Management of atrial fibrillation in older adults
BMJ 2024; 386 doi: https://doi.org/10.1136/bmj-2023-076246 (Published 17 September 2024) Cite this as: BMJ 2024;386:e076246
- Anna L Parks, assistant professor of medicine1,
- David S Frankel, professor of medicine2,
- Dae H Kim, associate professor of medicine3,
- Darae Ko, assistant professor of medicine3 4,
- Daniel B Kramer, associate professor of medicine5,
- Melis Lydston, knowledge specialist for research and instruction6,
- Margaret C Fang, professor of medicine7,
- Sachin J Shah, assistant professor of medicine8
- 1University of Utah, Division of Hematology and Hematologic Malignancies, Salt Lake City, UT, USA
- 2Cardiovascular Division, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, USA
- 3Hinda and Arthur Marcus Institute for Aging Research, Hebrew SeniorLife, Harvard Medical School, Boston, MA, USA
- 4Richard A and Susan F Smith Center for Outcomes Research, Beth Israel Deaconess Medical Center; Boston Medical Center, Section of Cardiovascular Medicine, Boston, MA, USA
- 5Richard A and Susan F Smith Center for Outcomes Research, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, USA
- 6Massachusetts General Hospital, Treadwell Virtual Library, Boston, MA, USA
- 7University of California, San Francisco, Division of Hospital Medicine, San Francisco, CA, USA
- 8Massachusetts General Hospital, Division of General Internal Medicine, Center for Aging and Serious Illness, and Harvard Medical School, Boston, MA, USA
- Correspondence to: S J Shah sshah{at}mgh.harvard.edu
Abstract
Most people with atrial fibrillation are older adults, in whom atrial fibrillation co-occurs with other chronic conditions, polypharmacy, and geriatric syndromes such as frailty. Yet most randomized controlled trials and expert guidelines use an age agnostic approach. Given the heterogeneity of aging, these data may not be universally applicable across the spectrum of older adults. This review synthesizes the available evidence and applies rigorous principles of aging science. After contextualizing the burden of comorbidities and geriatric syndromes in people with atrial fibrillation, it applies an aging focused approach to the pillars of atrial fibrillation management, describing screening for atrial fibrillation, lifestyle interventions, symptoms and complications, rate and rhythm control, coexisting heart failure, anticoagulation therapy, and left atrial appendage occlusion devices. Throughout, a framework is suggested that prioritizes patients’ goals and applies existing evidence to all older adults, whether atrial fibrillation is their sole condition, one among many, or a bystander at the end of life.
Introduction
Atrial fibrillation primarily affects adults over age 65 and is one of the most studied clinical conditions, yet a disconnect exists between the evidence and clinical care of older adults. Most randomized controlled trials (RCTs), guidelines, and recent reviews focus on atrial fibrillation in relative isolation. This approach is appropriate for robust older adults for whom atrial fibrillation is the single predominant disease.1 However, most people with atrial fibrillation are frail, have multiple comorbid conditions, and have at least one geriatric syndrome, and their individual priorities for treatment vary.2345 For this population, care of atrial fibrillation that applies existing evidence and guidelines without context may lead to more harm than benefit and may not cover what matters most to patients.. For example, in a hypothetical patient with five common chronic conditions, following clinical practice guidelines would entail 12 daily medications (some of which interact), 14 activities such as exercise and dietary recommendations (some of which conflict), multiple clinic visits, and thousands of dollars in out-of-pocket costs, with little mention of incorporating patients’ priorities.67 This review covers core advances in atrial fibrillation and in the management of older adults with multiple chronic conditions and uses a goals directed approach. We first characterize comorbidity and age related syndromes in patients with atrial fibrillation, as the rationale for an individualized approach to atrial fibrillation predicated on patients’ goals. Subsequently, we review evidence for the spectrum of atrial fibrillation care, from detection and preventive lifestyle interventions to symptom burden, rate and rhythm control, concomitant heart failure, anticoagulant therapy, and left atrial appendage closure. We take a structured approach to management and synthesize the evidence, where available, to apply it across the range of older adults—from fit and functional, to frail and multimorbid, to end of life.
