Comparative assessment of right, left, and biventricular pacing in patients with permanent atrial fibrillation

Transcription

1 European Heart JournaL (2005) 26, doi: /eurheartj/ehi069 Clinical research Comparative assessment of right, left, and biventricular pacing in patients with permanent atrial fibrillation M. Brignole 1 *, M. Gammage 2, E. Puggioni 1, P. Alboni 3, A. Raviele 4, R. Sutton 5, P. Vardas 6, M.G. Bongiorni 7, L. Bergfeldt 8, C. Menozzi 9, and G. Musso 10 on behalf of the Optimal Pacing SITE (OPSITE) Study Investigators 1 Department of Cardiology, Ospedali del Tigullio, Via don Bobbio, Lavagna, Italy 2 Department of Cardiology, University Hospital and University of Birmingham, Birmingham, UK 3 Department of Cardiology, Ospedale Civile, Cento, Italy 4 Department of Cardiology, Ospedale Umberto I, Mestre, Italy 5 Department of Cardiology, Royal Brompton Hospital, London, UK 6 Department of Cardiology, University Hospital, Heraklion, Greece 7 Department of Cardiology, Ospedale Cisanello, Pisa, Italy 8 Department of Cardiology, Karolinska University Hospital, Solna, Stockholm, Sweden 9 Department of Cardiology, Ospedale S Maria Nuova, Reggio Emilia, Italy 10 Department of Cardiology, Ospedale Civile, Imperia, Italy Received 3 July 2004; revised 7 October 2004; accepted 28 October 2004; online publish-ahead-of-print 20 December 2004 See page 637 for the editorial comment on this article (doi: /eurheartj/ehi234) KEYWORDS Atrial fibrillation; Heart failure; Bundle branch block; Catheter ablation; Resynchronization pacing Aims Left ventricular (LV) and biventricular (BiV) pacing are potentially superior to right ventricular (RV) apical pacing in patients undergoing atrioventricular (AV) junction ablation and pacing for permanent atrial fibrillation. Methods and results Prospective randomized, single-blind, 3-month crossover comparison between RV and LV pacing (phase 1) and between RV and BiV pacing (phase 2) performed in 56 patients (70+8 years, 34 males) affected by severely symptomatic permanent atrial fibrillation, uncontrolled ventricular rate, or heart failure. Primary endpoints were quality of life and exercise capacity. Compared with RV pacing, the Minnesota Living with Heart Failure Questionnaire (LHFQ) score improved by 2 and 10% with LV and BiV pacing, respectively, the effort dyspnoea item of the Specific Symptom Scale (SSS) changed by 0 and 2%, the Karolinska score by 6 and 14% (P, 0.05 for BiV), the New York Heart Association (NYHA) class by 5 and 11% (P, 0.05 for BiV), the 6-min walked distance by 12 (þ4%) and 4 m (þ1%), and the ejection fraction by 5 and 5% (P, 0.05 for both). BiV pacing but not LV pacing was slightly better than RV pacing in the subgroup of patients with preserved systolic function and absence of native left bundle branch block. Compared with pre-ablation measures, the Minnesota LHFQ score improved by 37, 39, and 49% during RV, LV, and BiV pacing, respectively, the effort dyspnoea item of the SSS by 25, 25, and 39%, the Karolinska score by 39, 42, and 54%, the NYHA class by 21, 25, and 30%, the 6-min walking distance by 35 (12%), 47 (16%), and 51 m (19%) and the ejection fraction by 5, 10, and 10% (all differences P, 0.05). * Corresponding author. Tel: þ ; fax: þ address: mbrignole@asl4.liguria.it & The European Society of Cardiology All rights reserved. For Permissions, please journals.permissions@oupjournals.org

2 Resynchronization pacing and atrial fibrillation 713 Conclusions Rhythm regularization achieved with AV-junction ablation improved quality of life and exercise capacity with all modes of pacing. LV and BiV pacing provided modest or no additional favourable effect compared with RV pacing. Introduction Pacing from the apex of the right ventricle is not considered optimal since it provides a non-physiological asynchronous contraction, resulting in a decrease in cardiac performance; 1,2 in addition, it causes an electrocardiographic pattern similar to left bundle branch block (LBBB). In one study 3 performed in patients with otherwise normal hearts and intermittent LBBB, the advent of LBBB was associated with a significant deterioration of cardiac function of 10 20%. In patients with permanent atrial fibrillation (AF) who have received atrioventricular (AV) junction ablation and pacing from the right ventricular (RV) apex, the beneficial haemodynamic effect of regularization and slowing of heart rhythm is thus assumed to be partly counteracted by the adverse haemodynamic effect of a non-physiological pacing mode. 4 The Optimal Pacing Site (OPSITE) study is a prospective randomized, single-blind cross-over comparison between right, left (LV) or biventricular (BiV) pacing for patients with permanent AF undergoing ablation and pacing therapy. The study consisted of an acute and a chronic evaluation. The results of the acute evaluation have been published previously. 5 In this report we present the results of the chronic evaluation. The main study hypothesis was that LV pacing and BiV pacing are better than RV pacing in improving quality of life (QoL) and exercise capacity in patients with permanent AF treated with ablation and pacing therapy. Secondary objectives were the comparison between two pre-defined subgroups of patients with preserved or depressed systolic function, and the comparison of the two modes of pacing with baseline measures in order to evaluate the effect of AV-junction ablation on quality of life and exercise capacity. Methods Patient selection The following patients were eligible for enrolment in the OPSITE study: (i) patients with permanent AF in whom a clinical decision had been made to undertake complete AV-junction ablation and ventricular pacing because of drug-refractory, severely symptomatic, uncontrolled high ventricular rate, and (ii) patients with permanent AF, drug-refractory heart failure, depressed LV function, and/or LBBB in whom a clinical decision had been made to undertake LV re-synchronization