• 
• 

Three Part Question

In [patients undergoing cardiac surgery] are [prophylactic beta-blockers] of benefit in reducing the incidence of [post-operative AF]?

Clinical Scenario

You are seeing a 75 year old diabetic man with triple vessel disease, for whom you are going to perform triple vessel coronary arterial bypass tomorrow. He has been suffering with angina for the past 9 years but it has progressively got worse. His left ventricular ejection fraction is 45%
You note that he is on 25mg of atenolol pre-operatively and his heart rate is 80 today. He reports that he has had palpitations occasionally in the past but not recently. You feel that he is almost destined to go into Atrial Fibrillation post-operatively but there is wide variation in this hospital with regard to beta-blocker prophylaxis with some patients being changed to sotalol pre-operatively, some having atenolol started on the day after surgery, and many patients with any impairment of their ejection fraction having all beta-blockers withdrawn for at least 4 days. You wonder whether there is any consensus in the literature.

Search Strategy

Medline 1966-July 2004 using the OVID interface.
[exp cardiac surgery.mp OR bypass.mp OR CABG.mp OR exp coronary artery bypass OR cardiopulmonary bypass.mp OR exp cardiovascular surgical procedures] AND [exp adrenergic beta-antagonists OR exp beta-blockers OR beta-adrenoceptor blocker.mp OR prophylactic drug therapy.mp] AND [exp atrial fibrillation OR postoperative AF.mp OR postoperative atrial fibrillation.mp OR supraventricular arrhythmia.mp OR supraventricular arrhythmias.mp].

