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

What is the optimal anticoagulation management of patients post cardiac surgery who go into atrial fibrillation?

Three Part Question

In [patients with atrial fibrillation following coronary arterial bypass graft] is [anticoagulation] of any benefit in terms of [stroke prevention]?

Clinical Scenario

You are the cardiothoracic registrar on call and you are asked to review a 65 year old gentleman who underwent coronary artery bypass grafting 3 days ago. He has gone into atrial fibrillation with a ventricular rate of 130 beats per minute, but his blood pressure is good at 105/70 . You commence him on Metoprolol, and the nurse asks you if you would like to fully anticoagulate him with full dose tinzaparin. He suffered a stroke 4 years ago and getting a repeat CVA was one of his main concerns prior to his operation and you are tempted to give him this anticoagulation right away. However, you are unsure if this is safe in someone so recently post-cardiac surgery and you are aware of no evidence for a reduction in stroke risk with anticoagulation post-cardiac surgery, so resolve to check up on the literature that night.

Search Strategy

Medline 1966-March 2004 using the OVID interface.
exp Cardiovascular surgical procedures/ OR cardiovascular surgical OR exp Thoracic surgery/ OR Thoracic OR exp Coronary Artery bypass/ OR coronary artery bypass OR OR coronary OR cardiac OR ] AND [exp anticoagulation/ OR OR exp warfarin/ OR OR exp heparin/ OR] AND [exp atrial fibrillation/ OR atrial OR OR exp atrial flutter/ OR atrial OR exp supraventricular tachycardia/] . This search was repeated in Cochrane Central Register of Controlled Trials.

Search Outcome

A total of 166 papers were found of which only 2 were directly relevant. The American Heart Association guidelines for management of atrial fibrillation were reviewed. No direct studies were relevant looking at the reduction of stroke after AF post-cardiac surgery, thus the references of these reviews and the AHA guidelines were checked which identifed 4 further papers. Finally 2 papers were suggested by journal club colleagues. These 11 papers are presented in the table below.