pacing. Patient exclusion criteria were as follows: (i) New York Heart Association (NYHA) class IV heart failure, despite optimized therapy; (ii) severe concomitant non-cardiac disease; (iii) need for surgical intervention; (iv) myocardial infarction within 3 months; (v) sustained ventricular tachycardia or ventricular fibrillation; (vi) previously implanted pacemaker; (vii) inabilty to obtain reliable RV and LV pacing and persistent AV block. Two different subgroups were pre-defined for analysis: patients with an ejection fraction.40% and absence of LBBB pattern (group A); and patients with heart failure, i.e. those with ejection fraction 40% and/or LBBB pattern (group B). Pacemaker implantation and ablation were allowed at different times but required to occur,6 weeks apart. Right ventricular leads were positioned in the RV apex. LV leads were positioned via the coronary sinus in a position considered most appropriate by the implanting physician; in case of failure of pacing through the coronary sinus, an epicardial lead was implanted. A conventional dual-chamber rate-responsive pacemaker was used; the atrial port of the pacemaker was connected to the LV lead and the ventricular port to the RV lead. The AV interval of the pacemaker (the effective LV RV interval) was programmed 30 ms in order to achieve (almost) simultaneous BiV pacing or 200 ms during the LV phase in order to avoid RV pacing but to assure a back-up RV pacing in case of LV pacing failure. The pacemaker was programmed VVIR during the RV phase. Lowest rate was set at 80 beats per min (b.p.m.). Study design and randomization The study was divided into two phases. Each phase consisted of a 3-month randomized crossover design: Phase 1: comparison between RV and LV pacing Phase 2: comparison between BiV and RV pacing Following ablation and pacing therapy the patients underwent randomization and phase 1 follow-up started immediately. Randomization was computer-generated, blocked by centre, and assignments were hidden from participants until the time of allocation. Two different randomization sequences were generated with phase 1 assignment linked to phase 2:. RV to LV to RV to BiV. LV to RV to BiV to RV Moreover, two different randomization groups were generated for subgroups A and B. Phase 2 started immediately after the end of phase 1. Endpoints Primary endpoints were the evaluation of QoL and exercise capacity performed at the end of each of the 3-month periods. Secondary endpoints were the evaluation of the effects of AV-junction ablation on QoL and exercise capacity and the comparison between the two pre-defined subgroups. Outcome measures These were assessed at the time of enrolment and at the end of each 3-month study period. QoL was measured by means of the Minnesota Living with Heart Failure Questionnaire (LHFQ), 6 the Specific Symptoms Scale (SSS), 4,7 the modified Karolinska questionnaire, 8 and the NYHA classification. 9 Exercise capacity was assessed by the 6-min walking test (average of two consecutive tests). 10

3 714 M. Brignole et al. Statistical analysis Endpoints could be assessed only in patients with no missing data after completion of both crossover periods. Paired and unpaired one-sided Student s t-test was used for comparison of continuous variables as appropriate. The z-test was used for intrapatient comparison of proportions. The treatment period interaction (residual or carryover effect) was tested by a Student s t-test applied to the individual sums of the first and second period data. A P-value,0.05 was considered significant. Sample size In patients with severely symptomatic chronic AF, uncontrolled by conventional drugs, the mean score of the Minnesota LHFQ was 44 and decreased to 32 after AV-junction ablation and pacing, 4 the highest score of the SSS (6.5 points) was related to effort dyspnoea and decreased to 4.5 after AV-junction ablation and pacing. 4 The minimum chance necessary to allow 99% confidence that a real change of the 6-min walked distance has occurred is 10% of the average performance, using the averaged results of two consecutive tests at baseline. 10 Based on these results, a 20% reduction of the Minnesota LHFQ and of the effort dyspnoea item of the SSS, and a 10% increase in the 6-min walked distance were pre-determined as the minimum requirements for superiority of LV-based pacing over RV pacing of sufficient clinical relevance. The sample size able to provide 90% power to show an intrapatient difference, with a probability of 95%, was 40 patients. Results Participant flow and follow-up Fifty-six patients underwent successful pacemaker implantation and AV-junction ablation and were randomized between July 2001 and January 2003 (Figure 1 ). Their clinical characteristics are shown in Table 1. RV leads were positioned in the RV apex in all patients. LV leads were positioned via the coronary sinus in the midposterolateral site in 51 patients and in the anterior site in three patients. In two patients, who had failed the coronary sinus approach, the lead was implanted in an epicardial mid-posterolateral position through a limited thoracotomy. Six patients died during the year of the study, accounting for an 11% mortality rate; of these, three died suddenly and three from heart failure. During phase 2 of the study, six other patients dropped out because of a progressive deterioration in their general condition, mostly due to worsening heart failure. Thus, overall, 12 patients (21%) had severe clinical events during the study period, six during RV pacing and six during LV/BiV pacing. Finally, three patients refused to complete the follow-up. In two patients, AV conduction resumed after a few days and a second ablation procedure was rapidly performed, which achieved persistent AV block; in one patient a transient increase in LV pacing threshold occurred during phase 1. These patients were included in Figure 1 Patient flow. SD, sudden death; CHF, congestive heart failure.

