J.W.M. van Eck; N.M. van Hemel; P. Zuithof; J.P.M. van Asseldonk; T.L.H.M. Voskuil; D.E. Grobbee; K.G.M. Moons

Europace.  2007;9(10):884-889.  ©2007 Oxford University Press

Abstract and Introduction

Abstract

Aims: Despite an annual rise in the numbers of patients receiving their first pacemaker (PM), the risks of the implantation procedure remains unclear. The purpose of this prospective study is to estimate the incidence of in-hospital events after first PM implantation and to determine the predictors of these events.
Methods and results: Patients with conventional pacing diagnosis were included in the Dutch multicentre FOLLOWPACE PM registry that prospectively documented patients' prognosis and quality of life, and PM events after first implantation. From these registry characteristics, implantation data and in-hospital findings were analysed as potential predictors for events in a sample of 1198 patients. In 111 patients studied, at least one serious in-hospital event occurred (incidence 10.1%, 95%CI: 8.9-12.3). Six variables, i.e. a lower body mass index, presence of heart failure in medical history, main indication for implantation, vena subclavia use for venous access, active atrial lead fixation, and the implantation of a dual chamber system, were found to be independent predictors of events after first PM implantation. The overall multivariable model yielded an ROC area of 0.65 (95%CI: 0.60-0.70).
Conclusion: This large prospective multicentre study identified six variables as independent predictors for serious in-hospital events after first implantation. This may assist the implanting cardiologist and surgeon to identify patients at higher risk, during and immediately after PM implantation.

Introduction

Although pacemakers (PMs) are implanted on a large and growing scale, current data on the incidence of early (in-hospital) events after PM implantation are inconsistent and have been obtained from studies that have only limited relevance for current practice.^[1-10] Recent large prospective multicentre studies on all types of events after PM implantation are lacking. Furthermore, there are even less studies that investigated independent predictors of in-hospital events after PM implantation.^[7-9,11,12] Knowledge on independent predictors of in-hospital events improves the identification of patients at a higher risk and facilitates making decisions on the use of a specific type of PM system and the conditions of the implantation procedure.

Given the ongoing increase in number of PM implantations,^[13,14] and the limited data on the incidence of clinically relevant (in-hospital) events, after first PM implantation and on the independent predictors of these events, we conducted a large prospective multicentre study on the prognosis of PM patients -- the FOLLOWPACE study.^[15] The first aim was to quantify the occurrence of clinically relevant in-hospital events related to a first PM implantation. Subsequently, we determined which patient-, PM-, or implantation- related factors are independent predictors of these in-hospital events.

Patients and Methods

Patients

FOLLOWPACE is a prospective multicentre longitudinal cohort study, executed in 23 (of the 104) PM centres in the Netherlands covering ~60% of the annual implantations. The design of the FOLLOWPACE study has been published previously.^[15] In brief, any consecutive patient aged 18 years or over, receiving a first PM for a conventional reason^[16] for chronic pacing therapy in one of the participating centres, is a candidate for the study. Patients are not eligible if they refused to sign an informed consent, were taking any investigational drug, or have a non-approved or investigational PM implanted. In addition, patients having diseases that are likely to cause death or significant morbidity during the study period, such as neoplasia and immune, infectious, or degenerative diseases, are excluded. In general, the aims of FOLLOWPACE are to assess the incidence of events (in-hospital and at 1 year) and quality of life after implantation, to quantify which factors measured at implantation (baseline factors) are the predictors of both outcomes, and to determine whether subsequent follow-up measurements have added prognostic value. The motive of FOLLOWPACE is to reduce the time and energy of patients and health care providers (i.e. cardiologists and PM technicians) while preserving the safety of pacing therapy and the prognosis of the patient. Moreover, FOLLOWPACE intends to contribute to a more evidence-based and efficient follow-up strategy for patients receiving a PM. The protocol for this study was approved by the Ethical Commission of the University Medical Center (UMC) Utrecht.

From each patient, all data on the potential predictors and study outcomes were measured according to a systematic protocol that was identical for each participating centre. Study data were directly entered online in an electronic case report form that was specifically designed for the study. We randomly validated the data of 75 patients by checking the electronically entered data with the original patient records. No major discrepancies were found. Only two inconsistencies in the baseline characteristics (i.e. missing value on cardiovascular history) were detected. No study outcomes were missing.

