Practice guidelines contain a well-defined threshold for starting anticoagulation as stroke prevention for AF but there is no clear standard for when to stop it.

According to a Research Letter in the JAMA Internal Medicine, the magnitudes of benefits and harms of anticoagulation are likely to change substantially as diseases such as dementia progress towards end of life.

The US authors reported a cross-sectional study of 15,000 elderly nursing home residents with advanced dementia and AF (with a CHA2DS2VASC score ≥2) who died between 2014 and 2017. The residents, mostly women (68.2%), had a mean age of 87.5 years.

The study found 33.1% of residents were receiving an anticoagulant in their last six months of life.

A higher CHA2DS2VASC (score >7, OR 1.38) and ATRIA (score >7, OR 1.25) scores, nursing home length of stay of at least 1 year (OR 2.68), not having Medicaid (OR 1.59), weight loss (OR 1.09), pressure ulcers (OR 1.37), and difficulty swallowing (OR 1.12) were associated with greater odds of anticoagulant use.

“With the notable exception of hospice use, most indicators of high short-term mortality, such as difficulty swallowing, weight loss, and pressure ulcers, were associated with greater odds of anticoagulant use.”

“Conversely, older age (80-89 y, OR, 0.82 [95% CI, 0.74-0.92]; ≥90 y, OR, 0.59 [95% CI, 0.52-0.66]), female sex (OR, 0.88 [95% CI, 0.81-0.95]), requiring restraints (OR, 0.79 [95% CI, 0.66-0.95]), and being enrolled in hospice (OR, 0.76 [95% CI, 0.70-0.83]) were associated with lesser odds of anticoagulant use,” the study said.

The study said without guidance for when to stop anticoagulation, clinicians were instead asked to engage in shared decision-making with patients and their families.

“Data about the benefits and harms of therapy are essential to that process,” they said.

They said that as dementia progresses and function is irretrievably lost, the potential benefits of preventing a stroke also become increasingly attenuated.

An editorial in the journal said the risks and benefits of therapeutic anticoagulation in severe dementia and other life-limiting illnesses have not been well studied.

“In real-world practice, many patients with severe dementia have limited life expectancy and would choose to focus on quality of life. However, avoiding the potential morbidity of stroke may still be within patients’ and families’ goals at the end of life.”

“Others might argue that for those with limited prognosis, drugs for chronic conditions that do not directly target symptoms, such as dyspnea or pain, increase the risk of adverse events without clear benefit.”

The editorial said the study’s findings highlight the lack of a rational strategy for managing anticoagulation in those with limited life expectancy owing to age or illness.

The issue is further complicated by lack of evidence in seriously ill and frail patients.

“Traditionally, the net clinical benefit of anticoagulation is driven by difference between ischemic stroke reduction and intracranial hemorrhage risk. A more patient-centered framework would expand this narrow definition of net clinical benefit. Consideration of the competing risk of death from other causes, such as dementia or cancer, decreases the net clinical benefit of anticoagulation and should be incorporated.”

They said bleeding events factored in should not be limited to intracranial hemorrhage as extracranial and so-called “nuisance bleeding” are common and can diminish quality of life.

They called for studies of decision-making aids and dose reduction or deprescribing in this population using this expanded net benefit definition.

National Heart Foundation/Cardiac Society of Australia and New Zealand AF guidelines say the net clinical benefit is in favour of treating older people and those aged more than 85 years.