The Problem
An 82-year-old woman with hypertension and diabetes is seen in the clinic for newly diagnosed atrial fibrillation (AF). She uses a walker and fell twice last year. She completes activities of daily living (ADLs) independently. Should she be prescribed an anticoagulant?
AF markedly increases the risk of ischemic stroke, which often results in disability and threatens independence.1 Anticoagulants reduce stroke rates but simultaneously increase hemorrhage rates. Unfortunately, predictors of benefit and harm overlap considerably; for example, advanced age predicts both stroke and hemorrhage risk. Guidelines give clear-cut therapeutic recommendations based on stroke risk but provide less specific guidance on balancing the increased hemorrhage risk.2 Although risk models help quantify bleeding risk, the consequences of bleeding (e.g., disability) in older adults are ill defined. The result is a quagmire.
The Pros and Cons of Anticoagulant Use in Older Adults with AF
Randomized controlled trials (RCTs) unequivocally demonstrate that anticoagulants reduce stroke rates, including in older adults.3 We have long understood that anticoagulants also increase the risk of intracranial hemorrhage, which is uncommon but devastating. By comparison, many consider the more common extracranial hemorrhage (e.g., gastrointestinal hemorrhage) to be a temporary inconvenience. While this is likely true for younger adults, bleeding may be more consequential for older adults.
Stroke prevention is important because strokes often result in a sudden loss of ADL independence. Now consider that hospitalization for any reason poses a risk to older adults’ independence: one in three develop lasting hospital-acquired disability.4 Even when the underlying disease is treated, hospitalization of older adults may result in lost ADL independence. Given the typical management for hemorrhage (e.g., procedures, not allowing patients to eat), such admissions are bound to result in hospital-acquired disability. Although stroke-associated disability has been reported in RCTs, hemorrhage-associated disability has not been commonly reported.
Older also adults prize quality of life.5 Anticoagulants impose costs to quality of life not captured in RCTs, including so-called nuisance bleeding, delays for urgent procedures, physician visits, medication interactions, and out-of-pocket expenses. Such costs are particularly burdensome to vulnerable older adults. They are hardly a niche population: 80% of older adults with AF are vulnerable with one or more geriatric syndromes (e.g., falls, frailty).6
What We Know
Guidelines recommend using the CHA2DS2-VASc score to estimate stroke risk and offer anticoagulants above a certain threshold — one that nearly all older adults with AF surpass.2,6 The recommendation implies that above the threshold, the benefits outweigh the risks; however, this threshold has not yet been tested with regard to disability and quality of life.
Faced with uncertainty, anticoagulant use in older adults with AF is far below what the guidelines suggest: only 45% consistently use anticoagulants.7 When surveyed about the decision not to treat specific patients in their care, physicians cite frailty, falls, and life expectancy — geriatric features that are under-addressed in RCTs.8,9
The ELDER-AF trial addressed anticoagulant use in older adults with AF.10 The investigators randomly assigned adults 80 years of age or older who were not considered candidates for therapeutic anticoagulation to placebo or low-dose edoxaban, a direct-acting oral anticoagulant. The trial again demonstrated that anticoagulants reduce stroke risk by 66%.
The trial also underscored the trouble with anticoagulants in older adults. Major bleeding, the primary safety outcome, was not statistically higher with treatment. However, this outcome excludes clinically relevant nonmajor bleeding, which is common and consequential.11 Clinically relevant nonmajor bleeding includes hemorrhages requiring interventions like hospitalization, endoscopy, and surgery. Using this more germane end point, anticoagulation increased bleeding risk by 65%. In a trial in which one sixth of all participants had major or clinically relevant nonmajor bleeding, no data were presented on the outcomes of bleeding events. We are left wondering — did bleeding result in lost independence or decrements in quality of life?
What We Need
A trial that compares oral anticoagulants to placebo for older adults with AF where function and quality of life are the primary outcomes would capture the broader impact of therapy that many older adults prioritize. Such a trial would answer straightforward questions: If the 82-year-old woman described at the beginning of this article uses an anticoagulant, will it improve her chances of remaining ADL independent? Will it sustain her quality of life?
The work ahead is to co-design and execute a trial with stakeholders — older adults with AF, caregivers, and physicians — whose decisions we seek to inform. This trial would recruit a population similar to the ELDER-AF trial: older adults with AF for whom physicians have misgivings about anticoagulants or patients who, for any reason, decided not to use anticoagulants despite their physician’s recommendation. Based on projections from disability rates after stroke and hemorrhage, and studies of quality of life in AF,12,13 this trial would need to randomly assign approximately 1600 patients to treatment with 18 months of follow-up to detect a 5% absolute change in ADL independence and a 0.15-SD change in the 36-Item Short Form Health Survey physical component score. The results of such a trial could provide clarity for millions of older adults with AF and their physicians who, with current data, struggle to balance the benefits and harms of anticoagulants.
ARYAN'S BLOG 