = long QT/tachycardia as the indication for pacing. The relationship between the CHD and EP diagnostic groups (Table 4) and type of cardiac surgery and EP diagnostic group (Table 5) were both statistically significant. Patients undergoing valve surgery, in particular, aortic valve surgery, were most likely to require pacing peri-operatively. Patients with shunt repair or those requiring complex surgical intervention had a greater likelihood of pacemaker requirement in follow-up than the former group.
Endocardial versus epicardial pacing
Of 168 patients, 106 (63%) were paced endocardially at first implant, and 62 (37%) epicardially. Over time, however, 24 converted from epicardial to endocardial pacing and three from endocardial to epicardial pacing.
An epicardial pacing system as the initialimplant was significantly associated with early surgically induced sinus node disease or AV block (60% vs. 40% patients, p = 0.015 by chi-square test). Patients in whom the initial implant was epicardial were also younger (17 ± 12 years vs. 35 ± 16 years, p < 0.001) and more likely to undergo repeat pacemaker surgical interventions in follow-up (74% vs. 57% patients with >1 pacemaker intervention, p = 0.023 by chi-square test). The time from implant to re-operation for patients receiving an initial epicardial versus endocardial implant (Accufix [Telectronics Pacing Systems, Colorado] leads excluded) was, however, not significantly different (median survival 7 vs. 8 years, respectively, p = 0.31). Lead complications—lead recalls, fractures, and so on—occurred in 36% and 26% of patients with epicardial and endocardial systems, respectively, but were not significantly different between the two groups (p = 0.64 by chi-square test). When patients with transvenous Accufix leads were excluded, there was a trend toward improved lead survival in patients receiving endocardial leads as their first implant (mean survival, 24.2 vs. 23.8 years, p = 0.078) (Fig. 1).
Mode of pacing
The mode of pacing and change in mode over time is summarized in Table 6. Seventy-two (42%) patients received dual-chamber implants at the outset and remained in this mode, while 23 (14%) were upgraded from ventricular to dual-chamber pacing in follow-up. Atrial pacing was observed in 10%, with only one patient upgraded to dual-chamber and one converted to ventricular pacing in follow-up. Pulse generator longevity (median time to replacement) was not significantly associated with either underlying congenital anatomic diagnosis (7 vs. 8 years, non-complex vs. complex, p = 0.18) or mode of pacing (7 vs. 8 years, ventricular vs. dual-chamber pacing, p = 0.27). Similarly, freedom from lead complications (Accufix leads excluded) was not significantly different for patients with physiologic (dual-chamber or atrial-based pacing) compared with ventricular pacing (p = 0.29) (Fig. 2).
Incidence of AT
Sustained AT were observed in 76 patients (45%). Five variables were evaluated as possible predictors for AT after pacing: history of AT before pacing, physiologic pacing (dual-chamber and atrial-based pacing), age at initial implant > the median age of 23 years, gender, and complex anatomy. Univariate analysis identified AT before pacing as significant (p < 0.001) and complex anatomy as of borderline significance (p = 0.10). Subsequent multivariate analysis, however, confirmed AT before implant was the only independent predictor of arrhythmia after implant (p < 0.001, risk ratio = 7.41). Physiologic pacing did not appear to be protective against subsequent AT when compared with ventricular pacing (Fig. 3).
Forty-five patients (27%) experienced lead-related complications, with some patients experiencing more than one lead-related event. Lead fractures were encountered in 17 patients (2 of whom also had leads on recall), while primary lead recalls accounted for lead replacement in 28. There were 22 lead complications in 1,582 patient years of follow-up or 1.4 events per 100 patient years of follow-up. As a consequence of anatomic and surgical constraints, transvenous active fixation atrial and/or ventricular leads were implanted in many of these patients, and concerns with the Accufix lead system accounted for the majority (27 patients) of these recalls. Three patients (2 post-Mustard repair, 1 atrio-pulmonary Fontan repair) were identified by transthoracic echocardiography as having thrombus on their pacing leads. Two of these patients were experiencing sustained AT at the time of detection, and the third had paroxysmal atrial flutter. Vegetations on the pacing leads occurred in one patient in long-term follow-up. Wound infections occurred in six (in association with lead fracture in 2, lead recall in 3) and pulse generator erosion in two (in association with lead recall in one, lead malposition and thrombus in the other).
A 67-year-old patient with established atrial flutter and a transvenous ventricular pacemaker suffered a stroke while awaiting device closure of an atrial septal defect. Although the stroke may have been attributable to the arrhythmia, a paradoxical embolus originating from the pacemaker lead could not be excluded.
Difficulty with vascular access was encountered in 25 patients (15%): 11 had venous obstruction, nine had anomalous veins, and venous access was precluded in five as a consequence of previous surgery. One patient experienced acute subclavian vein thrombosis. Five variables were evaluated as possible predictors for lead complication: age at first implant, number of cardiac surgeries, complexity of anatomic diagnosis, gender, and mode of pacing. None of these were independently predictive. The incidence of lead complications when Accufix leads were excluded was not significantly different for patients with physiologic (dual-chamber or atrial-based) compared with ventricular pacing.
