| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Eur J Cardiothorac Surg 2005;28:61-68
© 2005 Elsevier Science NL
a Department of Cardiothoracic & Vascular Surgery T, Aarhus University Hospital, Skejby Sygehus, Brendstrupgaardsvej, 8200 Aarhus N, Denmark
b Department of Clinical Epidemiology, Faculty of Health Sciences, Aarhus University, 8000 Aarhus C, Denmark
c Department of Cardiology B, Aarhus University Hospital, Skejby Sygehus, Brendstrupgaardsvej, 8200 Aarhus N, Denmark
Received 19 December 2004; received in revised form 4 March 2005; accepted 4 April 2005.
* Corresponding author. Address: Department of Cardiothoracic & Vascular Surgery T, Aarhus University Hospital, Skejby Sygehus, Brendstrupgaardsvej, 8200 Aarhus N, Denmark. Tel.: +45 89495483; fax: +45 89496016. (Email: morten.smerup{at}ki.au.dk).
 |
Abstract |
|---|
 Top Abstract 1. Introduction2. Methods3. Results4. DiscussionReferences |
|---|
Key Words: Pacemaker • Cardiac surgical procedures • Heart block • Mortality
|
1. Introduction |
|---|
|
TopAbstract1. Introduction2. Methods3. Results4. DiscussionReferences |
|---|
Previous studies have primarily focused on in-hospital risk of pacemaker implantation after congenital heart surgery and the data on long-term prognosis of patients as well as pacemaker systems after these procedures are sparse.
Furthermore, early as well as more recent studies report occurrence of late onset conduction disturbances years after congenital heart surgery [11,12]. Whether this phenomenon is caused solely by the type of surgery, is related to the underlying structural heart disease, or additional determinants are in play remains to be further elucidated.
We therefore performed a historical follow-up study of all patients who had pacemaker implantation due to post-surgical AV-block or sinus node dysfunction in the period 1981–2002 at Skejby Sygehus, Aarhus University Hospital, Denmark. The aim was to examine prognostic factors (possible predictors) of mortality, failure of systems (in terms of the opposite; event-free survival of systems) and timing of implantation after surgery in post-surgical pacemaker patients.
|
2. Methods |
|---|
|
TopAbstract1. Introduction2. Methods3. Results4. DiscussionReferences |
|---|
Mortality was defined as death at any time during the follow-up period. Event-free survival was defined as absence of explantation (death was registered as explantation of pacemaker system).
Survival time analysis for patients (mortality) as well as for 1st system (event-free survival) was performed after stratification by a number of potential and identifiable dichotomised predictors (sex, preoperative RACHS-1 score [13], age at operation, age at 1st implantation, epicardial/transvenous lead, historical time of 1st implantation (before or after 1995), historical time of surgery (before or after 1995) and early/late implantation after surgery) using Kaplan–Meier plots with survival time estimates for 75, 50 and 25% survival proportions if possible. Any differences between values were evaluated by Cox proportional hazards analysis in terms of relative risk (RR) and 95% confidence intervals (CI) before (RR-b) and after (RR-a) adjustment for confounding by other potential prognostic factors (15% ‘change-in-estimate’ method [14]). Only one of the variables ‘age at operation’ and ‘age at 1st implantation’, and similarly ‘historical time of surgery’ and ‘historical time of 1st implantation’ was included in the same Cox analysis due to causal interdependence. A maximum number of four independent variables were entered in Cox analysis due to limited number of observations. The Danish Data Protection Agency has approved all retrieval of information. Because the present study utilises already existing and approved databases, no approval from our Local Ethics Committee was needed according to Danish legislation. Statistical significance level was defined by 95% confidence intervals. StataTM 8.0© 1984–2003 Statistics/Data Analysis package (Stata Corporation, 4905 Lakeway Drive, College Station, TX 77845, USA) was used for statistical analysis of data.
|
3. Results |
|---|
|
TopAbstract1. Introduction2. Methods3. Results4. DiscussionReferences |
|---|
|
|
3.2.2. Age at implantation
Fig. 1C shows survival time estimates of PM-patients stratified by patient age at implantation of 1st system (below or above the median, 1487 days). Survival times are shown in Table 1. Thirty-seven percent of patients below median age were alive at end of follow-up (13.4 years) compared to 80% of patients above median age (follow-up 20.3 years). Patients below median age had a significantly shorter survival time (RR-a=0.22, 95% CI: 0.06–0.76); confounding by other potential prognostic factors (RACHS-1 score, historical time of 1st implantation and lead type) was negligible.