Enrolment for the FOLLOWPACE study is still continuing. The present analysis is based on the first 1198 patients with conventional indications for pacing.

Outcome

The outcome of the present analysis was the incidence of (one or more) clinically relevant events occurring during PM implantation or in the period until discharge from the hospital. Only clinical relevant events were considered including the occurrence of: ventricular arrhythmias; traumatic events (such as pneumothorax, haemothorax, pneumohaemothorax, or cardiac atrial/ventricular wall perforation); lead-related events (such as fractures, dislocations, disconnections from the PM, insulation, and pacing threshold problems); wound problems of the PM pocket (such as all visible haematomas, skin dehiscence, or infection); device-related events (such as device malfunctions or manufacturers recalls); acute myocardial infarction; endocarditis and sepsis; and in-hospital (all cause) mortality.

Predictors

Based on previous studies,^{[1,2,4-7,11,12,17]} we a priori selected 16 candidate predictors of in-hospital events ( Table 1 ). These included patient characteristics (such as age, gender, indication for pacing, body mass index (BMI), cardiac history, diabetes, and use of medication), and implantation- and PM-related characteristics (such as venous access, atrial or ventricular insertion site, number of leads and method of fixation, and programmed pacing mode at discharge). Programmed pacing mode at discharge was defined as last pacing mode programmed before hospital discharge.

Data Analysis

We first estimated the cumulative incidence of the outcome (occurrence of any clinically relevant in-hospital event) with corresponding 95% confidence interval (95%CI). We then estimated the univariable association between each candidate predictor and the outcome. Subsequently, we included all candidate predictors in an overall multivariable logistic regression model. Continuous predictors were analysed as linear terms, as there were no indications of non-linearity based on cubic spline analysis.^[18] This overall multivariable model was then reduced by deleting (one by one) the predictors with a P-value > 0.157 (the Akaike Information criterion) based on the log-likelihood ratio test to determine the independently contributing predictors. Where a selection of the important variables for a (prognostic) prediction model is required, a higher P-value than 0.05 is generally recommended.^[18,19] The predictive accuracy of the initial and reduced model was estimated by their calibration (reliability or goodness of fit) and discrimination. The calibration was tested using the Hosmer and Lemshow test. The models' ability to discriminate between patients with and without an event was estimated by the area under the receiver operating characteristic curve (ROC area). Prediction models commonly show too optimistic performance -- due to too extreme regression coefficients [odds ratios (OR)] of the predictors -- in the data set from which they are developed, so called over-fitting. This is certainly the case when many predictors relative to the number of events are analysed.^[18,19] To address this issue as much as possible with the available data, we used bootstrapping techniques (n = 100 bootstrap samples) -- repeating the entire modelling process including the backwards variable selection in each sample -- to internally validate the final model and to obtain insight in the optimism or over-fitting of the final model. The bootstrapping yields a shrunk ROC area (i.e. adjusted for optimism) as well as a shrinkage factor for the regression coefficients (log-odds ratios) of the selected predictors.^[18] The estimated regression coefficients of the selected predictors were multiplied by this shrinkage factor to adjust the found associations of the predictors for over-fitting and to better reflect their true association. Also, the adjusted model's performance obtained with the bootstrapping better reflects its performance that is expected in future PM patients. All analyses were performed using S-Plus Version 6.2.1 (Insightful Corp. Seattle, WA, USA).

Results

The majority of patients were males and the mean age at implantation was 73.7 years (second column, Table 1 ). Atrioventricular conduction disturbances were the main reason for PM implantation (40.4%). The venous access was in most cases through the subclavian vein (90.3%). Atrial leads were most likely to be fixated by an active mechanism (54.0%), and active fixation procedure was performed in 25.5% of all right ventricular lead insertions. In 73.5% of the cases, a dual chamber system was inserted. Hospitalization duration had a median of 3 days (25/75 percentiles 1/7 days).

In 111 patients at least one in-hospital event occurred ( Table 2 ), yielding a cumulative incidence of 10.1% (95%CI: 8.9-12.3). In 8 patients two events, mostly a combination of a traumatic complication and a haematoma occurred, and in one patient three events were reported (a haematoma followed by an infection and a fatal acute myocardial infarction). Eleven patients (0.9%) died during or shortly after the implantation procedure. In four cases, this was due to an acute myocardial infarction occurring the day after the implantation; in four other cases, a stroke the day after the implantation was the cause of death; and one patient developed aspiration pneumonia. In two other fatalities, the cause of death could not be determined. In four cases (0.5%), the implantation procedure was aborted due to malfunctioning X-ray devices (one patient) or difficult implantation procedures (three patients).