Of four deaths in follow-up, two were sudden and occurred in patients with previously repaired tetralogy of Fallot and Mustard repair, respectively. No further details were available regarding these deaths. One patient with previous atrial septal defect repair and associated pulmonary hypertension died of congestive heart failure. The circumstances of the remaining death are unknown.
Four patients had implantable defibrillators, and three had atrial ATD.
As a consequence of the ability to surgically repair or palliate patients with complex CHD, 85% of those born with congenital defects will survive to adulthood (2). Many of these patients will have a life-long need for pacing as a direct result of such surgery, but others may come to require pacing later in life. Yet others may require device implantation to provide anti-tachycardia pacing or to facilitate drug therapy of tachyarrhythmias. Improved lead design and rapidly evolving device technology have provided the clinician with more options in patient management and ever-expanding indications for device implantation (3,4). Adults with CHD, however, pose management and technical challenges not encountered in other paced groups.
Our cohort of 168 consecutive adult patients with CHD and device therapy had a median follow-up of 10 years. The patient with the longest duration of pacing, still actively followed, received her first pacemaker in 1963. The long duration of follow-up enabled us to determine the implications and outcomes of long-term pacing in this unique patient population. Two-thirds of patients required two or more pacemaker interventions in follow-up. Dual-chamber or atrial-based pacing was accomplished and maintained in the majority of patients (66%), with only four converting to ventricular pacing in follow-up. The incidence of lead complications was 3% per patient-years, and wound infections or pulse generator erosion was observed in 5% of patients overall.
Lead-related complications (lead fracture and recall) and difficulty with vascular access were commonly encountered, occurring in 27% and 15% of patients, respectively, with some patients experiencing more than one lead-related event. The incidence of lead complications (Accufix leads excluded) was not significantly different for patients with physiologic (dual-chamber or atrial-based) pacing compared with ventricular pacing. In a previous study from a pediatric population with CHD (5), atrial, as opposed to ventricular, lead placement was identified as predictive of subsequent lead failure. Dual-chamber or “physiologic” pacing in the adult population without CHD has traditionally been viewed as the preferred mode of pacing, as there is suggestive evidence this may be associated with a lower risk of stroke, atrial fibrillation, and death (6–8). A recent multicenter study comparing ventricular versus physiologic pacing, however, failed to demonstrate a significant difference in end points of heart failure, stroke, or death although physiologic pacing was associated with a small, late reduction in the risk of atrial fibrillation (9). Atrial arrhythmias, frequently encountered in the adult with CHD, contribute substantially to morbidity and mortality in this population (10). Any management strategy reducing the incidence of such arrhythmias would be a welcome contribution. In this cohort of adults with CHD, a substantial proportion of patients (45%) experienced sustained AT after pacing. Multivariant analysis identified only AT before pacing as predictive of arrhythmia subsequent to pacing. Although there was a progressive occurrence of AT for this group of patients in ongoing follow-up, we were unable to demonstrate a significant effect of the mode of pacing on the incidence of subsequent AT (Fig. 3). In a small cohort of Fontan patients, Fishberger et al. (11)found AV synchrony with dual-chamber pacing conferred long-term survival benefit compared with ventricular pacing, but this did not reach statistical significance.
The transvenous route was utilized as the initial approach for pacing in the majority of our patients (68%) and, of those implanted epicardially, a further 39% were subsequently converted to a transvenous system. Lead complications, although more frequent in those with epicardial systems, did not differ significantly between those with epicardial versus endocardial pacing. This was, in part, accounted for by the frequent use of active fixation atrial and/or ventricular leads, specifically the Accufix lead, in our patients with transvenous systems. The Accufix leads were subject to recall in the mid-1990s, secondary to a design flaw reported to be associated with fatalities (12). Active fixation leads are often required in this population either as a consequence of the underlying anatomic substrate or of the previous cardiac surgery—for example, to accomplish transvenous pacing in the non-trabeculated morphologic left ventricle after Mustard repair. When patients with Accufix leads were excluded, there was a trend toward improved lead survival in patients receiving endocardial leads as their first implant.
Patients with epicardial pacing systems were younger at initial implant and more likely to undergo repeat pacemaker surgery in follow-up. These findings are consistent with a recent report comparing these two approaches in a pediatric population (1). With a follow-up of 3 to 3.4 years in a population with a mean age of 5.3 years, Sachweh et al. (1)found lead-related re-operations more frequently in patients with epicardial leads as a consequence of higher initial and long-term pacing stimulation thresholds. The latter may no longer be applicable, however, to the current generation of epicardial systems, where improved lead design and technology have resulted in improved long-term stimulation thresholds. Two recent publications (13,14)attest to the long-term durability of steroid-eluting epicardial leads in pediatric patients with CHD.
