Things We Do for No Reason™: Discontinuing anticoagulation in older patients with atrial fibrillation and a high risk of falls

The hospitalist admits an 81-year-old woman with a history of atrial fibrillation (AF) and hypertension for back pain due to a lumbar compression fracture after a mechanical fall. Her home medications include: apixaban, amlodipine, and zolpidem. A full trauma work-up including a computed tomography head, does not identify any other injuries or bleeding. As the patient prepares for discharge, the hospitalist recommends discontinuing long-term anticoagulation (AC) due to her fall risk. The population of adults over age 65 in the United States continues to rise and will double by 2040 to over 80 million residents. Age is the most significant risk factor for AF, with a steep rise in disease prevalence after age 65. Patients with AF have a fivefold increased risk of stroke compared to patients without AF, and long-term AC prevents stroke in high-risk patients. The CHA2DS2-VASc score guides stroke risk assessment and initiation of AC treatment. Hospitalists initiate AC for males who score ≥2 and females who score ≥3. This suggests that patients aged 75 or older with AF would benefit from AC for stroke prevention. This demographic, however, also faces the highest fall risk. Over a quarter of community-dwelling adults over age 65 experience at least one fall annually. Ten percent of these falls result in severe injury and account for 2.8 million emergency room visits and 800,000 hospitalizations.1 Falls are also the primary cause of traumatic intracranial bleeding, which carries a high mortality rate of 15%. Falls significantly contribute to morbidity, mortality, and disability in older patients. This population also experiences other comorbidities: polypharmacy, visual impairment, gait impairment, and urinary incontinence that contribute further to fall risk. One study showed that AC in fall-prone patients increases major bleeding risk by 39%.2 Despite meeting treatment criteria, hospitalists often withhold anticoagulants due to bleeding and fall risk.3 By avoiding AC, we may reduce the potential for serious traumatic injuries due to falls. AC greatly increases intracranial hemorrhage (ICH), one of the most serious bleeding complications. In 2005, Gage et al. retrospectively studied patients with AF using Medicare data and found high fall-risk patients experienced traumatic ICH 4.1 times (95% confidence interval: 2.4–7.1) more often than low-risk fall patients. In this study, 33% of high-risk fall patients received warfarin. Patients on warfarin had a 30-day mortality from ICH of 51.8% compared to 33.6% in patients not on warfarin (p = .007).4 In a separate study of the National Trauma Data Standard registry, researchers found a 180% higher risk of death in patients who fell while on AC.5 Compared to warfarin, direct oral anticoagulants (DOACs) have comparable or lower major bleeding and ICH risk. Additionally, rivaroxaban and dabigatran are the fifth and tenth, respectively, most frequently implicated drugs leading to emergency room visits due to adverse drug reactions among older adults.6 These findings, coupled with the absence of universally available reversal agents, raise significant concerns regarding the use of DOACs in the older adult population. While AC increases the bleeding risk in older adults who fall, this risk does not outweigh the stroke prevention benefits of AC in patients with AF. The risk of ischemic stroke, one of the leading causes of death and disability worldwide in patients with AF, hovers between 2% and more than 10% annually. Warfarin and DOACSs reduce the risk of stroke by at least two-thirds.7 Observational studies suggest that those who stop AC have a two-to threefold increase in ischemic stroke risk.8 While bleeding risk may increase in those who fall while on AC, those who experience ICH and traumatic brain injuries (TBIs) on AC have similar outcomes to those patients not on AC. A retrospective analysis from a trauma registry comprising 33,710 patients across 90 centers nationally found no correlation between the use of AC and overall mortality in cases of ground-level falls and TBIs, though the majority of these patients had mild TBIs.9 In contrast, Gage et al. reported that those on AC with an elevated fall risk have an increased rate of traumatic ICH and worse outcomes as compared to patients not on AC. In that same study, however, anticoagulated patients at high risk of CVA with a CHADS2 ≥ 2 still had a 25% risk reduction in the composite outcome of stroke, any hemorrhage, myocardial infarction, and out-of-hospital death compared to those not on AC.4 The authors concluded that these fall-prone adults derived a net benefit from AC. This suggests that while AC may increase bleeding and its complications, overall composite patient outcomes favor AC therapy in older adults. Additionally, studies show that bleeding risk plays an outsized role in physicians