Search Outcome

A total of 113 papers were found from which 5 meta-analyses and systematic reviews represented the best evidence. In addition the American Heart Association guidelines were reviewed. Cross-checking reference lists and Journal Club suggestions provided an additional 2 papers. These papers are listed in the table.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Andrews et al 1991 USA	Meta-analysis of 18 RCTs addressing the efficacy of beta-blockers as prophylaxis after CABG, and 24 RCTs found looking at all drugs A total of 1,549 patients on beta-blockers were analyzed Subgroup analysis of preoperative treatment groups, postoperative treatment groups, patients treated with either low dose or high dose propranolol and patients treated with either propranolol or non-propranolol beta-blockers was also performed .	Meta-analysis (level 1a)	Incidence of supraventricular arrhythmia in the various study subgroups. All Beta-blocker patients.	8.7% . Control patients 34% (p <0.0001)	In this meta-analysis supraventricular arrhythmia included atrial fibrillation, atrial flutter, paroxysmal reentrant supraventricular tachycardia or paroxysmal atrial tachycardia. The patients in these studies were highly selected, predominantly young males with well preserved left ventricular function. Most beta-blocker trials either included patients on beta-blockers before surgery or excluded patients with ejection fractions below 30-50%. Trials also excluded patients with insulin dependent diabetes , history of AV block, sick sinus syndrome, bronchospastic lung disease or hypotension at the time of randomization.
			Pre-operative Beta-blockers	8.1% . Control patients 40.1% (p<0.0.0001)
			Post-operative beta-blockers	8.9%. Control patients 32.3% (P<0.0.0001)
			verapamil patients	18.2%. Control patients 18.2% (p =0.69)
			digoxin patients	14.2%. Control patients 17.6% (p =0.88).
Kowey et al 1992 USA	Meta-analysis of 7 randomized trials investigating the effectiveness of prophylactic beta-blockers in preventing supraventricular arrhythmia early after CABG. A total of 1418 patients were analyzed: 743 controls vs 675 patients who received beta-blocker therapy.	Meta-analysis (level 1a)	Supraventricular arrhythmia early after CABG	Beta-blocker patients 66/675 (9.8%). Control patients 150/743 (20.2%) P<0.001. Combination of Beta-blocker and digoxin 3/139 (2.2%) Control patients 45/153 (29.4%) P<0.001	Drug selection, dosing and monitoring were all variable in individual trial. Despite that most cases of postoperative SVA were due to AF, some were attributed to atrial flutter Variables that may have contributed to the results of individual trials were not taken into account. Data used in the study was 5 to 15 years old Mean age of pts 55, male to female 4.5:1
Eagle (American College of Cardiology Guidelines) 1999 USA	Systematic review of various issues regarding coronary artery bypass grafting. This report updated a previews review conducted in 1991	Systematic review Level 1a	Prevention of postoperative atrial fibrillation	Withdrawal of preoperative beta-blockers in the postoperative period doubles the risk of atrial fibrillation after CABG. Thus, early re-initiation of beta-blockers is critical for avoidance of this complication. Virtually every study of patients receiving beta-blockers prophylactically has shown benefit in lowering the frequency of atrial fibrillation. Most have used the drug in the postoperative period, but greater benefit may occur if beta-blockade is begun before the operation. Currently, the routine preoperative or early postoperative administration of beta-blockers is considered standard therapy to reduce the risk of atrial fibrillation after CABG.	Only highly selected papers referenced. No search strategy given. No level of evidence or grade of recommendation given.
Maisel et al 2001 USA	24 RCTS analysed	Systematic review (level 1a)	Reduction in AF	>75% reduction in AF with prohylactic B-blockade. Pre-operative beta-blocker therapy is more effective than post-operative therapy
Crystal et al 2002 USA	27 RCTs evaluating the role of beta-blockers in the prevention of postoperative AF were analyzed, involving a total of 3840 patients. 2 beta-blocker trials, including 1200 patients reported the effect on LOS	Meta-analysis (level 1a)	Primary outcome measure was the incidence of postoperative AF or atrial flutter. Two other outcome measures length of stay (LOS) and incidence of stroke were also analyzed.	All Beta-blocker patients 19%. Control patients 33%. (OR=0.39, P<0.00001). Sotalol patients 17% Control patients 37% (OR = 0.35 P< 0.00001). no significant heterogeneity between trials P=0.25. Sotalol patients 12% Other beta blockers 22% (OR, 0.50; 95% CI, 0.34 to 0.74). Amiodarone patients 22.5% control patients 37% (OR = 0.48 P< 0.00001). with no significant heterogeneity between trials P=0.55. Beta-blockers did not significantly reduced LOS (-0.66 day, 95% CI, -2.04 to 0.72). Incidence of stroke was 1.2% in all the treatment groups combined and 1.4% in controls (P=NS)	There was significant heterogeneity between individual beta-blocker trials (P=0.00001). However, analysis of the: role of specific beta-blockers used, individual sample size, the proportion of patients taking beta-blockers preoperatively, method of ECG monitoring, and source where the information for this meta-analysis was obtained from, revealed no reasons for heterogeneity. Individual studies included span 3 decades (1979-2001) Most studies excluded patients with impaired EF. Mean EF ranged from 43% to 68%.
Wurdeman et al 2002 USA	10 RCTs that studied either amiodarone (n=764) or sotalol (n=539) as Prophylaxis for AF (5 RCTs for each drug, no direct comparative studies) Mean EF ranged from 30% to 65%	meta-analysis (level 1a)	Incidence of AF	Sotalol group 21.5% reduction in AF compared to Placebo(p=<0.001). Amiodarone group 14.1% reduction in AF compared to Placebo(p=<0.001).	Neither drug reduced Length of Stay 3 sotalol studies were unblinded but only 1 amiodarone study was unblinded.
			Side effects requiring discontinuation	Sotalol group 9.7% (P=0.048). Amiodarone group 1.95% (P=NS)
Ferguson et al 2002 USA	The use of pre-operative B-blockers was assessed in the STS database from 1996 to 1999, among 629,877 patients from 497 sites for mortality and morbidity	Retrospective Cohort study (level 2b)	Unadjusted mortality	B-blocker patients 2.8%. Control patients 3.4%. (OR 0.8, CI 0.78-0.82)	6.4% of patients did not have any record of whether B-blockers were used. This is not a randomized trial and therefore caution should be used when interpreting the reasons for the benefits reported.
			Mortality adjusted for patient and centre risk factors	B-blockers had a lower mortality with OR 0.94, CI 0.91-0.97.
			Mortality for patients with EF <30%	B-blockers caused a higher mortality, OR 1.13, CI 0.96-1.33.
Zimmer et al 2003 USA	13 RCTs identified including 1,783 pts that specifically looked at prophylactic antiarrythmic studies that reported data on Length of stay (LOS) costs, CVA, or mortality.	Meta-analysis (level 1a)	Odds of AF	Odds of AF in all studies was reduced by 0.52(0.41 to 0.65)	65 trials excluded due to incomplete reporting of data Only one of the 13 RCTs looked at beta-blockers, 6 were in amiodarone, 5 were in atrial pacing, and one was in procainamide
			Length of stay	Length of stay was reduced by mean 1 day +/- 0.2 days in the treatment studies
			Costs	There was a mean $1,287 +/- $673 reduction in costs with prophylaxis
			Stroke or mortality	No differences in stroke or mortality between groups

