Stroke risk 'higher at start of warfarin treatment'

"Warfarin doubles risk of stroke in first week for irregular heartbeat sufferers," reports The Daily Telegraph.

Warfarin is a drug known to reduce the risk of stroke in people with atrial fibrillation, but the results of the study this headline is based on suggest that particular care should be taken when treatment is started.

Atrial fibrillation is the most common abnormal heart rhythm. This uncoordinated heart function causes blood to not be fully expelled with each heartbeat. As a result blood clots can form, and if a clot reaches the brain, it can block an artery, causing a potentially fatal ischaemic stroke.

Warfarin reduces the likelihood of blood clots forming. Previous studies have suggested that starting warfarin is associated with an increased risk of ischaemic stroke, and the researchers wanted to see if this was true.

The researchers compared people with atrial fibrillation who had a stroke with people who didn't have a stroke. They found that during the first 30 days of treatment, warfarin was associated with a 71% increased risk of stroke, with a peak in risk during the first week of treatment. However, after 30 days of treatment, warfarin was associated with a reduced risk of stroke.

The researchers suggest that the way that warfarin works causes a short period of excessive blood clotting.

However, in this study, people who took warfarin were compared with people who previously took no antithrombotic therapy. It is likely that the people who took warfarin were at higher risk of having a stroke than people who took no antithrombotic therapy.

Where did the story come from?

The study was carried out by researchers from McGill University and the Jewish General Hospital in Montreal, Canada, and Princeton University in the US. It was funded by Bristol-Myers Squibb and Pfizer Inc, two pharmaceutical companies.

It was published in the peer-reviewed European Heart Journal.

The research was covered well by the media.

What kind of research was this?

This was a nested case-control study. A nested case-control study compares cases and controls from a defined cohort (group) of people. In this study, people who had an ischaemic stroke (a stroke caused by something that stops blood flow to the brain) were compared with up to 10 people who didn't have a stroke.

The second group were matched on the basis of age, sex, when atrial fibrillation was diagnosed, and how long people had had atrial fibrillation from a group of people with the condition in the UK.

A nested case-control study has advantages over a full cohort study in that it can be cheaper and easier to perform.

A nested case-control study is a type of observational study and it therefore cannot show that warfarin caused a stroke, as there could be other factors (confounders) that could explain the association.

What did the research involve?

The researchers looked at medical records from people diagnosed with atrial fibrillation in the UK between 1993 and 2008 who had an ischaemic stroke. They compared these people with the medical records of people with atrial fibrillation who didn't have a stroke.

For each person who had a stroke, up to 10 people who hadn't had a stroke were analysed. People were matched on the basis of age, sex, when atrial fibrillation was diagnosed and how long they had had atrial fibrillation.

The researchers looked at whether warfarin use was associated with an increased risk of stroke. The researchers broke down warfarin use into less than 30 days of treatment, 31-90 days of treatment and over 90 days of treatment. Exposure to warfarin was compared to no use of any antithrombotic therapy for at least one year.

The researchers adjusted their analyses for:

• excessive alcohol use
• smoking status
• obesity
• CHADS2 score (a clinical estimate of the risk of stroke)
• peripheral artery disease
• myocardial infarction
• previous cancer
• prior bleeds
• venous thromboembolism (blood clots)
• valvular disease

They also adjusted for current use of:

• angiotensin-converting enzyme (ACE) inhibitors
• angiotensin receptor blockers
• antidepressants
• antipsychotics
• non-steroidal anti-inflammatory drugs (NSAIDs)
• statins

What were the basic results?

A total of 70,776 people had atrial fibrillation and were followed-up for 3.9 years on average. Of these, 5,519 people had a stroke during the period of the study. The overall rate of stroke was 2% per year.

Warfarin was associated with a 71% increased risk of stroke in the first 30 days of use (relative risk [RR] 1.71, 95% confidence interval [CI] 1.39 to 2.12) compared with no use of any antithrombotic therapy.

The researchers also modelled risk during the first 30 days of use. They found that risk peaked at three days after starting warfarin (RR 2.33, 95% CI 1.50 to 3.61).

However, warfarin use was associated with a decreased risk of stroke if it was taken for longer than 30 days. Warfarin use for 31-90 days was associated with a 50% decreased risk of stroke (RR 0.50, 95% CI% 0.34 to 0.75), and warfarin use for over 90 days was associated with a 45% decreased risk of stroke (RR 0.55, 95% CI 0.50 to 0.61), compared with no use of any antithrombotic therapy.

How did the researchers interpret the results?

The researchers conclude that, "patients initiating warfarin may be at an increased risk of stoke during the first 30 days of treatment".

They suggest that in the early days of warfarin use, the drug could cause excessive blood clotting. This effect only lasts for a short time. After this, "warfarin was strongly associated with a decreased risk of ischaemic stroke in patients who have used warfarin for more than 30 days."

Conclusion

This study has found that warfarin was associated with an increased risk of ischaemic stroke during the first 30 days of treatment. After 30 days of treatment, warfarin was associated with a decreased risk of stroke.

However, this study has a number of limitations that should be considered:

• All information was from patient records, which means that it wasn't subject to recall bias, but the information may not be complete – we do not know if, for example, people took the medication they were prescribed.
• There may be other factors (confounders) that explain the association seen. In particular, the baseline stroke risk may be higher among people treated with warfarin compared with people not treated with any anticoagulant. Although the researchers tried to adjust for a number of factors that are associated with stroke risk, the possibility that people receiving warfarin differed from people not receiving it remains.

Warfarin has been shown to be effective in reducing the risk of stroke in people with atrial fibrillation, but the results of this study suggest that care should be taken when treatment is started.

Further research will be required to confirm these findings and whether anything can be done to reduce the risk of stroke during the first 30 days. The researchers suggest that a heparin (another anticoagulant) bridging strategy at the initial phase of treatment could be investigated.

Read the latest official guidance on atrial fibrillation.