Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Is there a role for prophylaxis against atrial fibrillation for patients undergoing lung surgery?

Three Part Question

In [patients undergoing lung resection] is there a role for [prophylactic anti-arrhythmics] to reduce the incidence of [post-operative atrial fibrillation]?

Clinical Scenario

You are the thoracic registrar assessing a patient scheduled for lobectomy or possibly pneumonectomy the following day. Whilst explaining the risks of the operation you mention the risk of developing post-operative atrial fibrillation (AF), which increases with the extent of lung resected. You tell him that the reported incidence of AF post lung resection varies, with the lowest after a wedge resection (2-4%) increasing to 10-15% after lobectomy and 20-30% post-pneumonectomy. The patient asks whether you will be giving him any drugs to stop him getting AF. Your consultant never uses prophylaxis against AF for thoracic procedures but you wonder whether you should be giving something, as all your cardiac patients get Beta-blockers.

Search Strategy

Medline 1966-Jul 2004 and EMBASE 1980-Jul 2004 using the OVID interface.
[ OR exp pneumonectomy/ OR OR lung OR Thoracic OR lung OR pulmonary] AND [atrial OR exp Atrial Fibrillation/ OR Atrial OR OR exp Atrial flutter/ OR exp Tachycardia, Supraventricular/ OR OR arrhythmia$.mp]

Search Outcome

289 papers were found in MEDLINE and 168 papers were found in EMBASE of which 14 papers were deemed to be relevant. These are presented in the table.