Comment(s)

Andrews et al performed the first meta-analysis in this area. They found that 13 of 19 studies investigating the benefit of prophylactic beta-blockers showed a significant benefit in favour of giving prophylaxis. Pooling all these results showed a reduction in AF from 34% to 8.7% from studies involving 1,549 patients. Interestingly no difference was shown when pre-operative beta-blocker studies were compared to post-operative studies. No benefit was shown in 8 studies assessing either verapamil or digoxin as AF prophylaxis. They also showed that the mean ventricular rate was significantly lower in beta-blocked patients when they did go into AF, with a mean rate 24bpm slower than controls. They did caution that most patients in these studies were young, male and had good ejection fractions and had been on beta-blockers pre-operatively. Kowey et al in 1992 pooled data from 7 studies containing 1,418 patients, and found a reduction in AF from 20% to 9.8%. In addition they pooled data from 2 studies containing 292 patients that looked at prophylaxis with both Digoxin and Beta-blockers and concluded that combination therapy was better than beta-blockers alone with a p value of less than 0.01. Zimmer et al performed a meta-analysis of all anti-arrythmic strategies to look at the length of stay, costs, stroke and mortality. They found that there was an average 1 day less in hospital, and $1,300 less in costs using an anti-arrythmic prophylactic strategy, although they showed no difference in stroke or mortality. Although only 1 study was in beta-blockers, they justified extrapolation to all strategies on the basis that all strategies had the same endpoint of reduction of AF, and this meta-analysis quantifies for the first time the benefits of reducing the incidence of post-operative AF. Crystal et al in 2002 performed a comprehensive meta-analysis, including 52 RCTs on a variety of prophylactic therapies. They found that beta-blockers reduced AF from 33% to 19% from pooled data from 27 RCTs that recruited 3840 patients. This corresponds to a number needed to treat of 7. Pooling all strategies to reduce AF they found that hospital stay could be reduced by half a day, but no reduction in the incidence of stroke could be found. In sub-analyses they found that 4 trials with 900 patients compared Sotalol with other beta-blockers and found that sotalol significantly reduced the incidence of AF compared with other beta-blockers. They however caution that sotalol also has the potential to cause proarrythmic side effects despite this reduction in AF, but provide no figures to back this up. They also investigated the effect of either pre-operative or post-operative commencement of beta-blocker prophylaxis and found no difference between the two strategies. As a final note they found that both prophylactic amiodarone and biatrial pacing also significantly reduces the incidence of AF. Wurdemann et al in 2002 compared studies investigating sotalol with studies investigating amiodarone. They found no studies that directly compared both drugs. They found that sotalol reduced the incidence of AF by 21.5% compared to a reduction of 14% with amiodarone, but this difference between the drugs was not statistically significant. In addition a significantly higher percentage of patients receiving sotalol had their treatment stopped due to side-effects. No differences in length of stay were found. They concluded that both drugs were comparable in terms of their efficacy in reducing AF but amiodarone had fewer side-effects. Maisel et al performed a systematic review in 2001 and concluded that all patients should receive both pre-operative and post-operative beta-blockers prior to cardiac surgery, unless contraindicated, in which case amiodarone or biatrial pacing should be used. Of note they mainly quoted the meta-analysis from Andrews et al in making these recommendations. Ferguson et al performed a large retrospective analysis of the STS surgical database containing 629,877 patients to look at the mortality and morbidity associated with pre-operative beta-blocker use. This was not a randomized trial and therefore patients had not received beta-blockers randomly, so the authors used a propensity score for the risk of receiving a beta-blocker, derived from patient and centre associated risk factors. After adjustment they found that there was a slightly lower mortality in the pre-operative beta-blocker group, and a lower incidence of stroke ventilation and renal failure. They also found that patients with an EF less than 30% had a slightly increased mortality. This study was also not considering the effect of reduction in AF but only the benefit of Beta-blockers by any mechanism. The American Heart Association recommends that continuing beta-blockers and restarting them early post-op represents the optimal standard of care in coronary bypass graft surgery. In summary prophylactic beta-blockers clearly reduce the incidence of AF, from the results of 5 meta-analyses. In addition some benefits in terms of reducing length of stay, costs, mortality and morbidity have been shown in patients with a good ejection fraction, although the evidence for this is far less strong. In addition the benefits of pre-operative administration over early post-operative administration have not been clearly demonstrated or the relative benefits of any one beta-blocker over another.