In the univariable analysis, a lower BMI, presence of heart failure, presence of other cardiovascular diseases, indication for implantation, use of the vena subclavia for venous access and the implantation of a dual chamber system, were associated (P-value < 0.157) with a higher incidence of in-hospital events (column 3-5, Table 1 ).

The overall multivariable model with all 16 predictors yielded a ROC area of 0.70 (95%CI: 0.64-0.75) and a non-significant Hosmer and Lemeshow test (P-value = 0.34). Only six predictors appeared to be independently related to the outcome: BMI, presence of heart failure in medical history, main indication for implantation, vena subclavia for venous access, active atrial lead fixation, and the implantation of a dual chamber system ( Table 3 ).

The reduced model yielded a ROC area of 0.65 (95%CI: 0.60-0.70) after correction for over-fitting.

Discussion

The incidence of clinically relevant in-hospital events after first PM implantation in this large unselected sample of patients was 10.1% (95%CI: 8.9-12.3). Previous studies on the incidence of in-hospital events found an incidence varying from 4 to 7%.^[1-10] However, these results have been published more than a decade ago,^[1-3] were retrospective,^[3,6-10] were single-centre studies,^[1-3,6-10] were executed in a specific patient population such as elderly people,^[4,5] or focused on a particular event such as cardiac perforation^[7] or pocket haematoma.^[9] To put our results in a comparative setting, a registry study of the Mayo Clinics observed a cardiac perforation frequency of 1.2% of the 4280 patients receiving first PM implant.^[7] Extrapolating this incidence to our data would imply 14 cases of cardiac perforation. However, we only found one such event -- a ventricular perforation due to lead insertion. This low frequency of myocardial perforation cannot be ascribed to a less frequent application of active fixation leads in the right ventricular wall of the FOLLOWPACE population when compared with current literature.^[7] Another study showed an incidence of 4.9% of pocket haematoma after PM or implantable cardioverter-defibrillator, leading to prolonged hospitalization in 2.0% of the patients.^[9] Extrapolation would imply 58 (4.8%) cases, which is more than two-fold of our finding of 1.9%. If in our data only the indisputable complications are counted (10 in-procedural arrhythmias, 28 traumatic events, 32 serious lead problems, and 33 wound complications, all clearly related to the implantation procedure), the actual incidence of complications or major events is 8.6%, which is still higher than generally reported (4-7%).^[1-9] The most likely cause of our observed higher incidence is the closer documentation of all events in this carefully designed and planned prospective study.

Of all 16 variables studied, six were independent predictors for the occurrence of in-hospital events, i.e. lower BMI, presence of heart failure in medical history, main indication for implantation, vena subclavia for venous access, active atrial lead fixation, and the implantation of dual chamber system. To appreciate these results, several issues need to be discussed. Previously reported predictors of in-hospital events are history of thrombosis, atrial fibrillation, use of steroid therapy, absence or presence of anticoagulant treatment, lower BMI, presence of bradytachy syndrome, the use of temporary pacemaker, the method of the subclavian vein access, the active atrial lead fixation technique, the use of helical screw leads, and the experience of the implantation team.^{[5-7,10-12,20,21]} Most previous studies focused on the prediction of a particular event, such as cardiac perforation,^[7] atrial fibrillation,^[11] pocket-related complications,^[22] or pericarditis.^[20] Second, most did not apply multivariable analysis to determine which factors are independent predictors of in-hospital events after PM implantation, undermining the relevancy of these particular study outcomes.^{[5-7,10-12,20,21]}

This study confirms that a lower BMI is a predictor of in-hospital events after first PM implantation in accordance to previous studies; however, one study demonstrated that both a low BMI (< 20) as well as a high BMI (> 30) were predictors, although not statistically independent.^[5,7,10] We observed that per increase of 1 unit in BMI, the relative risk of in-hospital events decreases by 9% (95%CI: 1-11). The presence of heart failure in patients' history leads to a two-fold increase in the risk of developing an in-hospital event. This may be related to the more extensive cardiac history, the impaired physical condition of these patients, and the cardiac remodelling in patients with chronic heart failure. The risk of in-hospital events furthermore depends on the main indication for implantation with the lowest risk for patients suffering from atrial fibrillation with slow ventricular response.