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
De Decker et al,
Medline search for all papers on complications after non-cardiac thoracic surgery. Search strategy : Medline, and hand searching reference lists of papers. Double abstracting performedSystematic Review (level 1a)Efficacy of digoxin3 RCTs found, no evidence for reduction of AFEMBASE, SIGLE, CENTRAL, and many other widely available databases not searched No expert group Search not specific to atrial fibrillation Some earlier studies missed Table 2 incorrectly categorized Borgeat et al 1999 as a flecainide versus digoxin study, and felt that the Ritchie papers were in different patient groups
Efficacy of flecainide2 RCTs found, evidence for reduction in AF but only 30 patients in each study
Efficacy of B-blockade2 RCTS found reduction in AF , but one study was non significant
Efficacy of amiodarone11% incidence of ARDS in pts having prophylactic Amiodarone in only RCT set up to look at this
Efficacy of verapamil2 RCTs but high rate of complications
Efficacy of diltiazem1 RCT showed reduction in AF
Efficacy of magnesium1 RCT showing reduction in AF
Lanza et al,
Retrospective analysis of pulmonary resections over a 3 year period. One surgeon always used Amiodarone 200mg tds orally as AF prophylaxis during hospital stay,other surgeons did not All had continuous cardiac monitoring during inpatient stayRetrospective cohort study (level 2b)Incidence of AFAmiodarone group 3/31 (9.7%)
Control group 17/52 (33%)
No group differences found pre-operatively AF was not tightly defined, but recorded as absent or present from surgical records
ComplicationsNo ARDS or excess of complications in amiodarone group
Oka et al,
50 patients undergoing elective pulmonary resection at a single centre Bupivacaine group (n=23): Epidural bolus of 6-10ml of 0.25% bupivacaine then 3-5ml/hr for 3 days Morphine group (n=25): 2-3mg epidural bolus, then 0.2mg/hr Constant central monitoring for 3 daysDouble blind PRCT (level 1b)Post-operative tachyarrhythmia, defined as non-AF above 100bpm for >60 minsBupivacaine group 1/23 (4.3%)
Morphine group 7/25 (28%)
Patients not analysed by intention to treat, 2 excluded after becoming hypotensive
Complications2 bupivacaine patients withdrawn due to hypotension
11 of 25 pts receiving morhine reported pruritis
Amar et al,
330 adult patients undergoing pneumonectomy or over 60 and undergoing lobectomy in single institution Diltiazem group (n=167): Post operatively: 0.25mg/kg over 30 mins, then 0.1mg/kg/hr for 24 hrs , then diltiazem SR 120mg od oral. for 14 days Placebo group (n=163)Double blind PRCT (level 1b)Incidence of AF either >15 mins or requiring intervention
Patients received 3-4 days of holter monitoring
Diltiazem 25/167 (14%)
Placebo group 40/163 (26%)
p = 0.03
Good study design Power/sample size calculations. Samples stratified to ensure equal distribution of lobectomy and pneumonectomy between the groups 503 pts enrolled, 135 underwent lesser operation, 32 inoperable
Length of stayDiltiazem group 8.9+/-7.9 days
Placebo group 8.7+/- 7.8 days
Complications6 diltiazem and 1 placebo pts had interruption due to hypotension
Bayliff et al,
99 patients undergoing pneumonectomy, lobectomy, or oesophagectomy Propranolol group (n=49) 10mg oral 6hrly, from pre-op to 5 days post op Placebo group (n=50) placebo tablet Holter monitoring for 72 hours post-opDouble blind PRCT (level 1b)Incidence of any arrhythmiaPropranolol group 36/49 (72%)
Placebo group 31/50 (62%)
242 patients screened for eligibility NNT 7 to prevent one treated arrhythmia Results non-significant. There was a lower event rate than the sample size had been calculated for
Incidence of treated arrhythmiaPropranolol group 3/49 (6%)
Placebo group 10/50 (20%)
ComplicationsPropranolol group, hypotension 49% bradycardia 25%, pulm oedema 16%
Placebo group, hypotension 26%, bradycardia 4%, pulm oedema 8%
Jakobsen et al,
30 adult patients who had pneumonectomy, lobectomy or bi-lobectomy for lung cancer Metoprolol group (n=15) metoprolol 100mg pre-operatively and once daily Placebo group (n=15) Every patient received thoracic epiduralDouble blinded PRCT (level 2b)Incidence of AF lasting more than 30 seconds
Patients received 4 days of holter monitoring
Metoprolol group 1/15 (6.7%)
Placebo group 6/15 (40%)
p < 0.05
Small study population, 40 patients recruited, 10 withdrew or had a complication and was withdrawn. No intention to treat analysis performed No sample size calculation Holter monitoring was not universally used and only 3 pts had AF diagnosed on this, the remainder relied on clinical diagnosis or ECG
Complications1 patient on metoprolol developed pulmonary oedema
Amar et al,
70 adult patients undergoing pneumonectomy at single centre Digoxin group (n=35) 500, 250, 250 mcg iv first day post op 4 hrly, then 125-250mcg/day for 1 month Diltiazem group (n=35) 20mg iv, post-op, then 10mg iv qds for 1-2days then 180-240mg oral od for 1 monthUnblinded PRCT (level 2b)Incidence of AF lasting over 15 minutes
Patients received 3 days of holter monitoring
Digoxin group 11/35 (31%)
Diltiazem group 5/35 (14%)
(p = 0.09)
A control cohort with no Rx 11/40 (28%)