Clinical Bottom Line

Prophylactic Beta-blockers clearly reduce the incidence of AF with a number needed to treat of only 7 to prevent one episode of AF. The optimal beta-blocker or the benefits to patients with impaired ejection fraction are less clear.

References

1. Andrews, TC, Reimold, SC, Berlin, JA, Antman E. Prevention of supraventricular arrhythmias after coronary artery bypass grafting. A meta-analysis of randomized control trials. Circulation, 1991;84 [SIII]; III-236-III-244.
2. Kowey, PR, Taylor, JE, Rials, SJ, Marinchak A. Meta-analysis of the effectiveness of prophylactic drug therapy in preventing supraventricular arrhythmia early after CABG. American Journal of Cardiology, 1992;69:963–965.
3. Eagle KA, Guyton RA, Davidoff R. American College of Cardiology guidelines for coronary artery bypass graft surgery. J Am Coll Cardiol, 1999;99:1262–346.
4. Maisel WH, Rawn JD, Stevenson WG. Atrial Fibrillation after Cardiac Surgery. Ann Intern Med, 2001;135:1061-1073.
5. Crystal E, Connolly SJ, Sleik K, Ginger TJ, Yusuf Salim. Interventions on prevention of postoperative atrial fibrillation in patients undergoing heart surgery. A meta-analysis. Circulation, 2002;106:75-80.
6. Wurdeman RL. Mooss AN. Mohiuddin SM. Lenz TL. Amiodarone vs. sotalol as prophylaxis against atrial fibrillation/flutter after heart surgery: a meta-analysis. Chest 2002; 121(4):1203-10.
7. Ferguson TB, Coombs LP, Peterson ED. Pre-operative B-blocker use and mortality and morbidity following CABG surgery in North America. JAMA 2002;287(17):2221-2227.
8. Zimmer J. Pezzullo J. Choucair W. Southard J. Kokkinos P. Karasik P. Greenberg MD. Singh SN. Meta-analysis of antiarrhythmic therapy in the prevention of postoperative atrial fibrillation and the effect on hospital length of stay, costs, cerebrovascular accidents, and mortality in patients .. American Journal of Cardiology 2003. 91(9):1137-40.