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Hart et al,
9874 patients with lone AF from 16 trials identified from a search by the Cochrane Collaboration. Mean follow up 1.7 years. First part of meta-analysis 6 trials (N=2900) compared adjusted dose warfarin with placebo (5 trials) or control group (1 trial) Target INR ranged from 2.0-2.6Meta-analysis (level 1a)Stroke (including ischaemic and haemorrhagic)Warfarin reduced relative risk by 62% (95%CI, 48% to 72%)
Absolute risk reduction (ARR) 2.7% per year
Transient ischaemic attacks were not considered The percentage of participants who underwent neuroimaging or autopsy to reliably distinguish between ischaemic or haemorrhagic stroke varied, and this percentage was not always reported Well conducted meta-analysis with generally good homogeneity of results amongst the trials
Ischaemic strokeRelative Risk Reduction (RRR) of 62% (CI, 52% to 74%)
Intracranial haemorrhage0.3% vs 0.1% per year (P=NS)
Major extracranial haemorrhageRelative risk 2.4 (CI, 1.2 to 4.6) ARR 0.3% per year
All cause mortalityRRR 26% (CI, 4% to 43%) ARR 1.6% per year
Hart et al,
Second part of meta-analysis : 6 trials (N=3337) compared antiplatelet therapy with placebo. Aspirin dose ranged from 25mg twice daily to 1300 mg daily.Meta-analysis (level 1a)Stroke (including ischaemic and haemorrhagic)Aspirin reduced by 22% (CI, 2% to 38%) (P=NS)
ARR 1.5% per year
Ischaemic strokeReduction of 23% (CI, 0% to 40%)
Intracranial haemorrhageOccurred in 4 aspirin recipients (0.2% per year) and 3 placebo recipients. (difference not statistically significant)
All cause mortalityRRR 16% (CI, -5% to 33%)
Hart et al,
Third part of meta-analysis 5 trials (N=2837) compared adjusted dose warfarin with aspirinMeta-analysis (level 1a)Stroke (including ischaemic and haemorrhagic)RRR 36% (CI, -14% to 52%)
Ischaemic strokeRRR 46% (CI, -27% to 60%)
Intracranial haemorrhageRR 2.1 (CI, 1.0 to 4.6)
Major extracranial haemorrhageRR 2.4 (CI, 1.2 to 3.4)
AR 0.2% per year
All cause mortalityRRR 8% (CI, -21% to 30%)
Hylek et al,
A cohort of 13,559 patients with non-valvular atrial fibrillation identified from an integrated healthcare database. Severity of stroke was assessed and admission INR was recordedObservational cohort study (level 2b)30 day mortality of patients having a strokeNeither aspirin or warfarin 24% death rate
Warfarin with INR <15, 15% death rate
Warfarin with INR 1.5-1.9, 16% death rate
Warfarin with INR >2.0, 6% death rate
Aspirin only 15% death rate
Observational assessment of the effect of treatment It is possible that patients who had minor strokes did not seek medical attention and hence not enrolled on the study and also patients with catastrophic cerebrovascular events leading to out of hospital deaths would have been missed by this study
Systematic review of a wide range of issues in management of atrial fibrillation. This review updated a previous review conducted in 1999Systematic review (level 1a)Antithrombotic therapy in patients with AFAntithrombotic therapy (oral anticoagulant or aspirin) should be considered for all patients in AF, except those with lone AF, to prevent thromboembolism. To aim for INR of 2-3 in high risk patients
(Class I based on Grade A evidence)
AF post-CABG to be managed similar to non-surgical patients. (Class IIA based on grade B evidence)
Search strategies not given. Only 2 papers quoted, both pre 1990 in support of the increased risk of stroke in AF post CABG. Guidelines further state that anticoagulation with heparin or an oral anticoagulant is appropriate when AF persists more than 48 h
Single author review of the literature on the management of AF post-cardiac surgeryReview (level 2a)Anticoagulation after AFAHA guidelines quoted and the study by maisel. Recommends anticoagulation after 48 hours with warfarin . Sudies do not allow a recommendation for or against Heparin therapy and an individual risk benefit analysis should be madeThis is an expert review rather than a systematic review
Maisel et al,
Search of Medline from 1966-2000 of English language papers and checking of reference listsSystematic review (level 3a)Administration of WarfarinAll patients with atrial fibrillation persisting for more than 24 to 48 hours and without contraindication should receive anticoagulationThe search strategy for this review has major deficiencies, missing non-English papers, not searching Embase or grey literature, not contacting experts or performing double abstracting
Heparin administrationBecause the utility of heparin to prevent thrombus formation in post-CABG patients with atrial fibrillation is unknown and because the risk for postoperative bleeding is likely to be increased by using heparin therapy, routine use of heparin is unadvisable
52 patients from a cohort of 2630 patients undergoing On-pump CABG who suffered a stroke Stroke diagnosed either by clinical findings or CT AF diagnosed from clinical recordsRetrospective cohort study (level 3b)Incidence of AF prior to stroke19 of 52 patients had AF prior to stroke (36.5%)Unreliable diagnosis of AF , relying on clinical record of this event. No 24 hour monitoring protocol Control incidence of AF unknown in this cohort Enoxaparin was started in some patients post AF, but timings and frequency not documented
Timing of stroke after AFMean time after first onset of AF was 23 hours, (range 0-40 hours)
Villareal et al,
6477 patients in the Texas heart institute database who underwent CABG. 994 had AF 195 AF patients were matched with 195 controls AF defined as documentation of AF at any time in the hospital recordRetrospective cohort study and case-control study (level 2b)5 year survival in case matched patientsAF group 80% survival
Control group 93% survival p=0.0018
Flawed definition of AF, as its categorisation relied on clinician documentation rather than 24 hour ECG monitoring, thus many episodes may have been missed AF group were significantly higher risk in a large range of categories
Stroke incidence in cohortAF group 52%
Control group 1.7%
Cardiac arrest in cohortAF group 2.4%
Control group 0.8%
Stamou et al,
19512 patients from the Washington cardiac surgical database of patients undergoing on pump CABG, with <4 grafts from 1989 to 1999 333 patients had a postoperative stroke (neurological deficit >72 hours)Retrospective cohort study (level 2b)Independent predictors of stroke by multivariate analysisPost-operative AF increased odds of stroke, OR 1.7 (CI 1.4-2.2)Patients having a stroke also had more frequent CCF, diabetes, recent MI, previous CVA, renal failure, unstable angina and were older
Almassi et al,
3855 patients who underwent open cardiac surgery at 14 Veterans Medical CentersObservational cohort study (level 2b)Incidence of strokePatients with AF 5.26%
Patients without AF 2.44%
Stroke was not the primary outcome measure of this study and the AF group were a higher risk group of patients
Creswell LL et al,
3983 patients undergoing cardiac surgeryObervational cohort study (level 2b)Incidence of strokePostoperative AF 3.3% stroke
Sinus rhythm 1.4%; p < 0.0005
These results do not reflect the other risk factors for stroke in the 2 groups
Malouf et al,
141 patients undergoing CABG (56), valve (69) or congenital (16) cardiac surgery postoperatively by 2-D Echocardiography Group 1 (n=74) received full anticoagulation (warfarin 73; heparin 1) Group 2 (n=67) antithrombotics or no treatmentProspective cohort study (level 2b)Pericardial effusion of any sizeAnticoagulated group 43/74 (58%)
Control group 27/67 (40%)
p=0.043 by Fishers exact test
41 of the 74 anticoagulations had a period of excessive anticoagulation and these patients had an excessive incidence of effusion Selected cohort of patients
Large pericardial effusionsAnticoagulated group 24/74 (32%)
Control group 3/67 (4%) p<0.005
Tamponade requiring drainageAnticoagulated group 12/74 (16%)
Controls 0/67 (0%) p<0.001


