Author + information
- Received February 23, 2007
- Revision received June 18, 2007
- Accepted June 25, 2007
- Published online September 18, 2007.
- Elliot M. Tucker, MD⁎ (, )
- Lee A. Pyles, MD, FACC,
- John L. Bass, MD and
- James H. Moller, MD, FACC
- ↵⁎Reprint requests and correspondence:
Dr. Elliot M. Tucker, University of Minnesota, MMC 94, 420 Delaware Street SE, Minneapolis, Minnesota 55455.
Objectives This study sought to discover the incidence of permanent pacemaker (PPM) placement for atrioventricular conduction block (AV block) after operative repair of perimembranous ventricular septal defect (PMVSD) in a large multi-institutional database and in the subgroup of patients comparable to those considered for transcatheter device closure of PMVSD.
Background Atrioventricular conduction block is a complication of operative repair of PMVSD and of device closure of this defect. Earlier reports do not report the incidence of AV block by VSD type.
Methods The Pediatric Cardiac Care Consortium database was searched for all children who had operative PMVSD repair except those with abnormalities that increase risk of AV block. The patient group was searched for those with subsequent PPM placement for AV block. Demographic data and time to PPM placement were available for all patients.
Results Of 4,432 patients with PMVSD repair, 48 (1.1%) underwent PPM placement for AV block. The PPM group was more likely to have Down syndrome (41% vs. 18%; p < 0.001), was younger (mean age 14 vs. 26 months; p < 0.001), and had longer mean length of postoperative hospital stays (20 vs. 8 days; p < 0.001). The most significant risk factor for AV block was Down syndrome (odds ratio 3.62, 95% confidence interval 2.02 to 6.39; p < 0.005). Modal time to PPM placement was 7 days (range 0 to 4,078 days). Out of 1,877 patients comparable to those currently considered for device closure, 13 (0.8%) underwent PPM placement after PMVSD repair.
Conclusions Operative AV block and PPM placement occurred in 1.1% of patients in the total group and in 0.8% of patients comparable to those considered for device closure of PMVSD. A PPM placement is more likely in patients with Down syndrome. These data should be considered as devices are developed and in the future when counseling families about options for PMVSD closure.
Atrioventricular conduction block (AV block) is a well-described complication of operative repair of perimembranous ventricular septal defect (PMVSD). Large single-center studies report a 0% to 3% incidence of AV block after operative repair of ventricular septal defects (VSD), but these studies do not report the incidence of AV block by type of VSD (1–5). Recently, there have been reports of a 2% to 3.6% incidence of AV block requiring permanent pacemaker placement immediately and late after transcatheter device closure of PMVSD (6,7). The present study was performed to determine the incidence of permanent pacemaker (PPM) placement for AV block after operative repair of uncomplicated PMVSD and to identify associated risk factors. A secondary aim was to identify the incidence of PPM placement for AV block in the subgroup of patients who are comparable to those currently considered for transcatheter device closure of PMVSD (8).
The Pediatric Cardiac Care Consortium database (PCCC) has enrolled over 60,000 patients and has data for over 83,000 operations performed in North America since 1982 (9). The database was searched for all patients younger than 21 years of age with a primary diagnosis of PMVSD and operation performed primarily to repair this defect between 1982 and 2003. Patients with atrioventricular or ventriculoarterial discordance, ventricular outflow tract obstruction, subaortic fibromuscular ridge, aortic valve insufficiency or aortic cusp prolapse into the VSD, or multiple VSDs were excluded to simulate the population of patients who would currently be considered for transcatheter device closure of PMVSD. Patients with a coexistent atrial septal defect, previous pulmonary artery band placement, or previous repair of coarctation of the aorta or interrupted aortic arch were not excluded. This patient group was searched for all patients undergoing placement of a PPM for either second- or third-degree AV block after PMVSD repair. Patients developing AV block after a subsequent cardiac operation, unrelated to initial repair, were excluded from the PPM group, because the AV block was deemed unrelated to the PMVSD repair. The PCCC data, including age and weight at time of PMVSD operative closure, date of closure, cardiac center at which VSD closure occurred, length of postoperative hospital stay, hospital mortality, gender, and presence of Down syndrome, were collected for each patient. Time from operation to PPM placement was determined. The total dataset was analyzed to determine characteristics of patients requiring placement of PPM and risk factors for PPM placement. Two sub-analyses were also performed: 1) comparing early (1982 to 1992) and contemporary (1993 to 2003) eras for analysis of patient characteristics and outcome differences; and 2) analysis of the subgroup of patients with body weight ≥8 kg; patients meeting our inclusion criteria with body weight >8 kg would be comparable to those currently considered for transcatheter device closure of PMVSD.
