Evolution of pacemaker technology has resulted in the

Transcription

1 Relationship Between Pacemaker Dependency and the Effect of Pacing Mode on Cardiovascular Outcomes Anthony S.L. Tang, MD; Robin S. Roberts, MTech; Charles Kerr, MD; Anne M. Gillis, MD; Martin S. Green, MD; Mario Talajic, MD; Salim Yusuf, MD; Hoshiar Abdollah, MB ChB; Michael Gent, DSc; Stuart J. Connolly, MD; and the Canadian Trial of Physiologic Pacing (CTOPP) Investigators Background A recently completed trial, the Canadian Trial of Physiological Pacing (CTOPP), showed that physiological pacing did not significantly reduce mortality, stroke, or heart failure hospitalization, but it did show that atrial fibrillation occurred less frequently in patients with physiological pacing. Many pacemaker patients experience only transient bradyarrhythmias with an adequate unpaced heart rate (UHR) and are not pacemaker-dependent. The purpose of the present analysis was to determine if pacemaker-dependent patients have an increased benefit from physiological pacing compared with non pacemaker-dependent patients. Methods and Results Of 2568 patients included in the CTOPP trial, 2244 patients had a pacemaker dependency test performed at the first follow-up visit. The yearly event rate of cardiovascular death or stroke steadily increased with decreasing UHR in the ventricular pacing group, but it remained constant in the physiological pacing group. When the patients were subdivided to UHR 60 bpm or 60 bpm, there was an interaction between pacing mode treatment and UHR subgroup. The Kaplan-Meier plot confirmed a physiological pacing advantage only in the UHR 60 bpm subgroup. This differential effect was also present for the outcomes of cardiovascular death and total mortality. Conclusions This study demonstrated that UHR at first follow-up has an important influence on how pacing mode selection affects cardiovascular death and total mortality. Pacemaker-dependent patients with low UHR will probably be paced frequently and will likely benefit from physiological pacing. In contrast, non pacemaker-dependent patients will likely be paced infrequently and may not benefit from physiological pacing. (Circulation. 2001;103: ) Key Words: pacemakers trials mortality Evolution of pacemaker technology has resulted in the widespread use of dual-chamber pacing worldwide. 1 This has been based on the assumption that atrioventricular sequential pacing optimizes cardiac hemodynamics, improves exercise tolerance, and reduces mortality and morbidity. 2 However, until recently, data from randomized trials to support the choice of a specific pacing modality for an individual have been lacking. We recently completed a large clinical trial, the Canadian Trial of Physiological Pacing (CTOPP), which randomized patients with symptomatic bradyarrhythmias to physiological pacing (dual chamber/atrial pacing) or ventricular pacing. In this study, physiological pacing did not significantly reduce mortality, stroke, or hospitalization for congestive heart failure, but it was associated with a modest risk reduction for the occurrence of atrial fibrillation when compared with ventricular pacing. 3 Many patients receiving a permanent pacemaker for symptomatic bradyarrhythmias experience only transient bradyarrhythmias and have an adequate unpaced heart rate (UHR) most of the time. These patients are not pacemaker-dependent and will likely be paced for only a small proportion of the time. In contrast, pacemaker-dependent patients (ie, those with a slow UHR) will likely be paced most or all of the time. We hypothesized that the choice of pacing modality in pacemaker-dependant patients might affect cardiovascular outcomes more significantly than in non pacemakerdependent patients in whom the pacemakers function in a backup role. The purpose of the present analysis was to determine if pacemaker-dependent patients have an increased benefit with physiological pacing compared with patients who are not pacemaker-dependent. Methods CTOPP was conducted in 32 Canadian centers and randomized 2568 patients (1474 patients to ventricular pacing and 1094 patients to physiological pacing). Patients were eligible if they were receiving their first pacemaker for symptomatic bradyarrhythmias and did not have chronic atrial fibrillation. After obtaining informed consent, patients were randomized within 48 hours of a scheduled pacemaker Received February 1, 2001; revision received March 23, 2001; accepted April 6, From the University of Ottawa Heart Institute, Ottawa, Ontario (A.S.L.T., M.S.G.); McMaster University, Hamilton, Ontario (R.S.R., S.YL, M.G., S.J.C.); the University of British Columbia, Vancouver, British Colombia (C.K.); the University of Calgary, Calgary, Alberta (A.M.G.); Montreal Heart Institute, Montreal, Province of Quebec (M.T.); and Queen s University, Kingston, Ontario (H.A.), Canada. Correspondence to Dr Anthony S.L. Tang, University of Ottawa Heart Institute, 40 Ruskin St., Ottawa, Ontario, Canada, K1Y4W7. atang@ottawaheart.ca 2001 American Heart Association, Inc. Circulation is available at