Difficulty with vascular access was encountered in 15% of patients as a consequence of anomalous veins, previous surgery, or venous obstruction. The latter has been reported to occur in 30% to 45% of patients after transvenous pacing and is usually asymptomatic (15). Only one patient experienced acute venous thrombosis. The majority of our patients have had more than one transvenous pacing system implanted with as many as three to four leads traversing a single vein over the duration of follow-up. With the recent advent of laser extraction systems for lead removal (16), multiple leads in single veins should no longer be encountered, and new leads can be introduced even in the presence of venous obstruction. For patients with a history of previous cardiac surgery, it is helpful to document venous anatomy before pacemaker implant with venography. In those with more complex repairs, an operator familiar with the congenital anatomy should undertake the implant. In addition to issues of venous access, cardiac access via the transvenous route may also be precluded—for example, the patient with an atrio-pulmonary Fontan repair requiring ventricular pacing. Consideration should be given to intraoperative lead placement at cardiac surgical repair for conditions known to be associated with bradyarrhythmias and where transvenous access will be precluded after surgery (e.g., post-Fontan repair). Subsequent extensive adhesions may limit satisfactory epicardial atrial lead placement in particular.
Thrombus on pacing leads was observed in three patients, all of whom had a history of AT, and one additional patient suffered a stroke possibly attributed to the pacemaker. Although echocardiography would have been performed routinely in follow-up, it is certainly feasible that the incidence of thrombus has been underestimated. Thrombus on transvenous pacing leads in adults with acquired heart disease is an established complication of pacing (17). It is likely that CHD patients in whom enlarged atria and low-velocity flow states are frequently encountered (18)are at increased risk for such thrombus. In the Fishberger et al. series (11), no embolic complications were observed in 12 Fontan patients receiving transvenous atrial pacing, only two of whom were anti-coagulated with coumadin. Thrombus formation and embolic complications are not described in the Gillette et al. (19)report of pacing outcomes after Mustard repair, nor in other studies of long-term transvenous pacing in patients with CHD (20,21).
A large series of long-term pacing outcomes in this population has not been previously reported; thus, this was a preliminary study and is, therefore, hypothesis-generating. This was a retrospective study with data collection limited to variables available from clinical records. Certain complications that are more likely to be asymptomatic, such as lead thrombus, may be underestimated. The number of patients with implantable defibrillators and ATD was insufficient for separate analysis of the outcomes of these devices. A total of 16% of patients did not have follow-up data in the last two years of data collection. It was not possible to determine lesion-specific outcomes (such as comparing outcomes between different types of shunts) due to the small number of patients within each anatomic subgroup.
Historically, pacing in the patient with CHD was primarily required for AV block, after repair of septal defects and tetralogy of Fallot. In the current era, the incidence of surgical AV block in these lesions has decreased but has been offset by a corresponding increase in pacing after repair of increasingly complex defects. Furthermore, indications for pacing have evolved beyond bradycardia support to include hemodynamic optimization and therapy for tachyarrhythmias. This has important implications for the adult with CHD, in whom ventricular dysfunction and arrhythmias are frequently encountered. Planning for device therapy in these patients necessitates consideration of anatomic and surgical data with careful attention to coexisting arrhythmias, issues of vascular access, and intracardiac shunting. In this study, the incidence of lead complications (Accufix leads excluded) was not significantly different for epicardial versus endocardial pacing, nor for physiologic versus ventricular pacing. Disappointingly, the adult patient with CHD remains at risk for AT in follow-up regardless of the mode of pacing. Modifications of cardiac surgical techniques, the trend toward earlier age for surgical correction, and newer AT preventative algorithms may yet alter this outcome.
- atrial arrhythmias
- anti-tachycardia device therapy
- congenital heart disease
- permanent pacemaker
- Toronto Congenital Cardiac Centre for Adults
- Received August 15, 2003.
- Revision received December 17, 2003.
- Accepted December 23, 2003.
- American College of Cardiology Foundation
- Sachweh J.S,
- Vazquez-Jimenez J.F,
- Schondube F.A,
- et al.
- Moller J.H,
- Taubert K.A,
- Allen H.D,
- Clark E.B,
- Lauer R.M
- Israel C.W,
- Ehrlich J.R,
- Gronefeld G,
- et al.
- Hesselson A.B,
- Parsonnet V,
- Bernstein A.D,
- Bonavita G.J
- Cohen M.I,
- Bush D.M,
- Vetter V.L,
- et al.
- Wilkoff B.L,
- Byrd C.L,
- Love C.J,
- et al.
- Gillette P.C,
- Wampler D.G,
- Shannon C,
- Ott D
- Shahar E,
- Feldman S,
- Shem-Tov A,
- Hegesh J,
- Barzilay Z,
- Neufeld H.N