Methods
We reviewed the available literature to ensure comprehensive inclusion of available atrial fibrillation literature specifically relevant to older adults. However, we did not review all extant atrial fibrillation literature, and we did not do a meta-analysis to produce effect estimates. This review was guided by the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 statement.8 A medical librarian (ML) did electronic searches for published literature by using Ovid MEDLINE (1946 to present), Embase.com (1947 to present), Web of Science Core Collection (1900 to present), Cochrane Central Register of Controlled Trials (CENTRAL) via Ovid (1991 to present) and ClinicalTrials.gov (1999 to present) in May 2023. The search strategy incorporated controlled vocabulary and free text synonyms for the concepts of older adults, RCTs, and atrial fibrillation. The full database search strategies are documented in the web appendix. No restriction on language was applied. A date limit of 2010 to present was used, as well as adapted methods and age filters.910 All identified studies were combined and de-duplicated using EndNote and EPPI-Reviewer and then uploaded into Covidence systematic review software. At least two contributors reviewed each abstract, and authors of each section reviewed full text articles for relevance. The European Society of Cardiology (ESC) guidelines were updated in August 2024; we therefore reference these instead of the earlier 2021 ESC guidelines.
Epidemiology of atrial fibrillation and multimorbidity in older adults
Atrial fibrillation is a disease of aging. The burden of disease is concentrated in older adults—80% of adults with atrial fibrillation are aged 65 years or older (fig 1).17 Not only is atrial fibrillation more prevalent in older adults, but multiple studies show that the incidence of atrial fibrillation increases with age. Even after adjustment for differences in known risk factors, the incidence of atrial fibrillation increases exponentially as people age.1418 An analysis of the Framingham Study indicated that the lifetime incidence of atrial fibrillation among 65 year olds was 33%.19
Incidence of atrial fibrillation (AF) by age. Comparison is made of incidence of AF by age in previously published studies. In cases in which studies report on gender and age specific incidence, group average is used for purposes of comparison. *Data from primary care practices, Scotland, April 2001-March 2002.11 †Atherosclerosis Risk in Communities Study, US, 1987-90.12 ‡General Practice Research Practice, UK, 1993-2005.13 §MOnica Risk, Genetics, Archiving and Monograph Consortium, Denmark, Finland, Sweden, Norway, Italy, 1982-2010.14 ¶Rotterdam Study, community based cohort study, Netherlands, 1990-93.15 **Olmsted County, Minnesota, community based cohort study, US, 1980-2000.16 GP=general practitioner
As with many diseases for which incidence is concentrated in older adults, atrial fibrillation co-occurs with geriatric syndromes. Geriatric syndromes are clinical conditions resulting from impairments in multiple organ systems that are common in older adults and affect quality of life and life expectancy.20 For example, in a national cohort of older Americans with atrial fibrillation, 20% reported an injurious fall and 25% reported that they were dependent on others to perform one or more activities of daily living.4 Frailty is common in older adults with atrial fibrillation (prevalence 39-51%).2122 Older adults with atrial fibrillation also have a high burden of multimorbidity and polypharmacy, which are both associated with worse physical and mental health and quality of life.23
The cumulative burden of illness is illustrated in multiple epidemiologic studies showing a mortality rate of 20-25% in the first year after a new diagnosis of atrial fibrillation in older adults.242526 Taken together, the care of older adults with atrial fibrillation clearly requires a whole person approach that accounts for the full burden of disease and the varied health goals among older adults with multiple chronic conditions.
Tailoring atrial fibrillation management for older adults
Many people are familiar with the growing tide of atrial fibrillation and the complexities of caring for an aging population, but how best to adapt clinical care remains challenging. The most recently updated atrial fibrillation guidelines acknowledge these complexities, recommending integrated, multidisciplinary clinics, a “personalized package of care,” exploring patients’ values, and shared decision making.27282930 But how to implement this in practice remains difficult. Figure 2 shows a suggested approach to tailoring the management of atrial fibrillation in older adults according to multimorbidity, frailty, and prognosis.