There are several issues that must be addressed for this topic: does anticoagulation reduce the incidence of stroke in patients with atrial fibrillation, in patients who have just undergone cardiac surgery; is there also an increased risk of stroke; and if stroke risk is increased can anticoagulation reduce this incidence without an increase in bleeding complications? Addressing the issue of reduction in stroke risk in patients with atrial fibrillation, Hart et al analysed results from 16 trials by meta-analysis, 5 of which used warfarin against a placebo, 6 trials looked at antiplatelet therapy vs placebo and 5 trials looked at warfarin against aspirin. They demonstrated that warfarin reduced the relative risk of stroke both in comparison to placebo and aspirin and that warfarin is therefore by far the best long term treatment in patients with atrial fibrillation. The numbers needed to treat to prevent 1 stroke per year was 37 in the primary prevention group and 12 in the secondary prevention group (patients with a history of stroke) when compared to the placebo. These results were consistent for disabling and non-disabling strokes. It is interesting to note that though the incidence of intracranial haemorrhages was twice that of placebo, the difference was not statistically significant. The mean INR achieved was 2.0-2.6 in primary prevention trials and 2.9 in a single secondary prevention trial. Aspirin reduced the incidence of stroke by 22% when compared to the placebo and numbers needed to treat per year to prevent a stroke was 67 in the primary prevention group and 40 for secondary prevention. The difference of intracranial and extracranial haemorrhages in both groups was not statistically significant and all cause mortality was not significantly reduced by aspirin. In addition, since the above meta-analysis was performed, Hylek et al published a cohort study of 13,600 patients in the New England Journal of Medicine, showing that an INR above 2.0 significantly improved survival among patients with AF who suffer a stroke. The next issue is whether AF post cardiac surgery significantly increases the risk of stroke. Lahtinen et al reported that 19 of 52 strokes (37%) in their cohort of patients undergoing CABG were preceded by atrial fibrillation an average of 21 hours previously. However this is a small study and the incidence of AF in their patients who didn't have a stroke was not reported. Villareal et al reported that in a cohort of 6500 patients undergoing CABG, patients who went into AF had a much higher incidence of stroke (5.2% vs 1.7%) and also an increased risk of short and long term mortality. These patients were however significantly higher risk in a large range of categories including age, Heart failure, COPD and underlying coronary arterial disease and thus some caution should be used when analysing these figures. However adverse long term mortality persisted after case-control matching. Stamou et al performed a retrospective analysis of 19,500 patients who had undergone CABG, of whom 333 had suffered a stroke. Multivariate analysis showed that atrial fibrillation was an independent predictor of stroke, increasing the odds of stroke by 1.7. However multiple other high risk factors also predicted stroke and thus the stroke group was a much higher risk group than those who did not suffer a stroke. Almassi also showed a 5% stroke rate in patients with AF compared to 2.5% in the sinus rhythm group, performing a similar study to Stamou in a cohort of 3855 patients. Creswell et al found that the incidence of stroke was 3.3% if the patient was in AF compared to 1.4% in those with sinus rhythm in a cohort of 3983 patients. Unfortunately there are no studies that demonstrate that immediate or delayed anticoagulation of patients post cardiac surgery who go into AF significantly reduce this increased risk of stroke. However addressing the issue of the safety of immediate anticoagulation, Malouf et al performed a cohort study on 144 cardiac surgical patients, performing an echocardiogram on all these patients. They found a 16% incidence of tamponade requiring drainage in patients receiving early warfarinisation, with no such tamponades in controls. In addition 32% of the anticoagulated patients had a large pericardial effusion on echocardiography, compared to 4% in controls. As a caveat, these patients received warfarin not heparin, and a large number of these patients suffered a period of excessive anticoagulation at some stage, however despite this, their figures are a cause for concern. The American College of Cardiology/ American Heart Association/ European Society of Cardiology Guidelines for management of patients with atrial fibrillation suggest managing post CABG AF in a similar fashion to atrial fibrillation in non- surgical patients. They recommend use of antithrombotic treatment in high risk patients and a target INR of 2.0-3.0.They further recommend that anticoagulation with heparin or an oral anticoagulant is appropriate when AF persists more than 48 h. However they only quote two pre-1990 papers in support of this statement. Emile Daoud supported these recommendations in a review published in 2004, stating that warfarin should be started after 48 hours. He further recommends that in higher risk patients even if sinus rhythm returns, warfarin should be continued for 4 weeks as there is a delay in return of atrial contractility post AF. Maisel et al performed a review in 2001. The AHA guidelines were also supported for warfarinisation, but caution is advised in heparinisation, quoting anxiety over an increased risk of pericardial effusions Heparinisation is recommended only for the highest risk patients for stroke. Thus it is clear that chronic atrial fibrillation increases the risk of stroke and warfarinisation provides the optimal protection from this risk, with a number needed to treat of only 37 to save a stroke and this number drops to 12 if there is a history of stroke. It is clear that AF post cardiac surgery doubles the risk of stroke, but there are as yet no studies that have demonstrated a drop in this risk with immediate anticoagulation. In addition one study provides some evidence that there is a risk of pericardial effusions with early anticoagulation. The American Heart association supported by several other authors thus recommends warfarinisation while in AF, with an INR of 2-3 and anticoagulation within 48 hours of the onset of AF post cardiac surgery.