4 Resynchronization pacing and atrial fibrillation 715 the analysis. Pharmacological therapy remained stable from baseline and during the crossover phases (Table 2 ). Phase 1 Fifty-two patients completed phase 1. QRS duration was ms with RV pacing and ms with LV pacing. Compared with RV pacing, the improvement with LV was modest and only echocardiographic variables showed statistically significant changes (Table 3 ). Conversely, a great improvement was observed from baseline to both RV and LV modes for all variables. For example, the Minnesota LHFQ improved by only 2% from RV to LV pacing and 37 and 39% from baseline to RV and LV pacing, respectively. LV pacing led to no improvement in effort dyspnoea score of the SSS from RV to LV pacing, whereas the improvement from baseline to RV Table 1 Patients characteristics at enrolment Number of patients 56 Age, years Gender, males 34 (61) Duration of atrial fibrillation, years Number of hospitalizations per patient NYHA functional class Minnesota LHFQ, score Six-minute walking test, m Standard electrocardiogram: Mean heart rate, b.p.m LBBB 28 (50) Holter monitoring: Minimum heart rate, b.p.m Mean heart rate, b.p.m Maximum heart rate, b.p.m Standard echocardiogram: mean ejection fraction Associated structural heart disease Coronary artery disease 17 (30) Others 39 (70) Data are mean (standard deviation) or number (%). and LV pacing was 25%. The Karolinska score improved by 6% from RV to LV pacing, and 39 and 42% from baseline to RV and LV pacing, respectively. The NYHA class improved by 5% from RV to LV pacing, and 21 and 25% from baseline to RV and LV pacing, respectively. The 6-min walked distance increased by 12 m (4%) from RV to LV pacing, and 35 (12%) and 47 m (16%) from baseline to RV and LV pacing, respectively. Compared with RV, ejection fraction significantly improved by 5% and mitral regurgitation score decreased by 18% with LV pacing (Table 3 ). At the end of phase 1, 23 (44%) patients preferred LV pacing, 16 (31%) RV pacing, and 13 (25%) had no preference (P ¼ 0.04). Phase 2 Forty-one patients completed both phase 1 and 2. In phase 2, QRS duration was ms with RV pacing and ms with BiV pacing (P ¼ 0.001). Compared with RV pacing, BiV pacing showed a significant improvement in Karolinska questionnaire and NYHA class but not in the other measures of QoL. In absolute terms, however, the improvement was modest and much less than that observed from baseline to RV and BiV pacing for all variables (Table 4). For example, the Minnesota LHFQ improved by 10% from RV to BiV pacing, and 43 and 49% from baseline to RV and BiV pacing, respectively. The effort dyspnoea score of the SSS improved by 2% from RV to BiV pacing, and 38 and 39% from baseline to RV and BiV pacing. The Karolinska score improved by 14% from RV to BiV pacing, and 47 and 54% from baseline to RV and BiV pacing. The NYHA class improved by 11% from RV to BiV pacing, and 22 and 30% from baseline to RV and BiV pacing. The 6-min walked distance increased by 4 m (1%) from RV to BiV pacing, and 53 (18%) and 57 m (19%) from baseline to RV and BiV pacing. Conversely, the palpitation score of the SSS worsened significantly with both LV and BiV pacing. Compared with RV, ejection fraction significantly improved by 5% and mitral regurgitation score decreased by 25% with BiV pacing (Table 4 ). Table 2 Concomitant pharmacological therapy throughout the study Drug Baseline, n ¼ 52 (%) Randomization, n ¼ 52 (%) Phase 1, n ¼ 52 (%) Phase 2, n ¼ 41 (%) RV LV RV BiV Digoxin 38 (73) 32 (61) 34 (65) 32 (62) 24 (59) 24 (59) Diuretics 41 (79) 41 (79) 42 (81) 42 (81) 33 (80) 34 (83) Nitrates 10 (19) 9 (17) 10 (19) 10 (19) 11 (27) 11 (27) ACE-inhibitors 38 (73) 40 (77) 36 (69) 36 (69) 31 (76) 28 (68) Beta-blockers 28 (54) 27 (52) 27 (52) 25 (48) 20 (49) 19 (46) Calcium-antagonists 11 (21) 7 (13) 7 (13) 6 (12) 4 (10) 3 (7) Aspirin 4 (8) 4 (8) 3 (6) 4 (8) 3 (7) 5 (12) Warfarin 44 (85) 48 (92) 50 (96) 48 (92) 38 (93) 35 (85) Amiodarone 5 (10) 3 (6) 3 (6) 1 (2) 1 (2) 1 (2) Others 15 (29) 12 (23) 16 (31) 13 (25) 6 (15) 6 (15) ACE, angiotensin-converting enzyme.