2 3082 Circulation June 26, 2001 TABLE 1. Clinical Parameters of Patients With Pacemaker-Dependency Test at First Follow-Up Ventricular Pacing (n 1302) Physiological Pacing (n 942) Mean age, y Male sex, % New York Heart Association class 2, % Pacing indication, % Sinoatrial node disease Atrioventricular node disease Both 8 9 Unknown 6 7 Rate-adaptive pacing, % Medical history, % Myocardial infarction Diabetes Hypertension Stroke or transient ischemic attack Paroxysmal atrial fibrillation Medication, % Anticoagulant Antiplatelet agent Antiarrhythmic drug implant. Of the patients randomized to ventricular pacing, 99.1% received a ventricular pacemaker and 0.7% received either an atrial or dual-chamber pacemaker; 75.6% received rate-adaptive devices. Of the patients randomized to physiological pacing, 88.1% received a dual-chamber pacemaker, 5.2% received an atrial pacemaker, 5.6% received a ventricular pacemaker (mostly due to failure of atrial lead implantation), and 0.9% received no pacemaker; 43% received rate-adaptive devices. Patients were assessed 2 to 8 months after pacemaker implantation and annually thereafter. At each visit, pacemaker function and the occurrence of outcome events were assessed. The primary outcome was the first occurrence of either cardiovascular death or stroke. The secondary outcomes were cardiovascular death, death from any cause, stroke or systemic emboli, documented atrial fibrillation lasting 15 minutes, and admission to hospital with congestive heart failure. An event adjudication committee reviewed any reported outcome event in a blinded manner. At the first follow-up visit (2 to 8 months), a pacemakerdependency test was performed. The pacemaker was temporarily TABLE 3. Treatment Effect of Primary End Point by UHR P CV Death or Stroke TABLE 2. UHR at First Follow-Up UHR, bpm Ventricular Pacing, n (%) Physiological Pacing, n (%) (21) 209 (22) (13) 171 (18) (18) 188 (20) (48) 374 (40) Subtotal (UHR known) 1302 (88) 942 (86) The distribution of UHR category between the 2 treatment groups is significantly different (P ). programmed to VVI mode at a low rate of 40 bpm. A stable ventricular rate was then recorded as the UHR. Data analysis was based on the intention-to-treat principle. Proportional hazard modeling was used to determine the interaction between pacing mode treatment effect and UHR subgroups. Logrank tests were performed to compare the risk of the outcome events between treatment groups, and Kaplan-Meier plots were created. All probability values are 2-sided. Results Pacemaker Dependency Pacemaker-dependency tests were performed in 2244 of the 2568 patients: 1302 patients randomized to ventricular pacing (88%) and 942 patients randomized to physiological pacing (86%). Despite the slight difference in UHR ascertainment rate and the fact that this subgroup was defined by postrandomization criteria, the 2 treatment groups remained well balanced with respect to important baseline clinical factors (Table 1). Failure to determine the UHR in the remaining 324 patients (172 with ventricular pacing and 152 with physiological pacing) was due, in approximately equal proportions, to the following: (1) a primary outcome event occurred before follow-up in 57 patients (33%) in the ventricular pacing group and 47 patients (31%) in the physiological pacing group; (2) UHR was not assessed at the first follow-up visit in 63 patients (37%) in the ventricular pacing group and 49 (32%) in the physiological pacing group; or (3) first follow-up was missed by 52 patients (30%) in the ventricular pacing group and 56 (37%) in the physiological pacing group. The distribution of reasons for no UHR determination was consistent between treatment groups. UHR, bpm No. of Patients (%) Ventricular Pacing, Events (Rate/Year) Physiological Pacing, Events (Rate/Year) Relative Risk Reduction (95% CI) P for Treatment UHR* (22) 51 (6.9) 2.4 (4.1) 38.4 ( 2, 63) (15) 28 (6.4) 20 (4.2) 37.8 ( 13, 65) (19) 38 (5.9) 20 (3.9) 40.3 ( 6, 66) (45) 70 (4.1) 44 (4.3) 1.9 ( 50, 31) (55) 117 (6.4) 64 (4.1) 35.5 (12, 53) (45) 70 (4.1) 44 (4.3) 1.9 ( 50, 31) CV indicates cardiovascular. *P value for linear trend with categorized UHR.

