Intended for healthcare professionals


Using anticoagulation or aspirin to prevent stroke

BMJ 2000; 320 doi: (Published 08 April 2000) Cite this as: BMJ 2000;320:1008

Research was methodologically flawed

  1. Simon J Elli, consultant neurologist (,
  2. Renu Hans, senior house officer
  1. Neurosciences Trust, North Staffordshire Royal Infirmary, Keele University, Stoke on Trent ST4 7LN
  2. Department of Primary Care and General Practice, Medical School, University of Birmingham, Birmingham B15 2TT
  3. Department of Stroke Medicine, Guy's, King's and St Thomas's School of Medicine, London SE5 9PJ
  4. Julius Centre for General Practice and Patient-oriented Research, University Medical Centre Utrecht, PO Box 80045, 3508 TA Utrecht, Netherlands
  5. Pathology Laboratory, Warwick Hospital, Warwick CV34 5BJ
  6. Bristol Heart Institute, University of Bristol, Bristol Royal Infirmary, Bristol BS2 8HW
  7. Department of Social Medicine, University of Bristol, Bristol BS8 2PR
  8. Department of General Practice, University of Maastricht, PO Box 616, 6200 MD Maastricht, Netherlands
  9. Department of Neurology, University of Maastricht
  10. Department of Cardiology, University of Maastricht
  11. Department of Methodology and Statistics, University of Maastricht

    EDITOR—The paper by Hellemons et al is not justified in concluding that aspirin is the prophylactic choice in primary care for atrial fibrillation, if there is no clear indication for full anticoagulation.1

    The study is methodologically flawed. As clinicians, we ask ourselves: “Which patient in atrial fibrillation should be given anticoagulants?” This is a statistical question about the risks and benefits of aspirin or warfarin for that individual patient.

    In the power calculation Hellemons et al asked whether low anticoagulation (international normalised ratio 1.1-1.6) or aspirin should be used—but this is the wrong question. The choice should have been between aspirin and standard anticoagulation (INR 2.5-3.5). The increased incidence of major intracranial bleeding in the aspirin group compared with the anticoagulated groups (0.75% per patient year v 0.35%) calls into question the sagacity of using one tailed statistical tests.

    As the study was underpowered, the question of whether standard anticoagulation or aspirin was better in preventing major cerebral infarction cannot be answered. Although there is a trend towards full anticoagulation (hazard ratio 0.67), the 95% confidence intervals are so wide (0.11 to 4.1) that the result is meaningless.

    The arbitrary exclusion from standard anticoagulation of all people who were 78 years or older also undermines the study, for although it may have reduced the complication rate from anticoagulation, it will have also reduced the potential benefit.

    This paper highlights the problems in reporting “negative” or “no difference” studies. It has failed to show “no difference” between standard anticoagulation and aspirin prophylaxis in atrial fibrillation, as clinically important differences could well exist within the confidence limits. The study adds little to previous work that does demonstrate benefit from anticoagulation2 and may be misinterpreted as an excuse for a nihilistic approach to the prevention of embolic episodes in primary care. …

    View Full Text

    Log in

    Log in through your institution


    * For online subscription