Proposed approach to tailor clinical management of atrial fibrillation (AF) to older adults
The problem with extrapolating guidelines to older adults
The clinical practice guidelines for the management of atrial fibrillation make recommendations based on the best available evidence from clinical trials with a goal of reducing symptoms of atrial fibrillation and preventing complications, especially stroke and related morbidity and mortality. In general, disease based guidelines are appropriate for patients with a single predominant disease who share the same health outcome goals, such as those with atrial fibrillation who have few comorbidities and few competing health priorities, take a small number of concurrent medications, and have more than 10 years of remaining life expectancy.12731
Simply extrapolating guideline recommendations to all older adults is challenging because those with multimorbidity and frailty are under-represented in RCTs. Even trials that do include older adults may inadequately capture the heterogeneous health states of older adults, which then are not adequately covered in the guidelines.3233 Older adults’ multimorbidity and associated treatment burden (for example, polypharmacy and drug-drug interactions) can negatively affect daily function and quality of life and increase the risk of harm from drugs and procedural therapy for atrial fibrillation.4 Although preventing stroke and related morbidity and mortality remains an important goal, this benefit may be delayed or offset by treatment related adverse events, such as intracranial hemorrhage and major bleeding from oral anticoagulants and procedural complications after catheter ablation and left atrial appendage occlusion.25343536 Conversely, the paucity of data in older adults means that they may derive even greater benefit from treatment that has not yet been detected. Finally, older adults with multimorbidity and frailty often have competing health priorities and significant person-to-person variation in health goals.537
Approaches to achieving individualized care
One evidence based method to achieve whole person, individualized care in atrial fibrillation is to align treatment with patients’ health priorities.38 This approach acknowledges that older adults’ care preferences vary and that decision making requires explicit consideration of the wide array of care available for atrial fibrillation according to potential harms, burdens, benefits, and prognosis.6 To inform decision making, clinicians should do an assessment of each patient’s level of function and frailty. One practical means of assessing fitness and frailty is the Clinical Frailty Scale, which is a judgment based, nine point scale ranging from very fit to terminally ill.39 The most in-depth method is comprehensive geriatric assessment, an interdisciplinary, multidimensional assessment of an older person’s medical, psychological, and functional capacity.40 Estimating prognosis is notoriously difficult, but validated prognostic indices may be helpful for improving assumptions that influence clinical decisions. ePrognosis (https://eprognosis.ucsf.edu/) is one repository that links to prognostic tools with the highest quality evidence for prognostic accuracy for multimorbid older adults that can be helpful for estimating both life expectancy and health trajectory. With this information, the ultimate treatment decision can be determined by individuals’ specific goals. Importantly, care informed by priorities may lead to unanticipated outcomes—for instance, guideline directed care meets the goals of some multimorbid patients, and some fit patients may opt against it, underscoring the importance of individualized care.
One such approach example is Patient Priorities Care, which to date is the intervention shown to be most effective for reducing treatment burden and unwanted care in patients with multimorbidity.38 Using this approach, the patient’s health priorities and specific, measurable, achievable, relevant, and time-bound (“SMART”) goals are identified and shared by the clinical team, including the primary care physician, specialists, and other health professionals.41 Patients’ personal health outcome goals may include function (for example, walk two blocks without shortness of breath), life prolongation (for example, see my grandson graduate in five years), wellbeing (for example, reduce anxiety about developing stroke or major bleeding), or occupational role (for example, work for three more years).1 Then patients’ care preferences are assessed in terms of “cost” or the amount of workload that one is willing to do for the desired health outcome goal.5 The patient’s “cost” includes the financial cost and time, as well as treatment related workload, such as interaction with clinicians (for example, office visits and reconciling conflicting recommendations), medication management (for example, following a complex administration schedule before procedures and dealing with drug-drug interactions), self-management tasks (for example, adhering to a consistent dietary pattern while taking a vitamin K antagonist), diagnostic and laboratory testing (for example, monitoring of the international normalized ratio and kidney function), and discomfort, complications, and delayed recovery from invasive procedures.1 The patient and the care team make clinical management decisions in order to maximize the value to the patient, conceptualized as personal health outcome gain over the patient’s “cost.”5 Treatments and care activities that offer high value are prioritized. An important limitation of this process is that even with clearly defined goals, often few data are available to suggest which treatment strategy is most likely to achieve them.42
Treatment de-escalation
For older adults with atrial fibrillation nearing the end of life (that is, remaining life expectancy less than one year), palliative care focusing on symptoms (for example, palpitations and shortness of breath), quality of life, and comfort is appropriate. De-escalation of care should be considered when the risk of treatment related adverse events is deemed to outweigh the benefit or when the benefit is uncertain in the setting of advanced illness. The decision to stop anticoagulation should be individualized after consideration of the patient’s prognosis (for example, terminal illness), risk of major bleeding versus stroke within the remaining lifespan, factors influencing quality of life (for example, laboratory monitoring), and the patient’s and family’s preference.