3 Tang et al Pacemaker Dependency and Pacing Mode 3083 Figure 1. Yearly event rates for the composite end point of cardiovascular mortality and stroke are plotted against UHR. The yearly event rate steadily increased with decreasing UHR in the ventricular pacing group, whereas the yearly event rate of the physiological pacing group was essentially constant over UHR. Table 2 shows the distribution of the UHR in the 2 treatment groups. The categorization of UHR into the 4 rate ranges was determined a priori on the basis of clinical considerations and not by observed data. There was a marked difference in UHR between the 2 treatment groups, with the ventricular pacing patients having higher values. Effect of Pacemaker Dependency on the Primary Outcome Table 3 shows the outcome event rates of cardiovascular death or stroke subdivided by UHR and treatment group. The observed treatment effect of physiological pacing in the group of patients with UHR 60 bpm was slightly negative, whereas a consistent positive risk reduction of between 37% and 40% was observed in the other 3 (smaller) subgroups with lower UHR. A test of trend in the size of the physiological pacing effect over the ordered UHR categories was not conventionally significant (P 0.089), but there was a trend favoring physiological pacing. The observed event rates in Table 3 are displayed in Figure 1. The yearly event rate steadily increased with decreasing UHR in the ventricular pacing group, whereas the yearly event rate of the physiological pacing group was essentially constant over UHR. Although clearly data driven by the observed pattern of treatment effect with UHR, it is appropriate to combine the 3 lower UHR categories with consistent treatment differences into a single UHR 60 bpm subgroup to compare with UHR 60 bpm. In addition, a cut point set at 60 bpm is clinically relevant, because most pacemakers are programmed to the low rate of 60 bpm. The bottom of Table 3 shows the comparison between the 2 subgroups, which yielded P for the difference in treatment effect between the 2 UHR subgroups (ie, a relative rate reduction of 35.5% for UHR 60 bpm versus 1.9% for UHR 60 bpm). The 95% confidence interval for the UHR 60 bpm subgroup does not include zero, which supports the presence of a qualitative interaction. The Kaplan-Meier estimates of cumulative risk of cardiovascular death or stroke for these 2 UHR subgroups are shown in Figure 2. The treatment advantage of physiological pacing in the UHR 60 bpm subgroup emerged after 1 year, and the 2 curves continued to diverge over time. In contrast, the slight early advantage of physiological pacing in the UHR 60 subgroup failed to be sustained, and the 2 curves intertwine. Effect of Pacemaker Dependency on Other Outcome Events Table 4 shows a comparison of the physiological pacing treatment effect between the UHR 60 subgroup for the secondary outcomes considered in CTOPP. There was a significant UHR subgroup effect (ie, the P value for the interaction between pacing mode and UHR) for the outcomes of cardiovascular death and any death. Figures 3 and 4 show the Kaplan-Meier estimates of cumulative risk of cardiovascular death and total mortality, respectively, for the 2 UHR subgroups. The relative risk reductions were 43.8 and 38.2, respectively. There was no UHR subgroup effect for stroke/ emboli or congestive heart failure hospitalization. For the secondary end point, atrial fibrillation, there was a modest but significant relative risk reduction of 18.6% in the overall CTOPP trial. 3 In this analysis, a somewhat larger rate reduction was observed in the UHR 60 subgroup compared with the UHR 60 subgroup, but the difference was not statistically significant (P 0.22) because there was also a modest risk reduction in the UHR 60 subgroup. Discussion The main finding of this analysis of CTOPP data is that the UHR at the first follow-up visit seems to have an important influence on how pacing mode selection affects the combined primary end point of cardiovascular death and stroke, as well as the secondary end points of cardiovascular death and total mortality. On the combined primary end point of stroke or cardiovascular death, there is a strong trend of an interaction Figure 2. Kaplan-Meier estimates of cumulative risk of cardiovascular death or stroke for the 2 UHR subgroups. There was a significant treatment effect of physiological pacing in the UHR 60 bpm subgroup but not in the UHR 60 bpm subgroup.