3.2.3. Age at surgery
Fig. 1D shows survival time estimates of PM-patients stratified by patient age at surgery (below or above the median, 488 days). Survival times are shown in Table 1. Ninety-four percent of PM-patients above median age were alive at study termination. Patients below median age had a significantly shorter survival time (RR-a=0.06, 95% CI: 0.01–0.36); confounding by other potential prognostic factors (RACHS-1 score, historical time of 1st implantation and lead type) was negligible.
3.2.4. Lead type
Fig. 1E shows survival time estimates of PM-patients stratified by lead type (epicardial vs. transvenous). Survival times are shown in Table 1. Patients with epicardial leads had a significantly shorter survival time (RR-a=0.18, 95% CI: 0.04–0.87); there was possible confounding by RACHS category and age at surgery (see Table 1). Fig. 1F shows survival time estimates of PM-patients aged below median (488 days) at operation stratified by lead type (epicardial vs. transvenous). Again patients with epicardial leads had a significantly shorter survival time (RR-a=0.06, 95% CI: 0.01–0.36); confounding by other potential prognostic factors (RACHS-1 score) was negligible.
3.3. Event-free survival of pacemaker systems
Sixty-one failures of PM systems or deaths of patients occurred in the overall study population while 42 failures or deaths occurred in the 61 1st systems. Causes (overall study population) included: elective generator change (n=14, two of these were removals because of heart transplantation), haemodynamic upgrade (n=5), generator malfunction (n=2), generator protrusion (n=1), infection (n=5), battery failure (n=5), exit block (n=7), lead dislodgement (n=4), lead fracture (n=3), and death (n=15). Nineteen patients retained their 1st system at study termination, 29 patients had 
2 systems and 11 patients had 3 systems. Since implantation of numerous systems implies shorter longevity due to the study design (fixed follow-up time with progressive censoring), it was decided only to study characteristics of the first two implantations per patient. Longevity was assessed by means of estimates of event-free survival of systems, where events were defined as any cause of explantation (failure or elective) or death. Fig. 2
A shows event-free survival time estimates of 1st vs. 2nd system. Survival times for 75, 50 and 25% survival proportions were 1.5, 4.9 and 8.1 years and 1.0, 6.0 and 7.4 for 1st and 2nd systems, respectively. There was no difference between values (RR=0.85, 95% CI: 0.48–1.51).
|
|
3.4.2. Age at implantation
Fig. 2C shows event-free survival time estimates of 1st systems stratified by patient age at implantation (below or above median). Survival times are shown in Table 2. Patients below median age had a significantly shorter PM-longevity (RR-a=0.32, 95% CI: 0.16–0.63); confounding by other potential prognostic factors (RACHS-1 score) was negligible.
3.4.3. Age at surgery
Fig. 2D shows event-free survival time estimates of 1st systems stratified by patient age at surgery (below or above median). Survival times are shown in Table 2. Patients below median age had a significantly shorter PM-longevity (RR-a=0.38, 95% CI: 0.19–0.77); confounding by other potential prognostic factors (RACHS-1 score and lead type) was negligible.
3.4.4. Lead type
Fig. 2E shows event-free survival time estimates of 1st systems stratified by type of lead (epicardial (all) vs. transvenous). Survival times are shown in Table 2. Patients with epicardial leads had a significantly shorter PM-longevity (RR-a=0.40, 95% CI: 0.18–0.89); confounding by other potential prognostic factors (RACHS-1 score, age at 1st implantation and historical time of 1st implantation) was negligible. Fig. 2F shows event-free survival time estimates of 1st transvenous systems stratified according to patient age at implantation (below or above median). Again, patients below median age had a significantly shorter PM-longevity (RR-a=0.35, 95% CI: 0.16–0.75) and confounding by other potential prognostic factors (RACHS-1 score) was negligible.
3.5. Time interval from surgery to pacemaker implantation and choice of system
Fig. 3
shows the distribution of pacemaker implantations vs. interval from operation to 1st implantation (OI). Median OI was 19 (range 2–4842) days; however, the distribution seemed biphasic. We therefore sought to investigate potential and identifiable predictors of early and late implantation by logistic regression analysis using the median OI as cut-off value. RACHS-1 score of 1–2 (RR=4.72, 95% CI: 1.11–20.00), female sex (RR=0.13, 95% CI: 0.03–0.55) and operation before 1995 (RR=0.15, 95% CI: 0.04–0.62) were statistically significant predictors for late pacemaker implantation while patient age at operation (RR=0.36, 95% CI: 0.10–1.30) was not.
|
|
|
|
|
4. Discussion |
|---|
|
TopAbstract1. Introduction2. Methods3. Results4. DiscussionReferences |
|---|
Although post-surgical heart block after correction of congenital heart defects was a significant cause of mortality and morbidity in the past [1–3], refined surgical methods based on increased knowledge of the anatomy of the cardiac conduction system as well as the progressive development of smaller practically feasible pacemakers soon made promise for ‘the elimination of heart block as a complication of intracardiac repair of congenital cardiac defects’ [4]. However, studies have shown that pacemaker implantation in paediatric patients is in itself associated with significant morbidity due to pacemaker related complications and/or change of systems [4,9]. This study is the first to identify a broad range of independent predictors of mortality as well as pacemaker failure in patients with pacemakers due to surgically induced AV-block or sinus node dysfunction.