137 pts initially entered into the study. 67 excluded after inoperability or smaller operation performed Diltiazem dose of 10mg qds is lower than the therapeutic dose of 5-15mg/kg/hr
Complications2 diltiazem pts had treatment halted due to low blood pressure. 2 pts had dose dropped as outpatient
1 digoxin patient developed 2nd degree heart block
Terzi et al,
194 patients undergoing non-cardiac thoracic surgery Magnesium group (n=93) 2g (16mEq) MgSO4 in 100ml 5% Dext iv over 20 mins on thoracotomy and second dose after 6 hours, then 2g oral Mg for 2 days Control group (n=101) Given digoxin 500mcg oral post op if over 70 or post pneumonectomy Continuous ECG monitoring for 24-48hrs post-opUnblinded PRCT (level 2b)Incidence of atrial tachyarrythmiaMg group 10/93 (11%)
Placebo group 27/101 (27%)
There were 23 vs 18 pneumonectomies and 51 vs 40 lobectomies in the control group vs treatment group There were 5 more wedge resections in the magnesium group 6 patients removed from the study due to low pre-op Mg levels or emergency re-thoracotomy post-op Definition of atrial tachyarrhythmia and AF not described
Incidence of AFMg group 9/93 (10%)
Placebo group 23/101 (23%)
P value not given
ComplicationsNo hypotension or bradycardia in magnesium group
Van Miegham et al,
199 patients undergoing elective pneumonectomy or lobectomy Verapamil group (n=100) 10mg iv over 2 mins 1hr post op then 375mcg/min for 30 mins, then 125mcg.min for 3 days Control group (n=99)Unblinded PRCT (level 1b)Incidence of AF
Continuous ECG monitoring on an ICU over 3 days
Verapamil group 8/100 (8%)
Control group 15/99 (15%)
No sample size calculations. This was originally a 3 arm trial also including amiodarone, but there was an unacceptable incidence of ARDS with amiodarone and this arm was discontinued early
Complications9 verapamil pts had infusion stopped due to bradycardia and 14 stopped due to hypotension
Van Meigham et al,
Retrospective review of 242 pneumonectomies and 310 lobectomies, after their PRCT was halted They were evaluating amiodarone 150mg iv over 2 mins then 1,200mg/24hrs iv for 3 days.Retrospective cohort study (level 2b)Complications from the RCT3 patients out of 32 who received Amiodarone developed ARDS, two diedHigh doses of amiodarone were used in this study
Complications from cohort studyAmiodarone group 6/55 (11%) got ARDS
Non amiodarone patients 9/497 (1.8%) got ARDS
Ritchie et al,
111 patients undergoing elective thoracotomy, and pneumonectomy, lobectomy or open/close Digoxin group (n=58) 500mcg 6pm and 10pm night pre-op, then 250mcg with pre-med. 250mcg oral post-op for 9 days Control group (unblinded) n=53 4 days of constant on screen ECG then daily ECG alternate daysUnblinded PRCT (level 2b)Incidence of arrhythmiaDigoxin group 29/58 (50%)
Control group 19/53 (36%)
This is clearly the Annals study from 1990 except without the oesophagectomy patients. However a few extra patients have turned up in the digoxin groups, which is difficult to explain
Incidence of AFDigoxin group 14/58 (24%)
Control group 12/53 (23%)
ComplicationsOne death due to asystolic arrest on day 4
Borgeat et al,
30 patients undergoing thoracotomy and operations including lobectomy, pneumonectomy, segmentectomy Flecainide group (N=15) Post-op 2mg/kg over 60mins, then 0.15mg/kg/hr for 72 hrs Digoxin group (n=15) 10mcg/kg over 12 hours then 250mcg/24hr iv Holter monitoring for 72hrsSingle blind PRCT (level 2b)Occurrence of SVT for more than 10 beats or complex ventricular arrhythmiaFlecainide group 1/15 (7%)
Placebo group 7/15 (47%)
Different definition of SVT compared to their 1989 paper. Rate above 140 here but above 100 in previous paper
Occurrence of AFFlecainide group 1/15 (7%)
Placebo group 2/15 (14%)
ComplicationsNone reported
Ritchie et al,
140 patients undergoing elective Thoracotomy, and pneumonectomy, lobectomy, oesophogastrectomy, or open/close Digoxin group (n=64) 500mcg 6pm and 10pm night pre-op. then 250mcg with pre-med. 250mcg oral post-op for 9 days Control group (unblinded) n=66 4 days of constant on screen ECG then daily ECG alternate daysUnblinded PRCT (level 2b)Incidence of arrhythmia (duration of arrhythmia to qualify not reported)Digoxin group 29/64 (45%)
Control group 24/66 (36%)
Small study, only 73 patients underwent thoracoctomy for lung resection (lobectomy or pneumonectomy) No sample size calculation Numbers do not add up. In table 1 total number of patients is 130 but in text 140 patients were included in the study. No definitions for duration of arrhythmias that qualified for inclusion as a positive outcome
Incidence of AF or flutter (duration not reported)Digoxin group 18/64 (28%)
Control group 15/66 (23%)
Complications11.5% of patients had Digoxin levels above therapeutic range. One death due to asystolic arrest on day 4
Borgeat et al,
30 patients undergoing Non-cardiac thoracic surgery, including pneumonectomy, lobectomy, decortication. Flecainide group (n=14) Post-op 2mg/kg over 60mins, then 0.15mg/kg/hr for 72 hrs Placebo group (n=16) normal saline infusions Holter monitoring for 72hrsSingle blind PRCT (level 2b)Treatment failure defined as need to start new anti-arrhythmic drug, or dose increaseFlecainide group 0/14
Placebo group 6/16 (38%)
Poorly tolerated AFFlecainide group 0/14
Placebo group 3/16 (18%)
ComplicationsNo side effects reported