Three implantation procedure related characteristics were independent predictors. The use of the vena subclavia for venous access provokes a higher risk on events when compared with the use of the vena cephalica. Although more complicated, more time-consuming, and less applied, the surgical preparation of the vena cephalica appears less aggressive than the puncture of the vena subclavia for the access of one or two leads. This issue was already discussed in several publications emphasising that the lowest complication rate was observed with the cephalic vein cut-down technique. ^{[2,3,6,12,21]} In this study, passive fixation of the atrial lead is correlated with a higher risk on events after initial PM implantation. Active lead fixation may reduce the risk of early lead dislocation^[23] and repeat implantation procedure but may be associated with a higher risk of atrial and ventricular wall perforation.^[24-26] In contrast, our data suggest that active fixation of the atrial lead is correlated with a favourable outcome. However, regarding atrial lead dislocation, our data are not homogenous as in 16 cases with atrial lead dislocation -- nine had passive and seven had active fixation mechanism. In all 6 cases, where the ventricular lead dislocated from the initial pacing site, the leads had a passive fixation mechanism. In our dataset in 98.6% of the cases, where a dual chamber system was implanted, the programmed pacing mode at discharge was DDD pacing. Eberhardt and Wiegand stated that differences in complication rates between dual-chamber and single-chamber devices depend on operation time. The duration of the implantation procedure correlates to low or medium level of experience of the operator. ^[27,28] Our data confirm their results that the standard right ventricular pacing mode is the safest procedure in terms of preventing in-hospital events. This is in contrast to the results of a previously conducted prospective single-centre study (n = 1088) claiming that there is no risk difference between the implantation of a dual- and single-chamber system.^[1] Our data could not disclose that oral anticoagulation treatment during implantation nor the presence of bradytachycardias were independent risk factors for in-hospital events. The effect of oral anticoagulant treatment was not found in the analysis (OR: 1.22, 95%CI: 0.81-1.84). Furthermore, a recent prospective study of 47 patients anticoagulated with warfarin, undergoing PM and ICD implantation, showed no major bleeding or haematomas. ^[29] However, we did not measure adverse effects of heparin and platelet aggregation inhibition therapy as potential risk factors for developing in-hospital events. Finally, the contribution of the experience of implanters is still a matter of debate. According to Parsonnet et al.,^[12] when physicians perform less than 12 implantations annually this is correlated with a higher risk on complications. This registry reports complication rates of centres with a relatively large implantation experience (20% of implanting centres in the Netherlands representing 60% of implants). However in the Netherlands, in every institution, the minimum number of PM implantations per implanter is more than 20 per year, thus reducing the difference of experience as a predictor.

Methodological Issues

First, the number of events (n = 111) and the number of studied predictors (n = 16) can be a cause of concern. The rule of thumb is that, for every specific predictor, at least 10 events should be recorded. ^[30,31] This was not the case in our study; where we found a ratio of 1:8. This may have resulted in less stable estimates of the independent associations (ORs) of the predictors in the final model as well as of the model's discriminative ability. To address this issue as good as possible, we used bootstrapping procedures to carefully validate our model and adjust the associations and ROC area for optimism. However, despite our bootstrapping, we recommend that our predictors should be validated in new patients.

Second, although our dataset was fairly large, we explicitly chose for a composite endpoint including all clinically relevant events. The number of the different types of events was simply too small to draw statistically relevant conclusions on predictors of the component events. Investigation of predictors of a specific event would obviously require a much larger dataset. If, for example, we would have made it our goal to investigate the predictors of in-hospital haematomas after implantation, we would have needed a dataset of over 8000 patients to fulfil the rule of thumb that for every predictor, 10 events should be recorded.

Third, our aim was only to establish the important prognostic predictors of clinically relevant in-hospital events, and by no means causal associations. We therefore never used the term 'complication', as this implies a causal relation. The search on causal relationships between variables and the outcome was never our interest, and would require a very different type of study, notably a different analysis and presentation of results.

Fourth, a longer follow-up period (for example, including the first month after hospital discharge) is also indicated for future research to improve the relevance of our predictors.