Weight for age zscore was calculated based on age and gender using standard normal Centers for Disease Control and Prevention growth charts for all patients (10). Patient groups were compared using 2-sample ttests for continuous variables and chi-square tests for categoric variables. Cox proportional hazards regression was used to calculate hazard ratios for PPM placement. In the analysis, time to event was time from PMVSD repair to PPM placement in days. Patients were censored at death or at the end of the treating institution’s participation in the PCCC, whichever came first. The JMP version 6.0 (Cary, North Carolina) software was used for all analyses.
There were 4,432 patients with a PMVSD repair at 51 different cardiac surgical centers. Forty-eight (1.1%) of the 4,432 underwent PPM placement for AV block. There were 789 (18%) patients with Down syndrome. Hospital death occurred in 69 (1.6%) patients after PMVSD repair for the period 1982 to 2003. Comparison of the PPM group with the non-PPM group, respectively, reveals the following: a mean age of 14 versus 26 months (p < 0.001), weight for age zscore of −2.6 versus −2.2 (p = 0.09), mean length of postoperative hospital stay 20 versus 8 days (p < 0.001), and 44% versus 18% with Down syndrome (p < 0.001) (Table 1).
In univariate and multivariate regression analysis, the only significant risk factor for PPM placement is the presence of Down syndrome (odds ratio [OR] 3.62, 95% confidence interval [CI] 2.02 to 6.39, p < 0.005). Year of operation as a continuous variable, age in months at PMVSD repair, and weight for age zscore at operation did not affect the incidence of PPM placement (Table 2).
Analysis to determine if PPM placement varied as a function of pediatric cardiac center surgical volume revealed no correlation (p = 0.37).
Thirty-two patients (67%) underwent PPM placement within 2 weeks of PMVSD repair. Modal time to PPM placement was 7 days, with a range of 0 to 4,078 days. In 10 patients (21%), PPM was placed more than 1 month after PMVSD repair. Out of these 10 patients, 5 (50%) had Down syndrome. Five patients had PPM placement between 1 month and 6 months after PMVSD repair. Three of these remained hospitalized after PMVSD closure and had AV block, and PPM placement was for insufficient cardiac output associated with slow ventricular rate. One patient was discharged with second-degree AV block and another was discharged in sinus rhythm; both patients returned with AV block requiring PPM within 6 months of PMVSD repair.
The remaining 5 patients received PPM placement for AV block between 4 and 11 years after PMVSD repair. One of these patients had postoperative AV block with an adequate ventricular rate, but underwent PPM because of fatigue years later. Another patient had immediate postoperative AV block that resolved before hospital discharge, only to return 4 years later. The third patient did not have perioperative AV block but presented with a cardiac arrest and AV block at age 13 years. The remaining 2 patients had no perioperative AV block but presented with syncope 8 years after PMVSD closure and were found to have AV block.
Comparison of patients who had PMVSD repair in the early era (n = 1,053) with the contemporary era (n = 3,379) revealed: mean age did not change significantly 24.4 versus 25.9 months (p = 0.26), mean weight for age zscore increased from −2.55 to −2.06 (p < 0.001), mortality decreased from 3.2% to 1.0% (p < 0.001), and mean length of postoperative hospital stay decreased from 11 to 8 days (p < 0.0001) (Table 3).In this crude comparison, the incidence of PPM placement decreased from 1.5% to 0.7% (p = 0.02), but, as stated in the preceding, year of operation as a continuous variable is not a significant risk factor for PPM placement. Univariate and multivariate regression analysis of the contemporary era also revealed the presence of Down syndrome as the only significant risk factor for AV block and PPM placement (OR 2.91, 95% CI 1.29 to 6.40; p = 0.01) (Table 4).
There were 1,739 patients who weighed more than 8 kg and would be comparable to those currently considered for transcatheter device closure of PMVSD. Thirteen (0.8%) of these patients underwent PPM placement for AV block after PMVSD repair. There were 9 hospital deaths (0.5%) in this group. In the contemporary era, of the 1,315 patients weighing >8 kg there was a significant drop in PPM placement to 5 (0.4%; p < 0.005) and hospital mortality to 3 (0.2%; p = 0.02).
This large group of patients undergoing operative repair of PMVSD is from 51 centers over the course of 21 years. The data include all patients from the participating centers for the years of their participation in the PCCC. This mixture of centers with different case volumes and operations performed over many years is representative of the incidence of PPM placement after PMVSD repair in North America. It is possible that our database review could underestimate the incidence of PPM placement or death occurring after hospital discharge if these occurred outside of a PCCC participating institution, but the 1.1% incidence of PPM placement in our overall patient group and the 0.8% incidence in the contemporary era are comparable to published data for all types of VSD (1,2). In the present comparison of patient groups, patients who have PPM placement are significantly younger than those who do not have PPM placement, but as a continuous variable younger age in months is not a risk factor for PPM placement. Lower weight for age zscore also did not meet significance as an independent risk factor for PPM placement. As expected, length of postoperative hospital stay was longer in those receiving PPM.