4 3084 Circulation June 26, 2001 TABLE 4. Event Rate for Secondary Outcomes by UHR Secondary Outcome UHR, bpm Ventricular Pacing, Event Rate/y Physiological Pacing, Event Rate/y Relative Risk Reduction, % P for Treatment UHR* 95% CI CV death , , 22 Any death , , 67 Stroke/emboli , , 74 CHF hospitalization , , 32 Atrial fibrillation , , 43 *P for linear trend with categorized UHR. between UHR and pacing mode, although it just missed being significant (P 0.058). This is further supported by the observation that in patients with an UHR 60 bpm, physiological pacing (an annual event rate of 4.1%) conferred a relative risk reduction of 35.5% compared with ventricular pacing (annual event rate of 6.4%) and that the 95% confidence interval (12%, 53%) does not include zero. There was also a significant interaction between UHR and pacing mode on the secondary end points of cardiovascular death and total mortality; thus, for patients with UHR 60 bpm, physiological pacing conferred a relative risk reduction of 43.8% for cardiovascular death and 38.1% for total mortality. Permanent pacing was introduced in the late 1950s. The appreciation of the hemodynamic importance of atrial systole and advancement of pacing technology promoted the use of dual-chamber pacemakers beginning in the late 1970s. Currently, dual-chamber pacemaker usage varies worldwide, even in developed countries, from 70% in the Unites States to 23% in Japan. 4 Atrial or dual-chamber pacing has been demonstrated to be superior to ventricular pacing in deriving better hemodynamic parameters, especially in patients with a failing or hypertrophic ventricle. 2 Observational studies have suggested that atrial or dual-chamber pacing is associated with about a two-thirds lower risk of atrial fibrillation and a one-third lower risk of death compared with ventricular pacing. 5 Several clinical trials attempting to address the issue of pacemaker mode selection have been completed recently. 6 9 The results of these studies are not completely consistent, although the larger studies suggest that ventricular and physiological pacing are not different with respect to total mortality but that physiological pacing results in a lower rate of atrial fibrillation. The influence of pacemaker dependency on pacemaker mode selection was not addressed in these studies. The present analysis of the CTOPP data provides this important information on pacemaker dependency and pacemaker mode selection. Pacemaker-dependency testing is routinely done during follow-up in most pacemaker clinics. The purpose of this test is to evaluate the patient s underlying rhythm should a pacemaker malfunction occur unexpectedly and result in no heart rate support In the present study, we used this test to expose the unpaced ventricular rate, which was used as a indicator of how much the patient may be paced. Direct measurement of percentage paced is now available in many models of pacemakers. This parameter may be useful, but it is dependent on the low rate setting of the pacemaker. Furthermore, pacemakers with this feature were not uniformly used when the CTOPP study was performed. There are some limitations to this analysis of the CTOPP data. The pacemaker-dependency test was not performed before pacemaker implantations. The first test was performed at 6 months follow-up. As a result, 12% of the patients were not included in the analysis because these data were not available. This may potentially cause a bias to the results, although the missing data occurred equally in the 2 randomized groups. Randomization of the pacing mode did not take Figure 3. Kaplan-Meier estimates of cumulative risk of cardiovascular death for the 2 UHR subgroups. There was a significant treatment effect of physiological pacing in the UHR 60 bpm subgroup but not in the UHR 60 bpm subgroup.