Population level systematic screening
Given its high prevalence in older adults, a concerted effort has been made to determine the value of screening for atrial fibrillation. Although the case for screening following a stroke or transient ischemic attack is clear, the evidence for primary screening is unclear. RCTs have shown that screening can increase the rate of diagnosis of atrial fibrillation. Whether this translates into fewer strokes and better health has not been demonstrated (fig 3).43
Screening modalities and associated yield. ECG=electrocardiograph
Multiple, well conducted RCTs have shown that electrocardiography based screening is effective for identifying cases of atrial fibrillation in older adults beyond usual care.43 The LOOP Study represents the opposite end of the screening intensity spectrum in older adults—6000 adults aged 70 years and older were randomized to implantable loop recorder versus usual care.44 Over a median of 5.3 years of follow-up, atrial fibrillation was diagnosed in 32% of the screening group and 12% of the usual care group, representing an excess diagnosis rate of 5.6 per 100 person years. When an episode of atrial fibrillation lasting more than six minutes was detected, participants were advised to start anticoagulants. Although 93% of patients heeded the recommendation and started anticoagulants, the study failed to show a significant reduction in the primary outcome of strokes or systemic embolism (0.88/100 person years with screening versus 1.09/100 person years with usual care).44454647 The STROKESTOP study randomized all adults aged 75-76 years in two regions in Sweden, whereby half were invited to screen for atrial fibrillation by using a handheld single lead electrocardiograph for 30 seconds twice a day for 14 days.48 Those with 30 seconds or more of atrial fibrillation were referred to a cardiologist for an assessment and anticoagulation recommendation. The screening intervention reduced the composite outcome of stroke, systemic embolism, bleeding resulting in hospital admission, and all cause death by 2.3 per 1000 person years—a small but significant finding. Enthusiasm is tempered because of the small effect size and uncertainty inherent in pragmatic screening studies (for example, unmasked intervention, use of claims defined endpoints, and that people with existing atrial fibrillation were included in the analysis). LOOP, STROKESTOP, and several ongoing studies are the first of a new series of RCTs that seek to answer the ultimate question in screening for atrial fibrillation among older adults—does screening for atrial fibrillation reduce the rates of stroke and stroke related disability?4950 Wearable consumer devices such as smartwatches can also screen for atrial fibrillation. Non-randomized studies have shown a reasonable yield of diagnosis of atrial fibrillation when an irregular heartbeat is detected by the device’s pulse wave algorithm. The Apple Watch study triggered a notification of an irregular heartbeat in 0.52% of participants; among those who went on to complete a seven day electrocardiography patch, 34% were found to have atrial fibrillation.51 A similar yield was observed in a study of the FitBit device.47 In the Huawei Heart Study, the device screened more often, and the algorithm required a longer duration of irregular pulse rate to trigger a notification. The Huawei algorithm triggered a notification in 0.23% of participants, and 87% were found to have atrial fibrillation.52 Analogous to population level screening, this is an area of active investigation with improvements in study design, hardware, software, and ongoing trials to determine whether consumer device based screening improves outcomes in older adults (fig 4).