Finally, a ROC area of 0.70 before bootstrapping (95%CI: 0.64-0.75) and 0.65 (95%CI: 0.60-0.75) after correction for over-fitting may seem low. But we note that our study concerns a prognostic rather than a diagnostic setting: in the latter, indeed higher ROC areas are generally found.

Conclusion and Clinical Implications

This large prospective multicentre study shows an incidence of in-hospital events after first PM implantation of 10.1% (95%CI: 8.9-12.3). We identified six independent predictors consisting of three patient characteristics and three implantation- and PM-related predictors. These predictors can support implanting cardiologists, surgeons, and assisting technicians to identify the patients at higher risk of in-hospital events.

Aggarwal RK, Connelly DT, Ray SG, Charles RG. Early complications of permanent pacemaker implantation: no difference between dual and single chamber systems. Br Heart J (1995) 73:571-5. Chauhan A, Grace AA, Newell SA. Early complications after dual chamber versus single chamber pacemaker implantation. PACE (1994) 17:2012-15. Hargreaves MR, Doulalas A, Ormerod OJM. Early complications following dual chamber pacemaker implantation: 10-year experience of a regional pacing centre. Eur JCPE (1995) 5:133-8. Lamas GA, Orav EJ, Stambler BS, Ellenbogen KA, Sgarbossa EB, Huang SKS, et al. Quality of life and clinical outcomes in elderly patients treated with ventricular pacing as compared with dual-chamber pacing. N Engl J Med (1998) 338:1097-104. Link MS, Estes NA3, Griffin JJ, Wang PJ, Maloney JD, Kirchhoffer JB, et al. Complications of dual chamber pacemaker implantation in the elderly. J Interv Card Electrophysiol (1998) 2:175-79. Tobin K, Stewart J, Westveer D, Frumin H. Acute complications of permanent pacemaker implantation: their financial implication and relation to volume and operator experience. Am J Cardiol (2000) 85:774-6. Mahapatra S, Bybee KA, Bunch J, Espinosa RE, Sinak LJ, McGoon MD, et al. Incidence and predictors of cardiac perforation after permanent pacemaker placement. Heart Rhythm (2005) 2:907-11. Marquie C, De Geeter G, Klug D, Kouakam F, Brigadeau F, Jabourek O, et al. Post-operative use of heparin increases morbidity of pacemaker implantation. Europace (2006) 8:283-7. Wiegand UKH, Lejeune D, Boguschewski F, Bonnemeier H, Eberhardt F, Schunkert H, et al. Pocket hematoma after pacemaker of implantable cardioverter defribrillator surgery. Influence of patient morbidity, operation strategy and the perioperative antiplatelet/anticoagulation therapy. Chest (2004) 126:1177-86. Kiviniemi MS, Pirnes MA, Eranen HJ, Kettunen RV, Hartikainen JE. Complications related to permanent pacemaker therapy. Pacing Clin Electrophysiol (1999) 22:711-20. Gillis AM, Morck M. Atrial fibrillation after DDDR pacemaker implantation. J Cardiovasc Electrophysiol (2002) 13:542-7. Parsonnet V, Bernstein AD, Lindsay B. Pacemaker-implantation complication rates: an analysis of some contributing factors. JACC (1989) 13:917-21. Ector H, Rickards AF, Kappenberger L, Linde C, Vardas P, Oto A, et al. The world survey of cardiac pacing and implantable cardioverter defibrillators: calendar year 1997 -- Europe. Pacing Clin Electrophysiol (2001) 24:863-8. Birnie D, Williams K, Guo A, Mielniczuk L, Davis D, Lemery R, et al. Reasons for escalating pacemaker implants. Am J Cardiol (2006) 98:93-7. van Eck JWM, van Hemel NM, Grobbee DE, Buskens E, Moons KGM. FOLLOWPACE study: a prospective study on the cost-effectiveness of routine follow-up visits in patients with a pacemaker. Europace (2006) 8:60-4. Gregoratos G, Abrams J, Epstein AE, Freedman RA, Hayes DL, Hlatky MA, et al. ACC/AHA/NASPE 2002 Guideline update for implantation of cardiac pacemakers and antiarrhythmia devices: summary article. J Cardiovasc Electrophysiol (2002) 13:1183-99. Mugica J, Ritter P, Lazarus B, Bakdach M, Henry L, Podeur H, et al. Evolutions of early complications after ventricular single chamber pacemaker implants in a specialised centre: 9483 patients over 24 years. Eur JCPE (1993) 3:21-6. Harrell FE, Lee KL, Mark DB. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med (1996) 15:361-87. Chatfield C. Model uncertainty, data mining, and statistical inference. J R Stat Soc (1995) 158:419-66. Sivakumaran S, Irwin ME, Gulamhusein SS, Senaratne MPJ. Postpacemaker implant pericarditis: incidence and outcomes with active-fixation leads. PACE (2002) 25:833-7. Furman S. Subclavian puncture for pacemaker lead placement. Ann Thorac Surg (1986) 41:438. Milic DJ, Perisic ZD, Zivic SS, Stanojkovic ZA, Stojkovic AM, Karanovic ND, et al. Prevention of pocket related complications with fibrin sealant in patients undergoing pacemaker implantation who are receiving anticoagulant treatment. Europace (2005) 7:374-9. Lemke B, Holtmann BJ, Selbach H, Barmeyer J. The atrial pacemaker: retrospective analysis of complications and life expectancy in patients with sinus node dysfunction. Int J Cardiol (1989) 22:185-93. Ellenbogen KA, Wood MA, Shepard RK. Delayed complications following pacemaker implantation. PACE (2002) 25:1155-8. Ellenbogen KA, Hellkamp A, Wilkoff BL, Camunas JL, Love JC, Hadijs TA, et al. Complications arising after implantation of DDD pacemakers the MOST experience. Am J Cardiol (2003) 92:740-1. Oginosawa Y, Abe H, Nakashima Y. Right pneumothorax resulting from an endocardial screw-in atrial lead in an implantable cardioverter defibrillator system. PACE (2002) 25:1278-9. Eberhardt F, Bode F, Bonnemeier H, Boguschewski F, Schlei M, Peters W, et al. Long term complications in single and dual chamber pacing are influenced by surgical experience and patient morbidity. Heart (2005) 91:500-6. Wiegand UKH, Bode F, Bonnemeier H, Eberhardt F, Schlei M, Peters W. Long-term complication rates in ventricular, single lead VDD, and dual chamber pacing. PACE (2003) 26:1961-9. Al-Khadra AS. Implantation of pacemaker and implantable cardioverter defibrillators in orally anticoagulated patients. PACE (2003) 26(Pt. II):511-4. Concato J, Peduzzi P, Holford TR, Feinstein AR. Importance of events per independent variable in proportional hazards analysis. I. Background, goals, and general strategy. J Clin Epidemiol (1995) 48:1495-501. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol (1996) 49:1373-9.