The most significant risk factor for AV block identified was Down syndrome, which carried a 2.7% incidence of PPM placement in these patients. The anatomic relationship of the conduction system to PMVSD (11,12), and anomalies of the conduction system in patients with Down syndrome and endocardial cushion defects are well described (13), but no specific reports on the anatomy of the conduction system in patients with PMVSD and Down syndrome were found in our literature search. Our multivariable regression controlled for the effect of age and weight on the need for PPM, so the fact that patients with Down syndrome are frequently repaired at a younger age and smaller weight than other PMVSD patients does not explain their increased risk of PPM placement. This suggests that there may be a factor in patients with Down syndrome placing them at higher risk for AV block after operative repair of PMVSD. This warrants further study.
A previous PCCC study of the relationship between center volume and hospital mortality after VSD closure revealed no relationship (14). Likewise, we found no relationship between cardiac center surgical PMVSD closure volume and incidence of PPM for AV block. This may be related to the relative infrequency of AV block.
The finding of late AV block in which a PPM was placed more than 4 years after VSD repair is interesting and has been described before in patients both with and without Down syndrome (15–17). The present finding of 4 (0.1%) of 4,432 patients developing AV block more than 4 years after operative repair of PMVSD is important. This fact should be considered when counseling the patient, parents, and personal physicians of these patients about symptoms which might develop and need appropriate evaluation.
Comparison of patients having PMVSD repair in a contemporary era with those repaired in an earlier era revealed that, although age of repair has not changed significantly, contemporary patients weighed more than in the early era. Although hospital mortality rate, incidence of PPM for AV block, and postoperative length of stay were all better in the contemporary era, patient characteristics linked to PPM placement for AV block in the overall data set were unchanged. The decrease in both mortality and incidence of AV block supports the conclusion that the care of these patients has improved, but patients with Down syndrome continue to have increased risk for AV block.
The subgroup of patients who might meet inclusion criteria for device closure of PMVSD had a 0.8% incidence of AV block requiring PPM and 0.5% hospital mortality over both eras and a 0.4% incidence of PPM placement and 0.2% hospital mortality in the contemporary era. These figures should be considered as new devices are developed and tested and when counseling families about options for PMVSD closure.
Identification of patients with AV block requiring PPM placement that occurred after hospital discharge was only possible because of the PCCC’s maintenance of a Health Insurance Portability and Accountability Act of 1996-compliant unique patient identifier (9). The assignment of this unique patient identifier makes it possible to perform patient-specific longitudinal follow-up within the registry data. This sort of longitudinal follow-up within a multicenter registry is invaluable when assessing long-term outcomes after operative or catheter intervention for rare acquired or congenital heart diseases.
The present retrospective review has a few limitations. Patients that had heart catheterizations or cardiac surgery at the participating PCCC centers were entered into the database, but we had no further data about these patients unless they returned to a PCCC center for another heart catheterization or cardiac surgery. As stated in the preceding, the present data could underestimate the incidence of PPM for early or late AV block if patients have PMVSD repair at a PCCC institution but have PPM placement for later AV block at another facility after discharge. Similarly, mortality data were limited to hospital mortality. If late PPM placement or death is underestimated, this could alter our assessment of risk factors for PPM placement after PMVSD repair. The present study is also limited by the data collected in the PCCC. Proposed methods to avoid surgical AV block, such as tricuspid valve detachment (3,5), are not recorded in the database, and we cannot draw conclusions related to that technique’s effects on the incidence of AV block. The rigorous PCCC data collection and coding methods have been recently described and are an unlikely source of error (9).
The authors thank Virgil Larson for his assistance with data management in preparation of this manuscript.
- Abbreviations and Acronyms
- AV block
- atrioventricular conduction block
- Pediatric Cardiac Care Consortium
- perimembranous ventricular septal defect
- permanent pacemaker
- ventricular septal defect
- Received February 23, 2007.
- Revision received June 18, 2007.
- Accepted June 25, 2007.
- American College of Cardiology Foundation
- Bol-Raap G.,
- Weerheim J.,
- Kappetein A.P.,
- Witsenburg M.,
- Bogers A.J.
- Gaynor J.W.,
- O’Brien J.E. Jr..,
- Rychik J.,
- Sanchez G.R.,
- DeCampli W.M.,
- Spray T.L.
- Nygren A.,
- Sunnegardh J.,
- Berggren H.
- Fu Y.C.,
- Bass J.,
- Amin Z.,
- et al.
- Latham R.A.,
- Anderson R.H.
- Banks M.A.,
- Jenson J.,
- Kugler J.D.