Summary of recommendations for atrial fibrillation screening in older adults. ESC=European Society of Cardiology; USPSTF=United States Preventive Services Task Force
Lifestyle intervention for secondary prevention
Many older patients have potentially reversible risk factors that may influence the incidence, burden, and progression of atrial fibrillation. Obesity is a well established risk factor for development of atrial fibrillation. Observational studies that enrolled primarily older adults with atrial fibrillation in a supervised weight loss program suggest a dose dependent effect of weight loss on reversal of atrial fibrillation from persistent to either paroxysmal or no atrial fibrillation.5354
The relation between physical activity and incident atrial fibrillation is more complex.555657585960 A prospective study of older adults examining the association of exercise patterns and incidence of atrial fibrillation showed a U-shaped relation: although light and moderate intensity exercise was protective against development of atrial fibrillation, both inactivity and high intensity exercise were associated with increased incident atrial fibrillation.61 The Systolic Blood Pressure Intervention Trial (SPRINT), an RCT randomizing 8022 older adults at high risk of cardiovascular disease, showed that achieving systolic blood pressure <120 mm Hg led to a 26% lower risk of development of atrial fibrillation compared with targeting a goal of <140 mm Hg (hazard ratio 0.74, 95% confidence interval (CI) 0.56 to 0.98).62 Among adults of all ages, alcohol consumption has been confirmed as a trigger of atrial fibrillation, and abstinence reduces recurrence and symptoms of atrial fibrillation, whereas caffeine does not seem to have an effect.63646566
Data largely support obesity prevention, weight loss, moderate physical activity, control of blood pressure, and possibly alcohol avoidance to prevent and manage atrial fibrillation in functional patients (fig 5). By definition, many of these interventions were highly structured programs that required patients to be fit enough to participate, so how to adapt these findings to multimorbid/frail older adults should be a focus of future research. Lifestyle changes likely will not meet the goals of or may have limited benefit for people with short life expectancies.
Summary of recommendations for atrial fibrillation (AF) lifestyle interventions
Symptoms and clinical manifestations of atrial fibrillation
Symptoms of atrial fibrillation can be non-specific and intermittent and may manifest differently in older than in younger adults. Symptoms such as palpitations, lightheadedness, breathlessness, and chest discomfort are prevalent in all age groups, whereas older adults are more likely to experience fatigue or generalized weakness as their main complaint.67 Syncope is uncommon in the absence of additional conduction disease, such as significant conversion pauses or rapidly conducting accessory pathways. Attributing any of these symptoms to atrial fibrillation may be more complex in older adults with multiple comorbidities, as concurrent lung disease, heart failure, other arrhythmias, or medication effects may confound attribution to specific causes.
Alleviating symptoms is a key goal for many patients with atrial fibrillation.68 Multiple clinical trials have used clinician reported and patient reported outcome measures to document the effect of various treatment strategies for atrial fibrillation on quality of life. Validated patient reported measures can also be used clinically.69 One example of a clinician reported outcome is the European Heart Rhythm Association (EHRA) symptom classification for atrial fibrillation, which uses an ordinal scale (1=no symptoms; 4=daily activities suspended).33 The most widely used patient reported outcome, the Atrial Fibrillation Effect on QualiTy-of-life (AFEQT) questionnaire,70 measures domains including symptoms, social functioning, physical functioning, emotional functioning, treatment concerns, and treatment satisfaction. When an atrial fibrillation specific patient reported outcome measure was integrated longitudinally into clinical practice, worse scores correlated with higher atrial fibrillation burden and higher healthcare utilization, as well as greater use of rhythm control over rate control.71 Systematic measurement of patients’ experience can help to capture symptoms of atrial fibrillation and their broader effect on quality of life, articulate treatment goals, guide treatment recommendations, and track progress.
Patients with symptoms potentially attributable to atrial fibrillation—or its treatment with anti-arrhythmic or rate control agents—may benefit from specialty referral, including formal application of quality of life measures alongside other cardiovascular testing, to elucidate the contribution of atrial fibrillation to patients’ lived experiences.