Acknowledgements

The authors would like to thank all the patients included in this study and all cardiologists and pacemaker technicians of the following hospitals in the Netherlands; Bernhoven Hospital, Veghel; Amphia Hospital, Breda; Diaconessen Hospital, Meppel; Medical Center Alkmaar, Alkmaar; Hospital group Twente, Hengelo; Viecuri Medical Center, Venlo; Zaans Medical Center, Zaandam; St Antonius Hospital, Nieuwegein; Alysis Rijnstate Hospital, Arnhem; Vlietland Hospital, Schiedam; Deventer Hospital, Deventer; VU Medical Center, Amsterdam; Twenteborg Hospital, Almelo; Spaarne Hospital, Heemstede; Westfries Hospital, Hoorn; Atrium Medical Center, Heerlen; Rijnland Hospital, Leiderdorp; University Medical Center, Groningen; Maxima Medical Center, Veldhoven; Antonius Hospital, Sneek; Hospital De Tjongerschans, Heerenveen; Canissius Wilhelmina Hospital, Nijmegen; and Diaconessen Hospital, Leiden.

Funding Information

This study was financially supported by: Dutch College of Health Care (CVZ/VAZ grant number 01236); all Dutch pacemaker distributors and manufacturers; Netherlands Pacemaker Registry Foundation, Groningen Jacques H. de Jong Foundation; and Roger Crowson Foundation for Human Arrhythmias Studies.

Reprint Address

J.W.M. van Eck, Tel: +31 30 2506545; fax: +31 30 2505481. E-mail: j.w.m.vaneck-2@umcutrecht.nl .