Anticoagulation for patients with atrial fibrillation and risk factors for stroke
BMJ 2000; 320 doi: https://doi.org/10.1136/bmj.320.7244.1219 (Published 06 May 2000) Cite this as: BMJ 2000;320:1219
Warfarin reduces the risk by two thirds, but doctors still aren't prescribing it enough
- Stuart J Connolly, professor of medicine (coonostu{at}hhsc.ca)
Papers p 1236
The most clinically relevant advance in the management of cardiac arrhythmia in the past two decades has been that anticoagulant treatment substantially reduces the risk of stroke in patients with atrial fibrillation. The fact that each year 4% of patients with atrial fibrillation will have a stroke prompted several randomised trials of treatment with anticoagulants that consistently reported a reduction in the incidence of strokes with warfarin. A meta-analysis of these trials shows that the risk is reduced by two thirds.1
A subsequent analysis has shown that not all patients with atrial fibrillation have the same risk of having a stroke, and this has led to the idea of risk stratification.1 Major risk factors for stroke are age over 75, previous stroke or transient ischaemic attack, systemic hypertension, mitral stenosis, and left ventricular dysfunction. Patients with atrial fibrillation and any of these factors face a higher than average risk (5-15% per year) of a stroke.1 On the other hand, the risk of a stroke in patients with none of these risk factors is too low for warfarin to be beneficial.
Randomised trials evaluating aspirin, low fixed dose warfarin, and their combination clearly show the superiority of warfarin.2 On the basis of this strong evidence expert panels recommend that all patients with atrial fibrillation should be considered for treatment with anticoagulants and that if any major risk factors are present their treatment should aim to keep the international normalised ratio between 2.0 and 3.0.3
Surprisingly, given the evidence and expert opinion, most patients with atrial fibrillation are still not prescribed warfarin by their doctors. Surveys of practice patterns in several countries consistently report that only a quarter to a half of patients who have atrial fibrillation, with or without the risk factors for stroke, are actually given anticoagulants.4–6
Clinical trials are different from clinical practice
The reasons for such underprescribing are a matter of speculation. One of the important issues, now addressed by the paper of Kalra et al in this issue (p 1236), is whether the beneficial effects of warfarin seen in the randomised trials will also occur in clinical practice.7 Concern that results might not be so good may be why some physicians are not convinced that they should use warfarin. Clinical trials are run to optimise the chance of showing a benefit from treatment, usually with extra specialist personnel to monitor patients and side effects. Patients who enter clinical trials are more motivated than the average patient and more likely to be compliant with treatment. Screening procedures for trials are likely to be more rigorous than that done in everyday practice. Thus the findings of Kalra et al are an important contribution to our understanding of how trial results can transfer to clinical care. They screened 2457 patients in general practice clinics and found 248 (10%) patients with atrial fibrillation who had at least one major risk factor for stroke and did not have an excessive risk for haemorrhage. Although these patients had had atrial fibrillation for 11/2 years, only 31% were receiving anticoagulant drugs. Those not already receiving anticoagulants who were willing to be treated (n=167) entered this study and were given anticoagulants by local anticoagulation services according to local guidelines. Results over two years in these patients were compared with those patients given warfarin in randomised trials. Even though the patients participating in the trial were older and had more risk factors for stroke, their actual stroke rate over two years was similar to that seen in the trials. Their risk of haemorrhage (minor, major, and intracranial) was virtually identical. These findings clearly show that the results of the anticoagulation trials can be replicated in general medical practice.
Death and change influence patients' attitudes
These results emphasise the need to understand why anticoagulation remains underused. Many factors play a part, including patients' fears of haemorrhage and reluctance to have international normalised ratio testing on a regular basis. Howitt and Armstrong recently found that only half of patients eligible to take warfarin for preventing stroke were willing to do so after the benefits and risks had been explained to them. Patients' health beliefs and attitudes towards change and death had the greatest influence on whether they would accept treatment with warfarin.8
Undoubtedly it is vital to educate doctors. Several surveys have found a high rate of aspirin use in patients with atrial fibrillation even though antiplatelet therapy is much less effective in preventing stroke than warfarin. Doctors may also underuse warfarin because of reluctance to carry out regular testing of patients' international normalised ratios because of the low rate of remuneration for tests.
Currently we do not know why doctors and patients fail to implement the findings from anticoagulation trials. A better understanding of what factors actually influence patients' and doctors' behaviour in this area is needed, so that effective strategies can be implemented.
Acknowledgments
Dr Connolly has received speakers' fees from Dupont Pharmaceuticals.