5 Table 3 Baseline and phase 1 results in the 52 patients who completed the phase 1 period Baseline Phase 1 RV 1 vs. baseline LV vs. baseline RV 1 LV Difference (95% CI) P Difference (95% CI) P Difference (95% CI) P LHFQ, score (range 1 105) (24.7; þ 3.2) (211.7;222.9) (212.5; ) SSS score (range 1 10) Palpitations þ0.2 (20.3; þ 0.7) (25.3;23.1) (25.1;22.9) Effort dyspnoea (20.8; þ 0.8) (22.6;20.6) (22.8;20.4) Rest dyspnoea (20.7; þ 0.5) (22.2;20.6) (22.3;20.7) Effort intolerance (20.7;þ0.7) (22.7;20.9) (22.7;20.9) Easy fatigue (20.7; þ 0.7) (21.7;20.1) (21.7; 2 0.1) Chest discomfort þ0.3 (20.3; þ 0.9) (22.4;20.9) (21.7;20.3) 0.01 Karolinska, score (range 0 16) (21.1; þ 0.5) (23.9;22.5) (24.2;22.6) NYHA class (range 1 4) (20.3; þ 0.1) (20.7;20.3) (20.8;20.4) Six-minute walked distance, m (average of 2) þ12 (þ37; 2 13) 0.13 þ35 (þ65;þ5) 0.01 þ47 (þ75;þ26) Echocardiographic measures Ejection fraction, % þ2 (0; þ 4) 0.04 þ2 (21;þ5) 0.04 þ4 (þ2;þ6) LVEDD, mm (22;0) 0.06 þ1 (0;þ2) (21;þ1) 0.26 LVESD, mm (22;0) (23;þ1) (0;þ 4) 0.04 Mitral regurgitation, score (range 1 4) (20.5; 2 0.1) (20.2;þ0.2) (20.5;20.1) LVEDD, left ventricular end-diastolic diameter; LVESD, left ventricular end-systolic diameter. Table 4 Baseline and phase 2 results in the 41 patients who completed the study period Baseline Phase 2 RV 2 vs. baseline BiV vs. baseline RV 2 BiV Difference (95% CI) P Difference (95% CI) P Difference (95% CI) P LHFQ, score range 1 105) (26.4;þ1.0) (226.4;213.6) (227.8;217.6) SSS score (range 1 10) Palpitations þ0.8 (0;þ1.6) (25.7;23.3) (25.1;22.3) Effort dyspnoea (21.1;þ0.9) (23.8;21.4) (23.8;21.2) Rest dyspnoea (21.0;þ0.8) (22.4;20.6) (22.5;20.7) Effort intolerance þ0.6 (20.4;þ1.6) (23.6;21.6) ( ) Easy fatigue (20.5;þ0.5) (21.8;20.2) (21.5;20.3) Chest discomfort þ0.3 (20.5;þ1.1) (22.3;20.3) (21.8;20) 0.01 Karolinska, score (range 0 16) (21.2;0) (24.8;23.0) (25.2;24.4) NYHA class (range 1 4) (20.4;0) (20.7;20.3) (20.9;20.5) 0.02 Six-minute walked distance, m (average of 2) þ4 (þ16;28) 0.27 þ53 (þ81;þ28) 0.01 þ57 (þ92;þ22) Echocardiographic measures Ejection fraction, % þ2 (0;þ4) 0.02 þ2 (21;þ5) 0.16 þ4 (þ7;þ1) LVEDD, mm (22.5;þ0.5) 0.25 þ2 (þ1;þ3) 0.10 þ1 (0;þ2) 0.17 LVESD, mm (22;0) (23;þ1) (0;þ4) 0.02 Mitral regurgitation, score (range 1 4) (20.6;20.3) (20.2;þ0.2) (20.6;20.2) M. Brignole et al.

6 Resynchronization pacing and atrial fibrillation 717 Figures 2, 3, and 4 evaluate the effect of the two sequences of randomization and the treatment period interaction (residual or carryover effect) of the Minnesota LHFQ, the 6-min walked distance, and the ejection fraction. No statistical difference was observed except from baseline. The intrapatient comparison of the differences RV LV and RV BiV performed on the 41 patients showed no significant difference in any variable (Table 5). At the end of phase 2, 25 (61%) patients preferred BiV pacing, 8 (20%) RV pacing, and 8 (20%) had no preference (P ¼ 0.001). Subgroup analysis Among the patients who completed phase 1, 19 belonged to group A and 33 to group B (25 of group B had LBBB) (Table 6 ). The ejection fraction was 53+8 and %, and the QRS width was and ms, respectively, in groups A and B. The effects of LV vs. RV pacing were not statistically different in the two groups. Among the patients who completed phase 2, 17 belonged to group A and 24 to group B (18 of group B had LBBB) (Table 7 ). There was a trend towards a better outcome with BiV pacing in group A as suggested by the significantly higher improvement in easy fatigue score, NYHA class, 6-min walked distance, ejection fraction, and of the Minnesota LHFQ and Karolinska score, although the latter is not significant. Discussion Primary endpoints were QoL and exercise capacity. Although we observed a modest significant improvement in a few measures, especially with BiV pacing, in general we were unable to show a definite superiority of LV-based pacing over RV pacing. In absolute terms, the improvement was modest for any variable and inferior to the minimal level of clinical efficacy with the exception of a few parameters with BiV pacing in the subgroup of patients with preserved systolic function and narrow QRS. The study was powered to show a minimum benefit of clinical relevance, defined as an improvement of 20% of QoL scores and 10% of the 6-min walked distance; these values were chosen based on the results of previous trials (see Methods). The changes resulting from the present study were far lower. Even if some improvement of minor degree may still be present, its clinical utility seems modest. For example, the maximum possible improvement of the Minnesota LHFQ, corresponding to the upper 95% of the confidence interval, was 4.7 points, of a total of 105, for LV vs. RV pacing and 6.4 points for BiV pacing and the maximum improvement of the 6-min walked distance was 37 and 16 m, respectively. The same reasoning applies to the other outcome measures. The situation is almost certainly one in which some patients are showing marked clinical benefit, balanced by other patients with very little benefit. In Table 8 we have reported the percentage of patients who showed a benefit from one mode with respect to the other of sufficient entity to be clinically relevant. Up to a quarter of patients were better with RV pacing than with LV or BiV pacing. Only a quarter or less of the patients had a relevant objective benefit from LV or BiV pacing. Thus a heterogeneity effect of resynchronization therapy is present in patients with AF. In contrast, the improvement in QoL observed with the rhythm regularization obtained with AV-junction ablation and RV pacing was much more impressive, being in general from 3 to 10 times higher than that obtained from RV to LV or BiV pacing. For example, the effect of ablation and RV pacing on the Minnesota LHFQ determined an improvement of 37 43% whereas the additive effect of LV or BiV pacing was only a further 2 10%. Admittedly, the evaluation of AV-junction ablation was not controlled and a part of the total benefit was not due to ablation alone. A previous study has evaluated that a placebo effect of pacing may account for 9 40% of the total improvement of NYHA class and of the SSS; 7 in another study 4 the control arm treated with pharmacological therapy had an improvement of 7% in the NYHA class and of 16% in the Minnesota LHFQ. The improvement in QoL measures with LV and BiV pacing is consistent with that of the ejection fraction at the end of each 3-month crossover phase in this study which was 5 and 5% and that observed during the acute study performed immediately after Figure 2 Effect of the two sequences of randomization and the treatment period interaction (residual or carryover effect) on the Minnesota LHFQ.

7 718 M. Brignole et al. Figure 3 Effect of the two sequences of randomization and the treatment period interaction (residual or carryover effect) on the 6-min walked distance. Figure 4 Effect of the two sequences of randomization and the treatment period interaction (residual or carryover effect) on the ejection fraction. AV-junction ablation, which was 6%. 5 An improvement of ejection fraction was also observed by Leclercq et al. 11 and by Leon et al. 12 The improvement of ejection fraction was due to a reduction of both end-diastolic and end-systolic diameters (Tables 3 and 4 ) and suggests a beneficial influence of resynchronization pacing on the heart. It is possible that the duration of the present study is too short to show functional benefit over the benefit of rate control (as experienced in the ROVA trial 13 ) and that a longer observation period could allow the potential benefit of resynchronization therapy to become more manifest. Only a few other studies have evaluated the effect of LV-based pacing on QoL of the patients with AF and none have evaluated its effect on mortality. In the AF arm of the only other randomized clinical study, the MUSTIC trial, 14 performed on 39 evaluable patients, the intention-to-treat analysis did not show any statistically significant difference in either primary or secondary endpoints and efficacy analysis showed only slight significant difference in a few endpoints between BiV and RV pacing. It seems likely that the main reason for the modest effect of LV-based pacing in patients with AF is that it is additive to the powerful beneficial effect of rhythm regularization and slowing achieved with AV-junction ablation per se, which reduces the amount of the potential additional benefits obtainable through LV or BiV pacing, at least over the time frame of this study. The beneficial effect of ablation and RV pacing is well known from both haemodynamic 5,15 18 and clinical studies. 4,7,19 21 This situation does not occur in patients in sinus rhythm undergoing resynchronization therapy. Comparison with patients in sinus rhythm The effect of resynchronization therapy in patients in sinus rhythm seems to be different from the effect in those in AF. There is increasing evidence for the favourable effect of cardiac resynchronization pacing in patients with heart failure and intraventricular conduction delay who are in sinus rhythm In the sinus rhythm arm of the single-blind, crossover MUSTIC trial, 22 the Minnesota LHFQ improved by 13 points (32%) and the 6-min walked distance improved by 73 m (23%) with BiV pacing compared with no pacing. Auricchio et al., 23 in a single-blind crossover study, showed an improvement in the Minnesota LHFQ by eight points (29%), of 6-min walked distance by 47 m (12%) with LV pacing vs. no pacing. In the single-blind, parallel-controlled, randomized MIRACLE trial, 24 the Minnesota LHFQ score improved by nine points and

8 Resynchronization pacing and atrial fibrillation 719 Table 5 Phase 1 and phase 2 results in the 41 patients who completed the study period Phase 1 Phase 2 Diff. phase 1 vs. diff. phase 2 RV 1 LV P RV 2 BiV P P LHFQ, score (range 1 105) SSS score (range 1 10) Palpitations Effort dyspnoea Rest dyspnoea Effort intolerance Easy fatigue Chest discomfort Karolinska, score (range 0 16) NYHA class (range 1 4) Six-minute walked distance, m (average of 2) Echocardiographic measures Ejection fraction, % LVEDD, mm LVEDS, mm Mitral regurgitation, score (range 1 4) Table 6 Subgroup analysis, phase 1: comparison between group A and group B patients Group A (n ¼ 19) Group B (n ¼ 33) Diff. Gr. A vs. diff. Gr. B RV 1 LV P RV 1 LV P P LHFQ, score (range 1 105) SSS score (range 0 10) Palpitations Effort dyspnoea Rest dyspnoea Effort intolerance Easy fatigue Chest discomfort Karolinska, score (range 0 16) NYHA class (range 1 4) Six-minute walked distance, m (average of 2) Echocardiographic measures Ejection fraction, % LVEDD, mm LVEDS, mm Mitral regurgitation, score (range 1 4) the 6-min walked distance by 29 m in paced vs. not paced patients. In the double-blind, parallel-controlled MIRACLE ICD trial, 25 the Minnesota LHFQ improved by a median of 6.5 points, NYHA class by 1.0 point, treadmill exercise duration by 66 s whereas the 6-min walked distance increased by only 2 m. Apart from the specific effect of AV-junction ablation, there are other differences between patients in sinus rhythm and in AF that could explain different clinical results. In sinus rhythm patients, the optimal resynchronization occurs with the fusion of intrinsic RV activation and paced LV activation, which is dependent on AV synchronization. 26 Clearly, after ablating the AV-junction, intrinsic rhythm is not conducted and thus fusion is not possible. Typically, patients enrolled in sinus rhythm have very low ejection fraction values and wide QRS complexes. In our study the mean ejection fraction was % and only half had LBBB. Despite the fact

9 720 M. Brignole et al. Table 7 Subgroup analysis, phase 2: comparison between group A and group B patients Group A (n ¼ 17) Group B (n ¼ 24) Diff. Gr. A vs. diff. Gr. B RV 2 BiV P RV 2 BiV P P LHFQ, score (range 1 105) Specific symptom scale, score (range 0 10) Palpitations Effort dyspnoea Rest dyspnoea Effort intolerance Easy fatigue Chest discomfort Karolinska, score (range 0 16) NYHA class (range 1 4) Six-min walked distance, m (average of 2) Echocardiographic measures Ejection fraction, % LVEDD, mm LVEDS, mm Mitral regurgitation, score (range 1 4) Table 8 Analysis of heterogeneity: interpatient variability Differences between mode of pacing Phase 1 (n ¼ 52) Phase 2 (n ¼ 41) RV 1 better LV better P RV 2 better BiV better P LHFQ, difference 10 points, % pts 12 (23) 10 (19) (12) 5 (12) 1.00 SSS Palpitations, difference 2 points 8 (16) 6 (12) (15) 2 (5) 0.07 Effort dyspnoea, difference 2 points 10 (20) 13 (25) (17) 10 (24) 0.21 Rest dyspnoea, difference 2 points 6 (12) 5 (10) (5) 3 (7) 0.67 Effort intolerance, difference 2 points 7 (14) 8 (16) (17) 7 (17) 0.99 Easy fatigue, difference 2 points 7 (14) 6 (12) (5) 3 (7) 0.67 Chest discomfort, difference 2 points 3 (6) 2 (4) (10) 3 (7) 0.80 Karolinska, difference 2 points 13 (25) 12 (23) (10) 11 (27) NYHA, difference 1 class 5 (10) 11 (21) (10) 9 (22) 0.04 Six-minute walked distance, difference 50 m 5 (12) 13 (25) (12) 6 (15) 0.74 Data are number (%). that the mean LV ejection fraction was only moderately diminished, we observed a high mortality of 11% during the 1-year follow-up. In the MIRACLE trial, 24 patients in sinus rhythm had a mortality of 6.1% at 6 months with an ejection fraction of about 22%. Subgroup analysis The clinical results were similar or even better, with BiV pacing, in the patients with preserved systolic function and no LBBB than in the other subgroup. This finding is original and confirms the haemodynamic effect seen in the acute study. 