Randomized controlled trials were identified that studied the prophylactic effects of diltiazem, B-Blockers, digoxin, verapamil, flecainide, amiodarone, magnesium and epidural anaesthesia. However both the quality and sizes of these studies varied greatly, ranging from 330 patients to only 30 patients. The strongest paper was by Amar et al (200) that evaluated prophylactic diltiazem. They demonstrated a significant reduction in post-operative AF in lung resection patients with the incidence of 26% in control patients reduced to 14% in diltiazem patients. Furthermore, using a loading dose of 0.25mg/kg of diltiazem administered over 30 min followed by a continuous infusion (0.1 mg/kg/hr) for 18-24hr there were no significant increased adverse reactions to drug treatment in this study. This well conducted study gives a number needed to treat of 8 to prevent one episode of AF. This group also published a paper in 1997 comparing diltiazem with digoxin(Amar et al, 1997) in 70 patients. The incidence of AF reduced from 31% to 14% but due to the small sample size this was not a significant finding (p=0.09). Digoxin has also been studied in comparison to placebo. Ritchie et al peformed a study in 1990 in 140 patients undergoing pneumonectomy, lobectomy, or oesophagectomy, using an unblinded design. The incidence of AF was 36% in the control group and 45% in the digoxin group. They published 2 further papers but these are likely to be subset analyses of this first paper (Ritchie AJ, 1992; Ritchie AJ, 1993). Thus it seems that digoxin has no role in the prophylaxis of AF in lung surgery. A second large study was performed by Van Mieghem et al in 1996 patients in an unblinded PRCT. They evaluated prophylactic verapamil and found that the incidence of AF reduced from 15% to 8%. Unfortunately this was a non-significant finding and more significantly 23% of patients experienced bradycardia or hypotension as a result of verapamil and had to be withdrawn from the study. Interestingly Van Mieghem's study was originally a three-arm trial of verapamil, amiodarone and controls. However it was stopped early after 3 patients in the amiodarone group developed Adult Respiratory Distress Syndrome (ARDS), with 2 patients dieing of this complication. They then performed a retrospective review of 552 lobectomies and pneumonectomies(Van Mieghem W, 1996). 55 of these patients received amiodarone and 6 got ARDS (11%) but of 497 patients who did not receive amiodarone only 9 (1.8%) got ARDS. In contrast to this alarming finding with amiodarone, Lanza et al in 2003 performed a retrospective cohort study of their database, as one of their surgeons always gave 200mg tds orally of amiodarone to his patients while in hospital. They found that the control group had an incidence of 33% but only 10% of the amiodarone group got AF. They reported no complications in this study, but it was retrospective, and small. In addition the dose of amiodarone that caused ARDS was much larger in the Van Meighem study, with 1,200mg/24hrs being given iv for 3 days. Therefore while there is a possibility that Amiodarone may reduce AF if given prophylactically, there is a real risk that it will also cause ARDS and thus cannot be recommended at the present time. Flecainide was examined in two studies by Borgeat (1989, 1991), with 30 patients in each study. Both studies showed reductions from 40% to under 10% in AF, but as there were only 15 patients in each treatment group these findings were on the margins of significance. Beta-blockers have also been studied. Jakobsen et al (1997) performed a study with 15 patients randomized to metoprolol and 15 controls. They showed a significant reduction from 40% down to 7%. Bayliff et al (1999) randomized 49 patients to propranolol and 50 to controls. They unfortunately only showed some trends towards significance and the propranolol group suffered a 50% rate of hypotension. One study looked at morphine versus bupivacaine given by epidural in 50 patients (Oka T, 2001). The Incidence of AF reduced from 28% in the morphine group to 4.3% in the bupivacaine group and there were no excess of side effects in this group. Finally Terzi et al in 1996 performed an unblinded PRCT in 194 patients, allocating 93 patients to prophylactic magnesium. They demonstrated a reduction from 23% to an incidence of 10% of AF in the magnesium group, without side effects. Thus in summary digoxin and verapamil do not reduce the incidence of AF. Amiodarone may cause ARDS, and thus cannot be recommended, and flecainide and beta-blockers have been inadequately studied to make safe recommendations. Single randomized controlled trials have demonstrated evidence for bupivacaine epidural , magnesium and diltiazem for the prophylaxis of AF in patients undergoing non-cardiac thoracic surgery.