5 Tang et al Pacemaker Dependency and Pacing Mode 3085 Figure 4. Kaplan-Meier estimates of cumulative risk of all-cause mortality for the 2 UHR subgroups. There was a significant treatment effect of physiological pacing in the UHR 60 bpm subgroup but not in the UHR 60 bpm subgroup. pacemaker dependency into consideration. The UHR distribution between the ventricular pacing group and the physiological pacing group was not the same. However, for the patients included in this analysis, the baseline patient characteristics were very similar between the 2 treatment groups. Despite these limitations, the results suggest an influence of UHR with pacemaker mode selection on the outcome measures. Ongoing pacing trials such as the United Kingdom Pacing and Clinical Event Trial (UK-PACE) 13 and the Danish Multicenter Randomized Study on AAI or DDD Pacing in Sick Sinus Syndrome (DANPACE) 14 may provide prospective data on the effect of underlying heart rate and pacing mode selection on morbidity and mortality. The former trial randomizes elderly patients with heart block to physiological or ventricular pacing; the latter randomizes patients with sick sinus syndrome to atrial pacing or dual-chamber pacing, and the pacemakers used will be able to record the amount of pacing. Conclusions This study demonstrated that the UHR at the first follow-up visit has an important influence on how pacing mode selection affects the combined end point of stroke or cardiovascular death, as well as cardiovascular death and total mortality. Pacemaker-dependent patients with a low UHR will probably be paced frequently, and they will likely benefit from physiological pacing. In contrast, non pacemakerdependent patients will likely be paced infrequently and may not benefit from physiological pacing. In the latter group, a ventricular pacemaker serving as a backup will be sufficient. References 1. Jeffrey K, Parsonnet V. Cardiac pacing, : a quarter century of medical and industrial innovation. Circulation. 1998;97: Janosik D, Labovitz AJ. Basic physiology of cardiac pacing. In: Ellenbogen KA, Kay GN, Wilkoff BL, eds. Clinical Cardiac Pacing. Toronto: Saunders; 1995: Connolly SJ, Kerr CR, Gent M, et al. Effects of physiologic pacing versus ventricular pacing on the risk of stroke and death due to cardiovascular causes: Canadian Trial of Physiologic Pacing Investigators N Engl J Med. 2000;342: Gillis AM, Goldman BS, Yee R, et al. Cardiac pacing in Canada in 1998: working towards optimal pacing therapy: Canadian Working Group on Cardiac Pacing Can J Cardiol. 1998;14: Connolly SJ, Kerr C, Gent M, et al. Dual-chamber versus ventricular pacing: critical appraisal of current data. Circulation. 1996;94: Andersen HR, Thuesen L, Bagger JP, et al. Prospective randomised trial of atrial versus ventricular pacing in sick-sinus syndrome. Lancet. 1994; 344: Mattioli AV, Vivoli D, Mattioli G. Influence of pacing modalities on the incidence of atrial fibrillation in patients without prior atrial fibrillation: a prospective study. Eur Heart J. 1998;19: Andersen HR, Nielsen JC, Thomsen PEB, et al. Long-term follow-up of patients from a randomised trial of atrial versus ventricular pacing for sick-sinus syndrome. Lancet. 1997;350: Lamas GA, Orav J, Stambler BS, 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: Staessen J, Ector H, de Geest H. The underlying heart rhythm in patients with an artificial cardiac pacemaker. Pacing Clin Electrophysiol. 1982; 5: Crick JCP, Rokas S, Sowton E. Identification of pacemaker dependent patients by serial decremental rate inhibition. Eur Heart J. 1985;6: Rosenheck S, Bondy C, Weiss AT, et al. Comparison between patients with and without reliable ventricular escape rhythm in the presence of long standing complete atrioventricular block. Pacing Clin Electrophysiol. 1993;16: Lamas GA. Pacemaker mode selection and survival: a plea to apply the principles of evidence based medicine to cardiac pacing practice. Heart. 1997;78: Andersen HR, Nielsen JC. Pacing in sick sinus syndrome: need for a prospective, randomized trial comparing atrial with dual chamber pacing. Pacing Clin Electrophysiol. 1998;21: