Bi-atrial pacing significantly reduces the Incidence of atrial fibrillation ...

ARTICLE IN PRESS doi:10.1510/icvts.2004.097121

Interactive CardioVascular and Thoracic Surgery 4 (2005) 33–40 www.icvts.org

Best evidence topic – Arrhythmia

Bi-atrial pacing significantly reduces the Incidence of atrial fibrillation post cardiac surgery Andrew Ronalda, Joel Dunningb,* a Aberdeen Royal Infirmary, Aberdeen, AB25 2ZN, UK Department of Cardiothoracic Surgery, Freeman Hospital, Freeman Road, Newcastle upon Tyre, NE7 7AZ, UK

b

Received 2 September 2004; accepted 3 September 2004

Summary A best evidence topic in cardiac surgery was written according to a structured protocol. The question addressed was whether either right atrial or bi-atrial pacing effectively reduces the incidence of Atrial fibrillation post cardiac surgery. Altogether 458 papers were found using the reported search, of which 16 represented the best evidence on this topic. The author, journal, date and country of publication, patient group studied, study type, relevant outcomes, results, and study weaknesses were tabulated. We conclude that Right atrial pacing is of no benefit but bi-atrial pacing significantly reduces the incidence of atrial fibrillation with an odds ratio for benefit of 0.51 (95%CI 0.36–0.72) from 11 studies. 䊚 2005 Published by European Association for Cardio-Thoracic Surgery. All rights reserved. Keywords: Thoracic surgery; Atrial fibrillation; Review; Meta-analysis; Pacing

1. Introduction A best evidence topic was constructed according to a structured protocol. This protocol is fully described in the ICVTS w1x. 2. Clinical scenario You are concerned to note that the incidence of postoperative atrial fibrillation in your unit is almost 40% after elective cardiac surgery. You have read a recent review that suggests that atrial pacing may protect patients against atrial fibrillation and as all patients receive right atrial wires in your unit intra-operatively this seems to be a simple opportunity to reduce the incidence of AF without the inherent complications of pharmacological prophylaxis. Thus you resolve to explore the literature further with a view to implementing a departmental policy for postoperative atrial pacing. 3. Three-part question In wpatients who have undergone cardiac surgeryx does watrial pacingx decrease the incidence of wpostoperative atrial fibrillationx? 4. Search strategy Medline 1966-August 2004 and EMBASE 1980 to August 2004 using the OVID interface. wexp Cardiac Pacing, Artifi*Corresponding author: Tel.: q44 780 154 8122. E-mail address: [email protected] (J. Dunning). 䊚 2005 Published by European Association for Cardio-Thoracic Surgery

cialy OR exp Pacemaker, Artificialy OR pacing.mpx AND watrial.mp or biatrial.mp or bi-atrial.mp or bachman$.mpx AND wAtrial fibrillation.mp OR exp Atrial Fibrillationy OR Atrial flutter.mp OR exp Atrial Fluttery OR AF.mp OR supraventricular tachycardia.mp OR exp Tachycardia, Supraventricularyx AND wCABG.mp OR Coronary art$ bypass.mp OR Cardiopulmonary bypass.mp OR exp Cardiovascular Surgical Proceduresy OR exp Cardiac Surgical Proceduresy OR exp Coronary Artery Bypassy OR Cardiac Surgery.mpx. 5. Search outcome A total of 229 papers were found in Medline and exactly 229 papers were also found in Embase of which 14 were relevant. An additional 2 papers were found by checking reference lists (Table 1). 6. Results Crystal et al. w2x performed a meta-analysis in 2002 that looked at pharmacological and pacing strategies for the reduction of AF after Cardiac Surgery. They identified 10 of the 13 completed trials that we identified by our search strategies. They found that Biatrial pacing significantly reduced the likelihood of AF with an Odds ratio of 0.46 (95%CI 0.30–0.71), which was a significant result. They also identified that right atrial and left atrial pacing reduced the odds of AF but that these results were not significant (RA pacing OR 0.68, 95%CI 0.39–1.19 and LA pacing OR 0.57, 95%CI 0.28–1.16). The reported studies varied markedly, however, in their protocols and pacing strategies, with

ARTICLE IN PRESS A. Ronald, J. Dunning / Interactive CardioVascular and Thoracic Surgery 4 (2005) 33–40

34 Table 1 Summary of best evidence papers Author, date and country

Patient group

Study type (level of evidence)

Outcomes

Key results

Comments

Crystal et al., 2002 Circulation, Canada w2x

10 trials found that investigated pacing in the prevention of AF, found from searching Medline, Embase, CINAHL up to April 2001

Metaanalysis (level 1a)

Odds reduction of biatrial pacing

Biatrial pacing reduces AF, OR 0.46 (95%CI 0.30–0.71) (744 pts enrolled)

4 studies reported below in the table were not included in this meta-analysis

Odds reduction of right atrial pacing

Right atrial pacing reduces AF, OR 0.68 (95%CI 0.39–1.19) (581 pts enrolled)

Odds reduction of left atrial pacing

Left atrial pacing reduces AF, OR 0.57 (95%CI 0.28–1.16) (148 pts enrolled)

Odds reduction for fixed and flexible pacing algorithms

Fixed rate OR 0.58 (95%CI 0.32–1.02)

trials found from 9 to 100 patients in treatment groups.

Flexible pacing algorithm OR 0.62 (95%CI 0.38–1.01) Debrunner M et al., 2004, EJCTS, Switzerland w3x

80 patients with no previous history of AF undergoing valve surgery " CABG

PRCT (level 1b)

At least 2 min of AF within first 72 h of surgery confirmed by 2 independent physicians

Control RA wires (lateral wall RA) ns40;

18y40 (45%) control developed AF within first 3 days compared with 8y40 (20%) of BAP group (Ps0.02) Differences more marked in subgroup analysis of 25 patients on preoperative beta-blockers – 1y14 paced versus 6y11 unpaced (Ps0.0088)

Biatrial synchronous pacing (BAP) – RA (lateral wall RA); q LA (between insertion of pulmonary veins) wires

52y80 male Not blinded Possible confounding effects of betablockade in 30% of patients

12y18 control patients and 4y8 ‘paced group’ who developed AF had persistent AF

Control group paced via RA for haemodynamic compromise if rate -80ymin BAP group paced using algorithm 10 bpm above intrinsic rate up to max 110ymin from immediately postop to 72 h postop

Complications

No complicationsysequelae related to wire placement

Use of antidysrhythmic therapy at discharge

All patient in sinus rhythm at discharge — 17y40 control and 7y40 ‘paced group’ discharged on antidysrhythmic therapy (Ps0.015)

occurrence of AFG1 min measured by telemetry

Of 161 patients, 21y50 controls (42%); 29y60 RA pacing (48%); 19y51 Bachmann bundle pacing (37%) developed AF (intention to treat analysis). Differences between groups not statistically significant.

TelemetryyHolter ECG monitoring for 72 h Goette et al., 1999, Pacing & Clinical Electrophysiology, Germany w4x

161 patients with no history of AF randomized pre-CPB

PRCT (level 1b)

50y161 control – RA pacing wire placed but not used;

137 patients analysed according to pacing actually received (24 patients withdrawn from study for clinical reasons). 11y39 (28%) control, 25y54 (47%) RA paced and 11y44 (25%) Bachman Bundle- paced groups developed AF.

60y161 paced via lateral wall RA wire; 51y161 paced at Bachman Bundle –(functional bi-atrial pacing) AAI pacing immediately after surgery in two actively paced groups for 5 days @ 96 bpm

Surgeon randomisation led to unequal group sizes 24 patients withdrawn from study for clinical reasons (11 control; 6RA paced; 7 Bachmanbundle paced) Fixed pacing rate – loss of pacing @ intrinsic rates )96 bpm 5 days pacing; @ 96 h monitoring

Differences between groups versus control not statistically significant although comparison of incidence of AF in paced groups was, favouring Bachman- bundle pacing (P-0.05) No sequelae related to pacing wires

Gerstenfeld et al., 2001, Journal of Interventional Cardiac Electrophysiology, USA w5x

118 patients randomised to biatrial pacing (BAP) for 96 h post-op 60 controls and 58 in BAP group Treatment and control group both given beta-blockers as tolerated

Active group paced in AAI mode at 100 bpm

PRCT (level 1b)

AF)10 min

Control group 21y60 35% BAP group 11y58 19% BAP (P-0.05) Incidence of AF in isolated AVR group 35% control versus 21% BAP (Ps0.08) Beneficial effect of pacing greatest in those over 70 years – 55% control versus 25% BAP developed AF (P-0.05)

(Continued on next page)


35

Table 1 (Continued) Author, date and country

Patient group


Outcomes

Key results

Comments

Levy et al., 2000, Circulation, UK w6x

130 CABG patients sinus rhythm. Biatrial group (ns65) 4 days DDD biatrial pacing at base rate 80 bpm using an atrial resynchronised algorithm

Unblinded PRCT (level 1b)

Episode of AF lasting ) 1 h on pacemaker Holter

Biatrial pacing decreased monitored AF from 25y65 (38.5%) in control group to 9y65 (13.8%) (Ps0.001)

Underpowered for ITUyhospital stay

Biatrial pacing decreased any monitored or clinical AF from 26y65 (40%) in control group to 10y65 (15.4%). (Ps0.001)

Control (ns65) (DDD pacing with base rate 30 bpm) Temporary atrial epicardial leads placed lateral wall RA and Bachmann Bundle LA during surgery Pacing box functioned as Holter and was interrogated at

Daoud et al., 2000, Circulation, USA w7x

118 patients undergoing open heart surgery Right atrial pacing at 45 bpm (RA-AAI; ns39), right atrial triggered pacing at a rate of G10 bpm above native rate but not - 85 bpm (RA-AAT; ns38),

Double Blind PRCT (Level 1b)

Biatrial pacing lead failure rate 20% All patients on betablockers had these withdrawn postoperatively.

Clinically detected AF

Biatrial pacing decreased clinically detected AF from 21y65 (32.3%) in control group to 7y65 (10.8%) (Ps0.002) during study period

ITU and hospital stay

Mean and median ICU and hospital stay NS, although those in control group who developed AF spent significantly more time in ITU and hospital than those who didn’t (P 0.001–0.05).

Postoperative complications

Serious complication rate decreased from 35 to 13 (Ps0.001) in control versus paced group

Postoperative AF lasting more than 5 min detected on Holter and telemetric monitoring

4y41 (10%) Bi-AAT group developed AF, compared to 11y39 (28%) RA-AAI group (Ps0.03 versus Bi-AAT) and 12y38 (32%) RA-AAT group (Ps0.01 versus Bi-AAT)

Pacing wire failure (60%RA and 80% LA electrodes had failed by 5th postop day)

No statistical difference between incidence of AF with RA-AAI and RA-AAT groups (Ps0.8)

Most patient betablocked

simultaneous right and left atrial triggered pacing at rate G 10 bpm above native rate but not - 85 bpm (Bi-AAT; ns41).

No complications related to presence of pacing wires.

Mixed group of patientsy 100y118 CABG patients Delay in commencing pacing at beginning of study

Controls (ns39) (RA-AAI) at an inhibition rate of 45 bpm Pacing achieved via temporary epicardial leads placed on anterior-superior RA and posterior-inferior LA between coronary sinus and right inferior pulmonary vein Pacing commenced within 12 h of surgery and continues until 24 h before discharge. Fan et al., 2000, Circulation, China w8x

132 post-CABG patients with no history of AF randomised to receive overdrive atrial pacing via temporary epicardial pacing wires placed at the right atrial appendage andyor roof of left atrium at surgery.

Four groups – 32 pts biatrial pacing (BiA); 33 pts left atrial pacing (LA); 36 pts right atrial pacing (RA); 31 pts no pacing (control)

PRCT (level 1b)

Incidence of postoperative atrial fibrillation lasting )10 min or requiring urgent intervention assessed by continuous telemetry and daily ECGs

BiA group 4y32 12.5%; left atrial group 12y33 36.4%; right Atrial group 12y36 33.30%; control group 13y31 41.9% (P-0.05 for BiA against all other groups)

Length of hospital stay

Length hospital stay significantly less for BiA group (7.0 " 1.4 days) compared to Control group (9.6"4.2 days) Ps0.003

Premature termination of pacing in 12 patients due to technical problems (raised pacing thresholds andyor poor sensing).Their data processed in an ‘intention-to-treat’ analysis but no information on ‘spread’across 3 active treatment groups 70% patients male

Pacing rate 90 bpm unless intrinsic rate higher than this in which case rate set at 10 bpm more. Max paced rate 120 bpm. Overdrive pacing continued for 5 days

Continued perio -perative use of beta-blockade in 54% of patients. 12% had pre-operative Beta-blockers withdrawn



36 Table 1 (Continued) Author, date and country

Patient group


Outcomes

Key results

Comments

Greenberg et al., 2000, Journal of the American College of Cardiology, USA w9x

154 patients undergoing cardiac surgery

Single blinded PRCT (level 1b)

Postoperative atrial fibrillation lasting )1 h or resulting in haemodynamic compromise requiring electrical or chemical intervention

RAP group 3y40 (8%); LAP group 7y35 (20%); BAP group 8y31 (26%); control group 18y48 (37.5%)

Technical limitations of BAP equipment used

Excluded if in AF preop or were receiving anti-dysrhythmic therapy RA (posterior aspect RA) and LA (lateral wall just inferior to origin of right superior pulmonary vein) temporary epicardial leads inserted at surgery

Incidence of AF decreased from 37.5% (control) to 17% (paced patients). P-0.005 RAP had 79% decrease in AF compared to control (Ps0.002) Post op AF decreased in LAP and BAP groups compared to control but not statistically significant

4 patient groups – 40y154 right atrial pacing (RAP); 35y154 left atrial pacing (LAP); 31y154 biatrial pacing (BAP); 48y154 control (no pacing unless clinically indicated-pulse - 50 bpm)

Length of hospital stay (surgery to discharge)

Length of hospital stay decreased by 22% control versus paced groups (Ps0.003); RAP versus control hosp stay Ps0.01 No sequelae related to placementy removal of wires

Paced commence as soon as possible post-op for 72 h using rate of 100 bpm, unless native rate was 80–89 (105 bpm) or over 90 bpm (110 bpm) Continuous ECG and telemetry for 96 h. ECGs reviewed by ’blinded’ medical staff

Beta blockers given to all patients so unable to separate effect of that from pacing – may have influenced overall rate 115y154 male; 88% patients undergoing CABG alone; remainder valves or combined procedure Difficulty maintaining pacing protocol for duration of study in all groups – RAP 10%, LAP 23%, BAP 33% – had problems maintaining pacing protocol. 19% of control patients also required therapeutic pacing for medical reasons Monitoring– used constant ECG moni -toring with alarms rather than Holter

Blommaert et al., 2000, Journal of the American College of Cardiology, Belgium w10x

96 consecutive post-CABG patients haemodynamically stable in sinus rhythm without anti-dysrhythmics on 2nd POD.

PRCT (level 1b)

Treatment group paced via temporary epicardial wires placed high on the right atrium utilising dynamic overdrive pacing algorithm for 24 h versus control (not paced)

Primary end- point was AF sustained for 15 min identified by Holter ECG.

Data from ECG monitoring system reviewed by Cardiologist

Dynamic pacing algorithm set to pace in AAI mode with rate of 80–125 bpm to stimulate at rate just above patient’s own intrinsic rate

Pacing group AF 5y48 (10%); control group AF 13y48 (27%) (Ps0.036)

Small groups; mainly male (73y96)

Duration of AF episodes comparable (Ps0.27)

Concomitant use of beta-blockers in )50% patients (but no statistical difference between groups)

Multivariate analysis showed this difference only relevant for patients with LV ejection fraction over 50% (rate of AF 6% in paced versus 26% in control group). No significant difference (Ps0.27) in duration of AF episodes between groups

Day 2 chosen as time when incidence of AF at its highest but short monitoring period used so long term benefits cannot be commented on

All patients received pacing wires to high RA near sinus node and RV routinely Chung et al., 2000, Annals of Thoracic Surgery, USA w11x

100 post-CABG patients not in AFyflutter at time of postop randomisation

PRCT (level 2b)

Incidence of new AFyflutter in first 4 days post-CABG requiring pharmacologicalyelectrical treatment.

AAI group, ns51 pacing at least 10 beatsymin more than resting native rate (paced rates 90–110 bpm) via temporary atrial and ventricular epicardial leads placed during surgery. Control group ns49 Control group could receive protective anti-bradycardia pacing in VVI mode at 60 bpm at discretion of attending physicians

Incidence of AFy flutter at 7 days.

Commenced 1st postop day and continued to 4th postop day.

Assessment of atrial ectopic activity by atrial premature

AAI group AF 13y51 (25.5%); control group 14y49 (28.6%) (Ps0.90) by 4th Atrial depolarization frequency significantly higher in control patients who developed AF compared to those who did not (Ps0.023) AAI group 14y51 (27.5%); control group 14y49 (28.6%) unpaced had developed AF by Day 7 (Ps0.90) Peak incidence of AF day 2 (non-paced) and day 3 (paced)

APD frequency significantly higher in paced patients on 1st and 2nd postop days, and overall (Ps0.041, 0.001

Lots of exclusions prior to randomisation –172 recruited preop but 72 excluded prior to randomisation due to haemodynamic instability, rhythm disturbance (including development of AF) and lead failure Multiple problems in pacing group–(1) Pacing in AAI mode did not satisfactorily lead to overdrive pacing in ll subjects (2) 5y13



37

Table 1 (Continued) Author, date and country

Patient group


Outcomes

Key results

Comments

depolarization (APD) frequency

and 0.0001, respectively) compared to controls

‘Paced’ subjects who developed AF by day 4 were not actually pacing at onset of AF on telemetry review (3) 13y51 paced patients did not complete protocol due to a number of reasons including failure of pacing system, sinus tachycardia )110 bpm and other unspecified medical reasons. Trend towards higher periopera -tive inotrope use in control group AF only considered end-point if required treatment

Gerstenfeld et al., 1999, Journal of the American College of Cardiology, USA w12x

61 post CABG patients who were in preop sinus rhythm Three groups – 21y61 no atrial pacing (NAP); 21y61 right atrial pacing (RAP); 19y91 biatrial pacing (BAP)

Single Blind PRCT (level 1b)

Onset of sustained AF or atrial flutter for more than 10 min

Control group 7y21 (33%) Right atrial pacing, 6y21–29% Biatrial pacing 7y19 (37%) (P)0.7) On-treatment analysis – NAP 36%, RAP 29%, BAP 33% ( P)0.4) Less AF in beta-blockedypaced sub groups – NAP 38%, RAP 15%, BAP 0% (P-0.05).

Ventricular, RA (high lateral wall near SA node) and LA (posterior surface between right superior and inferior pulmonary veins) temporary epicardial leads attached during surgery to all patients

Complications

In 3 patients loss of atrial sensing led to the pacemaker apparently inducing AF by pacing during atrial repolarisa tion

Small groups; 45y61 male Not all patients beta-blocked 4 control patients required pacing for clinical reasons Cross over of 1 BAP, 1 RAP and 1 NAP patient for clinical reasons Problem of fixed rate pacing in patients with fast native rates

RAP and BAP patients paced at lower rate of 100 bpm DDD for 96 h postop; NAP paced at 50 bpm via ventricular wires Holter q telemetry monitoring Kurz et al. 1999 Pacing & Clinical Electrophysiology zerland w13x

21 cardiac surgery patients randomised to biatrial synchronous pacing (BSP) or control

Aborted PRCT (level 4)

One wire on LA (posterior aspect between pulmonary veins) and 2 on RA (right atrial appendage). Wires placed at surgery Paced @ 10 bpm at rate more than underlying rate up to max 110 bpm in AAI mode for 72 h

Occurrence of AF)2 min in first 72 h postop

3y12 BSP completed study without developing AF

Interval between surgery and AF and duration of AF

6y12 BSP developed sensing failure due to deterioration of P wave amplitude leading to asynchronous atrial stimulation. 5y6 developed AF 2y12 withdrawn due to excessive diaphragmatic stimulation

Therapeutic interven tions required

1y12 withdrawn due to electrode dislocation causing LV stimulation 2y9 controls developed AF

Episodes of AF)30 min or requiring treatment

Paced group 17.9%; control group 33.9% (RR 0.53, P-0.0001)

Study aborted after only 21 patients of planned 200 due to excess incidence of AF in BSP groups (explanations above)

Control RA paced if clinically necessary Continuous rhythm monitoring Orr et al., 1999 Pacing & Clinical Electrophysiology, UK w14x

230 pts undergoing first-time CABG surgery with no prior history of AF Randomised to postop synchronised biatrial pacing via temporary epicardial wires placed on to posterior LA and RA during surgery or control AAI mode

PRCT (level 1b)

89% of ‘paced’ group successfully paced for 96 h

No details of exact pacing algorithm used No details of AF detection method

No excess haemorrhage, infection or other sequelae



38 Table 1 (Continued) Author, date and country

Patient group


Outcomes

Key results

Comments

PRCT (level 2b)

AF)10 min requiring electrical or pharmacological intervention

Atrial pacing 11y43 (34%); control group 11y43 (34%) (Ps1.0)

No details of AF detection methods or placement of wires.

Case control study, (level 3b)

Incidence of AF

Paced group 16y52 (30%) Control group 13y52 (25%)

This is not a PRCT but a case control study. Numbers are small

Treatment group paced within 4 h to 96 h post-op Schweikert et al., 1998, Journal of the American College of Cardiology, USA w15x this is an abstract only

Au et al., 2003, Asian Cardiovascular and Thoracic Annals, HongKong, w16x

86 patients randomised to atrial pacing or control Atrial pacing algorithm using dual chamber rate responsive pacemaker with automatic mode-switching and overdrive facility to drive rate 80–130 bpm 52 patients undergoing CABG. Biatrial group (ns52) had wire attached to RA appendage and second wire onto roof of left atrium behind aorta.

PsNS No differences in hospital stay or morbidity

Control group (ns52) was a non randomized matched group Pacing at AAI rate 90 bpm or 10 bpm above native rhythm up to 140 bpm White et al. for AFIST-II, 2003, Circulation, USA w17x

160 patients undergoing CABG or valve surgery. 2=2 factorial design comparing both Amiodarone and pacing at Bachmanns bundle.

PRCT (level 1b)

Onset of AF of more than 5 min or causing haemodynamic compromise.

Amiodarone groups 1050 mg Amiodarone i.v. given by 24 h infusion 6 h post-surgery, followed by 400 mg tds orally for 4 days

Pacing groups 20y73 (27%); control groups 29y87 (33%) Ps0.523

Amiodarone groups 17y77 (22%); control groups 32y83 (39%)

53.8% stopped active pacing for a proportion of the study Technical difficulties with pacing in 17% of patients

Ps0.037

Pacing groups epicardial wires placed at Bachmanns bundle. Pacing started AAI mode, 6 h post-op at 80 bpm rising to 110 bpm to keep rhythm above native rate. Pacing stopped if native rhythm above 100 bpm

definitions of AF from 1 min of AF to 1 h. In addition, the placing of the wires varied, and the pacing strategies from fixed rates to complex flexible algorithms were used. A more recent review in this area identified 13 studies but did not perform an update of the meta-analysis w18x. We therefore elected to include all the individual trials in this topic so that all these various strategies could be compared. Debrunner in 2004 w3x studied 80 patients undergoing valve surgery with or without CABG. Patients were randomized to Biatrial pacing with an algorithm to keep pacing )10 bpm over the intrinsic rhythm for 3 days. Control patients received right atrial pacing with pacing set to 80 bpm. They demonstrated a reduction in AF from 45% to 20% in the biatrial pacing group, although the administration of beta-blockers was not controlled in this study, and a large number of patients had beta-blockers withdrawn post-operatively. Goette et al. w4x randomized 161 patients with a history of AF undergoing cardiopulmonary bypass. They randomized the patients into 3 groups, controls who had right atrial

pacing, which was only used if clinically indicated, a right atrial pacing group with active pacing for 5 days and biatrial pacing with wires placed at Bachmann’s Bundle and active pacing used for 5 days. They found no statistically significant results although 24 patients were withdrawn from the study for clinical reasons. Gerstenfeld et al. published 2 studies in 1999 and 2001 w5,12x, studying Biatrial pacing, right atrial pacing and controls in 61 patients, and later just comparing biatrial pacing with controls in 188 patients. In the smaller study no significant differences were found although there were only 6–7 occurrences of AF in each group. In their second larger study, the incidence of AF in the control group was 35% but in the biatrial pacing group the incidence was only 19%. On further analysis this difference was attributable only to patients over 70 years of age. Levy et al. w6x performed a large study in 130 patients undergoing first time CABG. Patients were randomized to biatrial pacing with wires in the right atrium and a second pair of wires at Bachmann’s Bundle, set to pace at 80 bpm.


The control group had a rate of AF of 40% but the biatrial paced group had an incidence of only 15%. This was significant for both monitored and clinically detected AF. Unfortunately, the study protocol required all patients on beta-blockers pre-operatively to have these withdrawn post-operatively. Daoud published a study in Circulation in 2000 w7x that compared control right atrial pacing, right atrial pacing at 85 bpm or 10 bpm above the intrinsic rhythm or biatrial pacing in a double blind fashion. The control group had an incidence of 28% and the right atrial pacing group had an incidence of 32% but the biatrial pacing group had an incidence of only 10%. This was a statistically significant finding. Of note 60% of right atrial wires and 80% of left atrial wires failed by the 5th post-operative day. Fan et al. in 2000 published a study in Circulation w8x that randomized 137 patients to 4 groups, Biatrial, Right atrial, left atrial and a control group. The protocol was a fixed rate of 90 pbm with the rate increased to 10 bpm above the underlying rhythm up to 120 bpm for 5 days. They found that the incidence of AF in the biatrial pacing group was 12.5% but the incidence in the RA, LA and control groups were 36%, 33% and 42%, respectively. Thus they concluded that Biatrial pacing was significantly superior to the other 3 strategies. In addition they found that adequate pacing was possible in all patients for the full 5-day duration of the study. Greenberg et al. w9x studied 154 patients, randomizing them to right atrial, left atrial, biatrial pacing and a control group. Pacing was set to 100 bpm, if the native rhythm was over 80, rate was increased to either 105 or 110 bpm, for 3 days. Assessment of ECG recordings was by blinded cardiologists. They found that the incidence of AF was 8% in the right atrial pacing group, which was significantly lower than left atrial pacing 20%; Biatrial pacing 26% or control 37.5%. Unfortunately, they had considerable problems with the left atrial and biatrial pacing, with 23% and 33% of patients unable to maintain pacing either due to diaphragmatic pacing or high thresholds. Blommaert et al. w10x investigated 96 patients undergoing CABG, randomized to a control group or a right atrial wire group. They used a novel programmed dynamic pacing strategy where the pacemaker had a lower rate of 80 bpm but if the native rhythm rose above this, the pacemaker automatically increased the rate up to a rate of 125 bpm, but kept the rhythm just above that of the native rhythm. This strategy was started on day 2 and continued for 24 h. The control group had an incidence of AF of 27% compared to an incidence of 10% in the pacing group (Ps0.036). Chung et al. w11x studied 100 patients who were at least 6 h post elective CABG. Forty-nine patients received AAI pacing at 90 bpm or 10 bpm above the native rhythm up to a rate of 110 bpm for 4 days. They found that the rate of AF was 26% for the AF group and the incidence in the control group was 29%, which was a non-significant finding. In addition they found that there was a significant increase in the atrial ectopic frequency in the paced group. This study had several problems in the pacing group. Eleven patients did not have successful overdrive pacing, and 5 of the 13 patients who went into AF in this group were not actually receiving pacing at the onset of AF.

39

In 1999 Kurz et al. w13x set out to perform a randomized controlled trial in 200 patients, randomized to biatrial pacing with a single wire in the left atrium and two wires on the right atrium. However, they had considerable problems in the pacing group mainly due to sensing failure in 50% of studied patients. In 5 or the 6 patients with pacing failure, this induced atrial fibrillation. In addition 2 patients were withdrawn due to excessive diaphragmatic stimulation and one withdrawn as a wire dislocated and started to cause ventricular stimulation. Schweikert published an abstract in the Journal of the American College of Cardiology w15x. They used atrial pacing with advanced overdrive pacing capabilities to study 86 patients undergoing CABG. They found that there were 11 patients in each group that developed AF and thus concluded that right atrial pacing does not prevent AF. Orr et al. w14x performed a study in 230 patients, randomized to biatrial pacing or controls. They found that the incidence of AF was 17.9% in the biatrial pacing group compared to an incidence of 33.9% in the control group, which was a highly significant result. Au et al. w16x performed a small case-control study that showed no difference between biatrial pacing and controls in two groups of 52 patients. This study was however very small and non-randomized. The AFIST-II trial w17x performed a 2=2 factorial design study in 160 patients looking at both post-operative Amiodarone prophylaxis and also Atrial Pacing at Bachmann’s Bundle. They used a pacing rate of 80 bpm increasing to 110 bpm, but pacing was stopped if the native rhythm increased to above 100 bpm. While they showed a significant improvement in AF for amiodarone, they showed no benefit for atrial pacing. The Pacing groups had an incidence of 27% and the control groups had an incidence of 33%. They also had many problems with the pacing algorithm, as 54% of patients had pacing stopped for a period during the trial including 17% stopped due to technical difficulties with the wires. Thus, in summary, of the 11 biatrial pacing studies (including 2 that used Bachmann’s Bundle pacing), 6 found significant benefit and 5 found no significant benefit. We combined these results using the DerSimonian and Laird Random Effects Model (Fig. 1a and 1b). This showed that there was a significant benefit to biatrial pacing with an odds ratio of 0.51 (95%CI 0.36–0.72). Of the 8 right atrial pacing studies, 2 found a significant benefit and 6 found no benefit. When the results were again combined by metaanalysis no benefit was found. While there is a significant benefit to biatrial pacing, several of the papers reported technical difficulties, with loss of sensing, diaphragmatic pacing and LV pacing which led to a number of patients being withdrawn from their respective studies. Thus, if biatrial pacing is used, much care must be used when placing the wires. In addition there were many different algorithms for pacing, although most seemed to pace at 80–90 bpm, raising this higher if the native rhythm went above 80 bpm. Also the number of days that pacing was used varied. AF incidence generally peaked around day 2, thus 3–5 days of pacing may be prudent.

ARTICLE IN PRESS 40

A. Ronald, J. Dunning / Interactive CardioVascular and Thoracic Surgery 4 (2005) 33–40

w4 x

w5 x

w6 x

w7 x

w8 x

w9 x

w10x

w11x

w12x

Fig. 1. (a) Forest plot of Biatrial pacing for the prevention of AF. (b) Forest Plot of Right Atrial Pacing for prevention of AF.

w13x

w14x

7. Clinical bottom line w15x

Right atrial pacing is of no benefit but biatrial pacing significantly reduces the incidence of Atrial fibrillation. References w1x Dunning J, Prendergast B, Mackway-Jones K. Towards evidence-based medicine in cardiothoracic surgery: best BETS. Interact Cardiovasc Thorac Surg 2003;2:405–9. w2x Crystal E, Connolly SJ, Sleik K, Ginger TJ, Yusuf S. Interventions on prevention of postoperative Atrial Fibrillation in Patients Undergoing Heart Surgery: A Meta-analysis. Circulation 2002;106:75–80. w3x Debrunner M. Naegeli B. Genoni M. Turina M. Bertel O. Prevention of atrial fibrillation after cardiac valvular surgery by epicardial, biatrial

w16x

w17x

w18x

synchronous pacing. wClinical Trial. Journal Article. Randomized Controlled Trialx European Journal of Cardio-Thoracic Surgery 2004:25(1):16–20. Goette A. Mittag J. Friedl A. Busk H. Jepsen MS. Hartung WM. Huth C. Klein HU. Pacing of Bachmann’s bundle after coronary artery bypass grafting. wClinical Trial. Journal Article. Randomized Controlled Trialx Pacing & Clinical Electrophysiology 2002:25(7):1072–8. Gerstenfeld EP. Khoo M. Martin RC. Cook JR. Lancey R. Rofino K. Vander Salm TJ. Mittleman RS. Effectiveness of bi-atrial pacing for reducing atrial fibrillation after coronary artery bypass graft surgery. Journal of Interventional Cardiac Electrophysiology 2001:5(3):275–83. Levy T. Fotopoulos G. Walker S. Rex S. Octave M. Paul V. Amrani M. Randomized controlled study investigating the effect of biatrial pacing in prevention of atrial fibrillation after coronary artery bypass grafting. Circulation 2000:102(12):1382–7. Daoud EG. Dabir R. Archambeau M. Morady F. Strickberger SA. Randomized, double-blind trial of simultaneous right and left atrial epicardial pacing for prevention of post-open heart surgery atrial fibrillation. Circulation 2000:102(7):761–5. Fan K. Lee KL. Chiu CS. Lee JW. He GW. Cheung D. Sun MP. Lau CP. Effects of biatrial pacing in prevention of postoperative atrial fibrillation after coronary artery bypass surgery. Circulation 2000:102(7):755– 60. Greenberg MD. Katz NM. Iuliano S. Tempesta BJ. Solomon AJ. Atrial pacing for the prevention of atrial fibrillation after cardiovascular surgery. Journal of the American College of Cardiology 2000:35(6):1416–22. Blommaert D. Gonzalez M. Mucumbitsi J. Gurne O. Evrard P. Buche M. Louagie Y. Eucher P. Jamart J. Installe E. De Roy L. Effective prevention of atrial fibrillation by continuous atrial overdrive pacing after coronary artery bypass surgery. Journal of the American College of Cardiology 2000:35(6):1411–5. Chung MK Augostini RS Asher CR Pool DP Grady TA Zikri M Buehner SM Weinstock M McCarthy PM. Ineffectiveness and potential proarrhythmia of atrial pacing for atrial fibrillation prevention after coronary artery bypass grafting. Annals of Thoracic Surgery 2000: 69(4):1057–63. Gerstenfeld EP Hill MR French SN Mehra R Rofino K Vander Salm TJ Mittleman RS. Evaluation of right atrial and biatrial temporary pacing for the prevention of atrial fibrillation after coronary artery bypass surgery, Journal of the American College of Cardiology 1999:33(7):1981–8. Kurz DJ Naegeli B Kunz M Genoni M Niederhauser U Bertel O. Epicardial, biatrial synchronous pacing for prevention of atrial fibrillation after cardiac surgery. Pacing & Clinical Electrophysiology 1999:22(5):721–6. Orr W, Tsui SSL. Synchronised biatrial pacing after coronary artery bypass surgery Pacingand Clinical Electrophysiology 1999;22:755. Schweikert RA, Grady TA. Atrial pacing in the prevention of atrial fibrillation after cardiac surgery: results of the 2nd postoperative pacing study (POPS-2). Journal of the American College of Cardiology 1998;31:117A wAbstractx. Au WK, Chiu SW, Sun MP, Cheung LC, Cheng LC. Biatrial Pacing to prevent Atrial Fibrillation after Coronary Artery Bypass, Asian Cardiovascular and Thoracic Annals 2003:11;222–225. White CM, Caron MF, Kalus JS, Rose H, Song J, Reddy P, Gallagher R, Kluger J. Atrial fibrillation Suppression Trial II. Intravenous plus oral amiodarone, atrial septal pacing, or both strategies to prevent postcardiothoracic surgery atrial fibrillation, (AFIST-II). Circulation, 108 Suppl 1: II200–6. Archbold RA, Schilling RJ. Atrial Pacing for the prevention of atrial fibrillation after coronary artery bypass graft surgery: A review of the literature. (Review) Heart 2004;90:129–33.