Clinical Bottom Line

Patients post cardiac surgery require warfarinisation while in atrial fibrillation with an INR of 2-3, and full anticoagulation should be commenced within 48 hours of the onset of AF due to a doubling of their risk of stroke.


  1. Hart RG, Benavente O, McBride R, Pearce LA. Antithrombotic Therapy To Prevent Stroke in Patients with Atrial Fibrillation: A Meta-Analysis. Ann Intern Med 1999;131:492-501.
  2. Hart RG, Benavente O, McBride R, Pearce LA. Antithrombotic Therapy To Prevent Stroke in Patients with Atrial Fibrillation: A Meta-Analysis. Ann Intern Med 1999;131:492-501.
  3. Hart RG, Benavente O, McBride R, Pearce LA. Antithrombotic Therapy To Prevent Stroke in Patients with Atrial Fibrillation: A Meta-Analysis. Ann Intern Med 1999;131:492-501.
  4. Hylek EM, Go AS, Yuchiao C, Jensvold NG, Henault LE, Selby JV, Singer DE. Effect of Intensity of Oral Anticoagulation on Stroke Severity and Mortality in Atrial Fibrillation. N Engl J Med 2003;349:1019-1026.
  5. A Report of the American College of Cardiology/AmericanHeart Association Task Force on Practice Guidelines and theEuropean Society of Cardiology Committee for PracticeGuidelines and Policy Conferences ACC/AHA/ESC Guidelines for the Management of Patients With Atrial Fibrillation. J Am Coll Cardiol 2001;38:1266i-1266Ixx.
  6. Daoud EG. Management of atrial fibrillation in the post-cardiac surgery setting. Cardiol Clin 2004;22:159-166.
  7. Maisel WH, Rawn JD, Stevenson WG. Atrial fibrillation after cardiac surgery. Ann Intern Med 2001;135:1061-1073.
  8. Lahtinen J, Biancari F, Salmela E, Mosorin M, Satta J, Rainio P, Rimpila J, Lepoja M, Juvonen T. Postoperative Atrial Fibrillation is a Major Cause of Stroke After On-Pump Coronary Artery Bypass Surgery. Ann Thorac Surg 2004;77:1241-1244.
  9. Villareal RP, Hariharan R, Liu BC, Kar B, Lee V, Elayda M, Lopez A, Rasekh A, Wilson JM, Massumi A. Postoperative Atrial Fibrillation and Mortality After Coronary Artery Bypass Surgery. J Am Coll Cardiol 2004;43:742-748.
  10. Stamou SC, Hill PC, Dangas G, Pfister AJ, Boyce SW, Dullum MKC, Bafi AS, Corso PJ. Stroke after Coronary Artery Bypass. Stroke 2001;32:1508-1513.
  11. Almassi GH, Schowalter T, Nicolosi AC, Aggarwal A, Moritz TE, Henderson WG, Tarazi R, Shroyer AL, Sethi GK, Grover FL, Hammermeister KE. Atrial fibrillation after cardiac surgery: a major morbid event? Ann Surg 1997;226:501-511.
  12. Creswell LL, Schuessler RB, Rosenbloom M, Cox JL. Hazards of postoperative atrial arrhythmias. Ann Thorac Surg 1993;56:539-549.
  13. Malouf JF, Alam S, Gharzeddine W, Stefadouros MA. The role of anticoagulation in the development of pericardial effusion and late tamponade after cardiac surgery. Eur Heart J [erratum appears in Eur Heart J 1994 Apr;15(4):583-4] 1993;14:1451-1457.