withholding AC and that most patients with AF will accept higher bleeding risks for a decrease in stroke probability.10, 11 Although a higher fall risk likely produces an increased risk of ICH, multiple models have quantified the number of falls needed to outweigh the stroke prevention benefit from AC in AF. A Markov decision analytic model estimated that one would have to fall 295 times per year for the risk of ICH to outweigh the benefits of AC for stroke prevention.12 More conservative estimates suggest that an individual on warfarin would have to fall 35 times per year for the risk to outweigh the benefit. Apixaban decreases bleeding risk when compared to warfarin. Similar modeling estimates an older adult on apixaban would have to fall 458 times per year for the risk of bleeding to outweigh the stroke benefit.13 Given the available evidence, the European Society of Cardiology Guidelines advise that “a history of falls is not an independent predictor of bleeding on an oral anticoagulant” and an increased risk of falls does not outweigh the benefits of AC in older patients with AF.14 Hospitalists may appropriately discontinue AC in certain clinical scenarios. Patients with a high bleeding risk from other comorbidities may have a disproportionate risk-to-benefit ratio that favors discontinuation of AC. Validated clinical decision tools such as HAS-BLED and ORBIT can characterize an individual's bleeding risk when making decisions around AC.15, 16 However, researchers validated these calculators for warfarin, which limits their use for DOACs. To address this, the “DOAC Score” bleeding risk calculator was reported to have a superior performance to HAS-BLED in quantifying bleeding risk in those taking DOAC therapy.17 These validated tools do not include fall risk as a risk factor for bleeding, highlighting the lack of evidence of falls as a significant contributor to overall bleeding risk. While these tools predict which patients have a higher bleeding risk, no score or calculated bleeding risk represents an absolute contraindication to AC. Rather, hospitalists should use this information when counseling patients on the risks and benefits of continued AC therapy. Consider deprescribing AC in the setting of severe frailty and at the end of life when goals of care shift to comfort and symptom focused. Physicians may recommend stopping AC when a patient's poor prognosis makes the risks outweight the benefits, however, these recommendations lack consensus and evidence. In these scenarios, engage in a patient-centered discussion using shared decision-making to determine the risks and benefits of discontinuing AC therapy. On the whole, the benefits of AC for stroke prevention outweigh the bleeding risks from falls in older patients with AF. Although a significant association between ICH and fall risk exists, current evidence indicates AC therapy reduces cardiovascular mortality, including ischemic stroke, in older patients who have a high fall risk and a moderate to high risk for stroke. Do not use fall risk alone as a reason to withhold or discontinue AC from older adults with AF. Consider using bleeding risk calculators (HAS-BLED, ORBIT, DOAC score) to determine the risk-benefit profile of continuing AC. Utilize shared decision-making with patients by understanding their preferences of balancing the benefits of stroke prevention against bleeding risk when making AC recommendations. Work to prevent future falls and fall-related injuries by referring to evidence-based fall prevention therapy programs and home safety evaluations. Finally, do not substitute aspirin for AC. Aspirin has no role in preventing stroke in AF and should not be used as an alternative to therapeutic AC to mitigate bleeding risk. Do not routinely discontinue AC based on fall risk alone for older patients with AF and moderate to high stroke risk. Use shared decision-making for select patient populations (high bleeding risk, history of serious bleeding, strong patient preference, and end-of-life) who may not derive as much benefit from AC. Decrease fall risk by referring high-risk patients to evidence-based fall prevention therapy programs and home safety evaluations. The hospitalist prepares discharge recommendations for the patient and refers her to physical therapy for balance training and occupational therapy for home safety assessment. Through shared decision-making, the hospitalist and patient opt to continue her apixaban given her stroke risk. Hospitalists routinely treat older patients who have fallen or have a high fall risk and have AF with moderate to high stroke risk. In the absence of serious bleeding, data does not support withholding or discontinuing AC for a history of falls or concern for future falls. Physicians' apprehensions about the bleeding risk linked to anticoagulant therapy lack substantial evidence. Underuse of AC in this population may worsen cardiovascular outcomes. The authors declare no conflict of interest.