The RACHS-1 score has been proposed as a new method of risk adjustment for congenital heart surgery [15]. It has been shown to predict length of hospital stay as well as mortality in paediatric surgical patients [16], and the present finding may substantiate the further applicability of this classification system also after PM implantation. Age is a well-described predictor of mortality following surgery for congenital heart defects [17]. The phenomenon may potentially be explained by a vulnerability of younger infants to surgery and may furthermore be tentatively understood in light of the fact that emergent and complex cases present earlier. It is worth to emphasise, however, that complexity of preoperative diagnosis as well as RACHS-1 score was divided evenly between age groups, thus underscoring age as an independent predictor. Interestingly, the presence of epicardial lead did not appear to predict mortality before adjustment for RACHS-1 score and age below median at surgery. However, stratified analysis of mortality related to type of lead was carried out for patients below median age at operation and here presence of epicardial leads was shown to be significantly associated with raised mortality (RR=0.06) in this age group.
The observed overall mortality of patients was considerably greater than that of a background western paediatric population (www.danmarksstatistik.dk, November 2004). Whether this is due to the underlying congenital heart defect alone or augmented by having a pacemaker cannot be elucidated by the present data since this study does not comprise a control group of patients with surgically corrected congenital cardiac defects but without postoperative conduction/rhythm disturbances.
4.2. Prognostic factors, pacemaker systems
RACHS-1 score of 3–4, age below median at implantation, age below median at surgery and presence of epicardial leads were all shown to be significant predictors of early system failure, while system type (DD- and AA- vs. VV-), sex, historical time of surgery as well as implantation and postoperative time interval before implantation had no predicative value.
Longevity assessments of the 1st system have ranged from less than 40 up to 76% at 5 years in previous comparable studies [4,9,8]. The present results do not differ significantly from this; at 5 years, approximately 50% of our patients still have their 1st system. Furthermore, data does not suggest any difference in longevity of the first two systems. Due to low number of events and since implantation of numerous systems would tentatively imply shorter observed longevity due to the study design (fixed follow-up time with progressive censoring), we cannot perform a similar survival analysis of 3rd and 4th systems. One study [4] showed that first reoperation was a risk factor for failure of subsequent systems, but the study did not include analysis of longevity of these.
Our study is the first to show a correlation between high preoperative RACHS-1 score and failure of pacemaker systems. At 1.9 years after implantation of the 1st system, 50% of patients in RACHS-1 categories 3–4 had their pacemakers explanted due to either failure or death. This information may be a significant aid in selecting proper follow-up intervals as well as therapy in general (e.g. choice of lead type) for these patients at high risk of repetitive procedures. Low age at surgery as well as at implantation is like previously observed for mortality also a significant predictor of failure of 1st pacemaker system. This is especially interesting since one study [8] showed no increase in failure after implementation of transvenous pacemaker technology in small children (at a median follow-up time of 3.5 years). This study proposed several a priori inherent risk factors in young children: limited venous access, high thrombosis- and infection rates as well as high pulse rates with faster battery depletion. Our finding warrants a reconsideration of these tentative risk factors; however, due to a relatively high heterogeneity and low number of individual events we were not able to compare any of these outcomes between age groups.
A number of studies have shown the overall superiority of transvenous vs. non-steroid eluting (conventional-) epicardial leads in children [18–20]. In our study, the presence of epicardial leads (of which none were steroid eluting) was found to significantly predict failure of 1st system. Lead type was equally distributed between age groups as well as RACHS-1 categories (Table 4). Although pacemaker characteristics were not included in the present study it is plausible that the higher pacing and sensing thresholds of epicardial leads cause a higher failure rate. These findings substantiate the current practice at our institution; in general, we try to use transvenous systems when these are practically feasible; we use no definite lower age limit of this. However, recent publications have substantiated the use of steroid eluting epicardial leads, especially in small children with limited venous access and potentially many years of pacemaker dependency [21,22].
Also, interestingly, our study showed that age below median (1487 days) at 1st implantation of transvenous systems (excluding the epicardial leads) significantly predicts failure of these. Epicardial lead type showed a trend towards predicting system failure in age groups above and below median (Table 2). However, this did not reach statistical significance, possibly due to low numbers.