Clinical Bottom Line

Single randomized trials have demonstrated a benefit for diltiazem, bupivacaine epidural and magnesium for prophylaxis against AF in patients undergoing non-cardiac thoracic surgery, with a number needed to treat of between 4 and 8 with these regimes.


  1. De Decker K, Jorens PG, Van Schil P. Cardiac complications after noncardiac thoracic surgery: an evidence-based current review. Ann Thorac Surg 2003;75:1340-1348.
  2. Lanza LA, Visbal AI, DeValeria PA. Low-dose oral amiodarone prophylaxis reduces atrial fibrillation after pulmonary resection. Ann Thorac Surg 2003;75:223-230.
  3. Oka T, Ozawa Y, Ohkubo Y. Thoracic epidural bupivacaine attenuates supraventricular tachyarrhythmias after pulmonary resection. Anesth Analg 2001;93:253-259.
  4. Amar D, Roistacher N, Rusch VW, et al. Effects of diltiazem prophylaxis on the incidence and clinical outcome of atrial arrhythmias after thoracic surgery. J Thorac Cardiovasc Surg 2000;120:790-798.
  5. Bayliff CD, Massel DR, Inculet RI, et al. Propranolol for the Prevention of Postoperative Arrhythmias in General Thoracic Surgery. Ann Thorac Surg 1999;67:182-186.
  6. Jakobsen CJ, Bille S, Ahlburg P, et al. Perioperative metoprolol reduces the frequency of atrial fibrillation after thoracotomy for lung resection. J Cardiothorac Vasc Anesth 1997;11:746-751.
  7. Amar D, Roistacher N, Burt ME, et al. Effects of diltiazem versus digoxin on dysrhythmias and cardiac function after pneumonectomy. Ann Thorac Surg 1997;63:1374-1381.
  8. Terzi A, Furlan G, Chiavacci P, et al. VS Prevention of atrial tachyarrhythmias after non-cardiac thoracic surgery by infusion of magnesium sulfate. Thorac Cardiovasc Surg 1996;44:300-303.
  9. Van Mieghem W, Tits G, Demuynck K, et al. Verapamil as prophylactic treatment for atrial fibrillation after lung operations. Ann Thorac Surg 1996;61:1083-1085.
  10. Van Mieghem W, Coolen L, Malysse I, et al. Amiodarone and the development of ARDS after lung surgery. Chest 1994;105:1642-1645.
  11. Ritchie AJ, Danton M, Gibbons JRP. Prophylactic digitalisation in pulmonary surgery. Thorax 1992;47:41-43.
  12. Borgeat A, Petropoulos P, Cavin R, et al. Prevention of arrhythmias after noncardiac thoracic operations: flecainide versus digoxin. Ann Thorac Surg 1991;51:964-967.
  13. Ritchie AJ, Bowe p, Gibbons JRP. Prophylactic Digitalization for Thoracotomy: A reassessment. Ann Thorac Surg 1990;50:86-88.
  14. Borgeat A, Biollaz J, Bayer-Berger M, et al. Prevention of Arrhythmias by Flecainide After Noncardiac Thoracic Surgery. Ann Thorac Surg 1989;48:232-234.
  15. Ritchie AJ, Tolan M, Whiteside M, et al. Prophylactic digitalization fails to control dysrhythmia in thoracic esophageal operations. Ann Thorac Surg 1993;55:86-88.