Atrial fibrillation: warfarin reduces the risk of stroke and death.

Clinical bottom line (level 1a)

  1. Patients with atrial fibrillation who take warfarin, have fewer strokes (NNF = 23 for 1-2 years) and are less likely to die (NNT = 42 at 1-2 years) than those who take placebo or nothing.
  2. The effect on major bleeding is unclear. Around 2% of patients have a major bleed.
  3. The risk of stroke increases with:
    • increasing age
    • previous TIA or stroke
    • history hypertension
    • diabetes mellitus
Atrial Fibrillation Investigators : Archives of Internal Medicine 1994; 154: 1449-1457
Expires June 2003

The study

?blinded concealed randomised trial with intention-to-treat
Setting: five university hospitals, USA, Canada, Denmark

4253 patients (aged mean 69 years, 70% male) atrial fibrillation
Note:
  • This study is, in fact, the combination of five randomised controlled trials (AFASK, SPAF, BAATAF, CAFA, SPINAF).
  • Only two trials were double-blind, but they all had similar inclusion and exclusion criteria.
  • Two trials used an INR range of 2.0 to 4.2 (the others used prothrombin ratio).
  • Multivariate regression analysis was performed on risk factors.


    • Control Group: (n = 1236, 1236 analysed): placebo or control
    Experimental Group: (n = 1225, 1225 analysed): warfarin

    100% followed for 1.2-2.3 years
    Outcome notes:
    • stroke : focal neurological deficit >24 hours
    • major haemorrhage : intracranial bleed, hospital admission or required two or more units of blood

    The evidence

    Outcome Time to outcome CEREERRRR
    (95% CI)    		ARR
    (95% CI)    		NNT
    (95% CI)
    stroke 1-2 years 81
    (6.55%)    		 27
    (2.20%)    		 66%
    (48% to 78%)    		 4.35%
    (2.74% to 5.96%)    		 23
    (17 to 36)
    death 1-2 years 99
    (8.01%)    		 69
    (5.63%)    		 30%
    (5% to 48%)    		 2.38%
    (0.39% to 4.37%)    		 42
    (23 to 260)
    major haemorrhage 1-2 years 18
    (1.46%)    		 24
    (1.96%)    		 -35%
    (-147% to 27%)    		 -0.50%
    (-1.53% to 0.52%)    		 -200
    (NNT = 190 to infinity;
    NNH = 65 to infinity)

  • Significant risk factors were:
    • increasing age
    • previous transient ischaemic attack or stroke
    • history of hypertension
    • diabetes mellitus

    • Comments

    1. There was no test for heterogeneity between studies. There was no formal meta-analysis performed- raw data combined directly.
    2. Warfarin is a risky drug, because it causes to bleeds, and is costly to patients in terms of time and inconvenience. Because of this it is useful to estimate individual absolute stroke risk to aid treatment decisions.
    3. The bleeding rate in these trials was very low, and is generally reckoned to be at least double in routine practice. The effectiveness and cost-effectiveness of anticoagulation drops dramatically if bleeding rates are high.

    Citation

    1. Atrial Fibrillation Investigators , : Risk factors for stroke and efficacy of antithrombotic therapy in atrial fibrillation: analysis of pooled data from five randomized controlled trials. Archives of Internal Medicine 1994; 154: 1449-1457
    Search Terms: atrial fibr* and anticoag* in Cochrane
    Contributor: Chris Ball and Clare Wotton, June 2000
    Reviewer: Rowan Harwood

    Clinical Question.
    Patient atrial fibrillation
    Intervention or Exposure warfarin
    Comparison placebo
    Outcome stroke, bleeding