A striking but logical feature of figures 5–10 is that all systems, regardless of predictors fail at approximately 12 years; ‘the curves meet’, and the logical explanation of this is of course that any pacemaker eventually has to be explanted.
4.3. Time interval from surgery to implantation
The time interval from surgery to pacemaker implantation showed a biphasic pattern.
RACHS-1 score of 1–2, operation before 1995 and female sex were statistically significant predictors for late pacemaker implantation while patient age at operation was not. The fact that the two former had a predicative value is presumably fairly straightforward. High RACHS-1 score would favour early censoring due to death (as shown by others [15,16] and indicated by the present study) and furthermore, otherwise benign rhythm disturbances would possibly be focus of pacemaker therapy in face of declining cardiac function as suspected in high RACHS-1 score patients. Patients receiving surgery in the late period who will later develop late rhythm disturbances would be likely to do so after expiry of the follow-up period due to time bias. We have no immediately intelligible explanation to the fact that female sex predicts late pacemaker implantation after congenital heart surgery. However, we note the fact that sex was equally distributed between age groups, historical periods and RACHS-1 score, thus stating its role as an independent predictor. The bulk of early implantations were in the immediate postoperative period, while late implantations were more scattered with six occurring in an intermediate period from postoperative day 20 to 100. All patients in the early period were reported to have AV-block immediately after surgery; approximately half of these patients received a PM between postoperative day 10 and 20. This is interesting since a recent study showed that 95% of patients with postoperative AV-block who regain conduction do so by postoperative day 9 [23]. Perceivably, there would be minimal benefit from delaying pacemaker implantation beyond this period. All six ‘intermediate’ patients were discharged with sinus rhythm but still developed pacemaker need from AV-block (n=5) or sinus node dysfunction (n=1). The present study does not permit analysis of predictors of intermediate term pacemaker implantation in patients discharged with sinus rhythm. Therefore this finding merely warrants attention from attending paediatricians and general practitioners as well as the parents who are the ones to observe the children in this period.
In the late period (after postoperative day 100), sinus node dysfunction is the predominant cause of pacemaker implantation (12 out of 18) and furthermore there is an apparent over-representation of patients operated upon by surgical methods no longer in use. The Circumclusion procedure for atrial septal defects, devised by Søndergaard in 1955 [24] yielded four while Mustard procedures yielded seven late events of sinus node dysfunction requiring pacemaker therapy. While the latter phenomenon has been described by others [12], it is novel, albeit understandable, that Circumclusion, which involved blunt dissection of the septal tissue between the two atria below the defect could provide a structural substrate for supraventricular rhythm disturbances.
4.4. Type of pacemaker system
Not surprisingly, the present data shows a significant upgrade of systems to DDD- and AA-mode after change/failure of 1st system. This may in part be due to the general development of pacemaker technology; however, in light of the fact that DDD-mode was preferably used in older children it falls nicely in line with our institution's intention to refine pacemaker therapy to individual patient needs. Moreover, data did not support a significant change of strategy through historical time.
4.5. Limitations
We must emphasise the fact that although all information has been gathered in a prospective manner, analysis of historical data is associated with a number of sources of error. Firstly, it was only possible to adjust for possible confounding by registered variables. This study did not include significant additional information of co-morbidity, preoperative health status (except for RACHS-1 score), duration of surgery, length of ICU- and hospital stay, ECG-at-discharge, long-term postoperative health status or pacemaker system characteristics (in terms of pacing and sensing characteristics) since this information could not be sampled in a controlled manner. Secondly, in most analyses the entered dependent and independent variables were artificially dichotomised using the median value due to low numbers, and it is therefore impossible for the authors to conclude beyond this, i.e. describe linear or other exposure–effect relations as well as any residual confounding.
4.6. Conclusion
We found that RACHS-1 score of 3–4, lower age at surgery (less than the median: 488 days), lower age at implantation of 1st pacemaker system (less than the median: 1487 days) and the presence of epicardial electrodes all significantly predict patient mortality as well as failure of 1st pacemaker systems after pacemaker implantation. Somewhat surprisingly, 1st and 2nd systems had similar event-free survival rates. These findings may have important implications for preoperative risk stratification of patients and counselling of parents to patients with congenital heart disease.
|
Acknowledgments |
|---|
|
References |
|---|
|
TopAbstract1. Introduction2. Methods3. Results4. DiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  H. Olgun, T. Karagoz, A. Celiker, and N. Ceviz Patient- and lead-related factors affecting lead fracture in children with transvenous permanent pacemaker Europace, July 1, 2008; 10(7): 844 - 847. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P.A. Zartner, R.P. Handke, A.M. Brecher, and M.B.E. Schneider Integrated home monitoring predicts lead failure in a pacemaker dependent 4-year-old girl Europace, March 1, 2007; 9(3): 192 - 193. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