5 In contrast, in sinus rhythm patients, a benefit was seen in patients with QRS duration.150 ms but not in those with a QRS duration of ms. 23 However, this finding needs to be verified in a larger population. Comparison between LV and BiV pacing Theoretically, in the absence of fusion between intrinsic RV activation and paced LV activation, pre-excitation of the left ventricle alone may create delayed activation of the septum and right ventricle, which might worsen ventricular pump function, analogous to the deficits generated by RV-only pacing. In the present study, BiV pacing resulted in a greater shortening of the QRS duration, possibly indicating better ventricular synchronization. As a result, BiV pacing may be more effective than LV pacing in AF patients. The study was, however, not designed for this comparison. A comparison between

10 Resynchronization pacing and atrial fibrillation 721 phases 1 and 2 cannot be performed because of the different populations due to several drop-outs in phase 2. Among the patients who completed both phases of the study, we were unable to find any statistical difference between LV and BiV pacing except for an advantage of BiV but not for LV pacing in subgroup A patients (Table 7 ). Finally, the study is probably underpowered to show small differences between the two modes of pacing. Thus, we believe that it is prudent not to draw any conclusion in this respect. Study power Although the study population was small, it is unlikely that even with a larger population we would have observed clinically important favourable results on the primary endpoints of the study. The improvement in QoL was lower than expected (see Sample size section). For example, the increase in the 6-min walked distance was 4% (95% CI 24% to þ10%) in phase 1 and 1% (95% CI 25% to þ7%) in phase 2; the decrease in the LHFQ score was 22% (95% CI þ11% to 216%) in phase 1 and 210% (95% CI þ4% to 24%) in phase 2. Limitations Selection bias might have been introduced into the comparison of BiV and RV. Fifteen patients dropped out from the study before the termination of phase 2. The study population included for the comparison between RV and BiV in the second phase depends on the outcome of the first phase and is therefore not necessarily the population of interest. Even if not significant, some treatment period interaction could have been present. On the other hand, the similarity of measures between the two RV periods (one in phase 1 and the other in phase 2) suggests stability of clinical condition. Therefore, the clinical significance of the study remains, i.e. BiV pacing provided modest additive clinical benefit to that already achieved with AV-junction ablation and BiV pacing benefits are not very different from those obtained with LV pacing alone. Conclusions and perspectives Rhythm regularization achieved with AV-junction ablation improved QoL and exercise capacity with all modes of pacing. LV and BiV pacing provided modest or no additional favourable effect compared with RV pacing during the 3-month observation period. Results are heterogeneous due to a large interpatient variability, suggesting the need for different methods of patient stratification. The favourable effect of BiV pacing in the subgroup of patients with preserved systolic function and narrow QRS is interesting but needs further validation. A longer observation period could have enhanced the adverse haemodynamic effect of non-physiological RV pacing, thus allowing the potential benefit of LV-based pacing therapy to become manifest. However, the study confirms that ventricular rate control and RV pacing is a very effective and attractive therapy for patients with permanent AF and refractory heart failure. 4,13 LV or BiV pacing cannot be recommended as a first line treatment for all patients with AF and should probably be restricted and delayed being offered to patients who have no benefit from RV pacing alone or have deterioration in their clinical condition late after ablation. Delayed BiV upgrading was very effective in a prospective uncontrolled study. 12 Alternatively, if the predictive value of tissue-doppler echocardiography or other imaging techniques is also confirmed in patients with AF as it has been in sinus rhythm, 27 these techniques could help provide a better stratification at the time of ablation and pacing therapy. Appendix Study organization: The OPSITE study is independent from industry. It is officially endorsed by the Working Group on Pacing of the European Society of Cardiology. The Steering Committee designed the study protocol as described in this document. All data were collected and processed by the Study Co-ordinating Centre in Lavagna. The Steering Committee was supported by a limited grant from Vitatron, The Netherlands, and St Jude Medical, Italy. Both companies agreed not to interfere with the scientific issues of the study. Steering committee: M. Brignole (co-chair), M. Gammage (co-chair), P. Alboni, A. Raviele, R. Sutton, P. Vardas. Executive committee: M. Brignole, M. Gammage. Data and statistical analysis: M. Brignole, M. Gammage, E. Puggioni. Participating centres and investigators (number of patients in brackets): Ospedali del Tigullio, Lavagna (17): Puggioni E, Lupi G, Brignole M; Ospedale S Chiara, Pisa (8): Soldati E, Bongiorni MG; Karolinska Hospital, Stockholm (7): Gadler F, Bergfeldt L; University Hospital, Eraklion (6): Simantirakis EN, Vardas P; Ospedale S Maria Nuova, Reggio Emilia (6): Tomasi C, Menozzi C; Ospedale Civile, Imperia (3): Mureddu R, Leoncini M, Musso G; Ospedale Umberto I, Mestre (3): Corrado A, Gasparini G, Raviele A; Queen Elizabeth Hospital, Birmingham (3): Gammage M; Ospedale Civile, Cento (2): Scarfò S, Alboni P; Ospedale S Pietro Igneo, Fucecchio (1): Del Rosso A. References 1. Zile M, Blaustein A, Shimizu G et al. Right ventricular pacing reduces the rate of left ventricular relaxation and filling. J Am Coll Cardiol 1987;10: Ausubel K, Furman S. The pacemaker syndrome. Ann Intern Med 1985;103: Sadaniantz A, Laurent L. Left ventricular doppler diastolic filling patterns in patients with isolated left bundle branch block. Am J Cardiol 1998;81:

11 722 M. Brignole et al. 4. Brignole M, Menozzi C, Gianfranchi L et al. Assessment of atrioventricular junction ablation and VVIR pacemaker versus pharmacological treatment in patients with heart failure and chronic atrial fibrillation. A randomized controlled study. Circulation 1998;98: Puggioni E, Brignole M, Gammage M et al. Acute comparative effect of right and left ventricular pacing in patients with permanent atrial fibrillation. J Am Coll Cardiol 2004;43: Rector TS, Kubo SH, Cohn JN. Patients self-assessment of their congestive heart failure. Part II: content, reliability and validity of a new measure, the Minnesota Living with Heart Failure Questionnaire. Heart Fail 1987;3: Brignole M, Gianfranchi L, Menozzi C et al. Influence of atrioventricular junction radiofrequency ablation in patients with chronic atrial fibrillation and flutter on quality of life and cardiac performance. Am J Cardiol 1994;74: Linde-Edelstam C, Norlander R, Uden A-L. Quality of life in patients treated with atrio-ventricular synchronous pacing compared to rate modulated ventricular pacing: a long term, double blind cross-over study. Pacing Clin Electrophysiol 1992;15: The Criteria Committee of the New York Heart Association. Nomenclature and Criteria for Diagnosis of Diseases of the Heart and Great Vessels. 9th ed. Boston, MA: Little, Brown & Co; 1994; p Opasich C, Pinna D, Mazza A et al. Reproducibility of the six-minute walking test in patients with chronic congestive heart failure: practical implications. Am J Cardiol 1998;81: Leclercq C, Victor F, Alonso C et al. Comparative effects of permanent biventricular pacing for refractory heart failure in patients with stable sinus rhythm or chronic atrial fibrillation. Am J Cardiol 2000;85: Leon A, Greenberg J, Kanaru N et al. Cardiac resynchronization in patients with congestive heart failure and chronic atrial fibrillation. J Am Coll Cardiol 2002;39: Stambler B, Ellenbogen K, Zhang X et al. ROVA investigators. Right ventricular outflow tract versus apical pacing in pacemaker patients with congestive heart failure and atrial fibrillation. J Cardiovasc Electrophysiol 2003;14: Leclercq C, Walker S, Linde C et al. Comparative effects of permanent biventricular and right-univentricular pacing in heart failure patients with chronic atrial fibrillation. Eur Heart J 2002; 23: Clark D, Plumb V, Epstein A et al. Hemodynamic effects of irregular sequences of ventricular cycle lengths during atrial fibrillation. J Am Coll Cardiol 1997;30: Daoud E, Weiss R, Bahu M et al. Effect of irregular ventricular rhythm on cardiac output. Am J Cardiol 1996;78: Herbert WH. Cardiac output and the varying RR interval of atrial fibrillation. J Electrocardiol 1973;6: Gosselink M, Blanksma P, Crijns H et al. Left ventricular beat-to-beat performance in atrial fibrillation: contribution of Frank Starling mechanism after short rather than long RR interval. J Am Coll Cardiol 1995;26: Wood M, Brown-Mahoney C, Kay GN et al. Clinical outcomes after ablation and pacing therapy for atrial fibrillation. Circulation 2000;101: Fitzpatrick AD, Kourouyan HD, Siu A et al. Quality of life and outcomes after radiofrequency His-bundle catheter ablation and permanent pacemaker implantation: impact of treatment in paroxysmal and estabilished atrial fibrillation. Am Heart J 1996;131: Rosenquist M, Lee M, Mouliner L et al. Long term follow-up of patients after transcatheter direct current ablation of the atrioventricular junction. J Am Coll Cardiol 1990;16: Cazeau S, Leclercq C, Lavergne T et al. Effects of multisite biventricular pacing in patients with heart failure and intraventricular conduction delay. N Engl J Med 2001;344: Auricchio A, Stellbrink C, Butter C et al. Clinical efficacy of cardiac resynchronization therapy using left ventricular pacing in heart failure patients stratified by severity of ventricular conduction delay. J Am Coll Cardiol 2003;42: Abraham W, Fisher W, Smith A et al. Cardiac resynchronization in chronic heart failure. N Engl J Med 2002;346: Young J, Abraham W, Smith A et al. Combined resynchronization and implantable cardioversion defibrillation in advanced chronic heart failure. JAMA 2003;289: Verbeek XA, Vernooy K, Peschar M et al. Intraventricular resynchronization for optimal left ventricular function during pacing in experimental left bundle branch block. J Am Coll Cardiol 2003; 42: Penicka M, Bartunek J, De Bruyne B et al. Improvement of left ventricular function after cardiac resynchronization therapy is predicted by tissue doppler imaging echocardiography. Circulation 2004; 109;