Rate and rhythm control
The recently published 2023 American College of Cardiology/American Heart Association/American College of Clinical Pharmacy/Heart Rhythm Society (ACC/AHA/ACCP/HRS) guideline for the diagnosis and management of atrial fibrillation recommended a substantial shift toward greater and earlier use of rhythm control over rate control in both paroxysmal and persistent atrial fibrillation compared with earlier guidelines.72 In the trials informing these guidelines, rhythm control improved a range of clinical outcomes with reassuring safety data, and older adults were relatively well represented, although the generalizability to the frail, multimorbid population remains a concern.
These guidelines give a strong recommendation for rhythm control in patients with heart failure and a moderate recommendation in symptomatic atrial fibrillation and in the first year after diagnosis of atrial fibrillation, among other groups.72 The guidelines draw on rigorous clinical trials leveraging modern technology, which consistently favor rhythm control over rate control for clinical outcomes. For example, the Early Rhythm-Control Therapy in Patients with Atrial Fibrillation (EAST-AFNET4) trial randomized 2789 people with atrial fibrillation (mean age ~70; 43% with at least mild cognitive impairment, 88% with hypertension, 28% with heart failure, and 12% with chronic kidney disease) who had been given a diagnosis of atrial fibrillation in the previous year to early rhythm control with anti-arrhythmic drugs or catheter ablation versus usual care.61 Usual care was limited to management of symptoms, predominantly attempted with rate control alone. The trial was stopped at a median follow-up of five years, with patients randomized to early rhythm control experiencing a significantly lower rate of death from cardiovascular causes, stroke, or admission to hospital for worsening of heart failure or acute coronary syndromes (3.9 v 5.0 per 100 person years; hazard ratio 0.79, 95% CI 0.66 to 0.94). These findings were consistent across components of the composite endpoint, including death from cardiovascular causes. Notably, however, quality of life measures such as the EHRA score did not differ between groups.
While EAST-AFNET4 included both anti-arrhythmic drugs and ablation as treatment strategies, selected at the investigator’s discretion, other clinical trials have established that catheter ablation is superior to drug treatment in the maintenance of sinus rhythm and has low procedural complication rates. The Effect of Catheter Ablation vs Antiarrhythmic Drug Therapy on Mortality, Stroke, Bleeding, and Cardiac Arrest Among Patients with Atrial Fibrillation trial (CABANA73) randomized 2204 patients with symptomatic atrial fibrillation (median age 68; 15% with heart failure, 25% with diabetes, 23% with sleep apnea, and 19% with coronary artery disease) to catheter ablation versus drug therapy, including rate and/or rhythm control. In the intention-to-treat analysis, the primary endpoint of death, disabling stroke, serious bleeding, and cardiac arrest was not significantly reduced (hazard ratio 0.86, 95% CI 0.65 to 1.15). The secondary endpoint of death and hospital admission for cardiovascular disease was significantly reduced in the group randomized to catheter ablation (hazard ratio 0.83, 95% CI 0.74 to 0.93). Furthermore, both atrial fibrillation-free survival and quality of life were higher among patients randomized to catheter ablation.74 CABANA’s findings are consistent with meta-analyses showing superior freedom from symptomatic atrial arrhythmias and reduced health resource use with catheter ablation than with anti-arrhythmic drugs.75
Safety data from CABANA inform the risk-benefit discussion, as both clinicians and patients may be wary of ablation owing to perceived risk of complications and recovery time. Cardiac tamponade occurred in 0.8% of patients undergoing catheter ablation. Minor hematomas (2.3%) and pseudoaneurysms (1.1%) were the most common adverse events. Importantly, the rate of complications during catheter ablation in CABANA was not significantly increased among older adults.76 In the medical therapy group, 1.6% had thyroid disorders and 0.8% had additional arrhythmias. These results make clear that both ablation and medical therapy have adverse effects, but the incidence overall is low. The generalizability of these findings outside of highly selected settings with experienced proceduralists remains unknown.
Early data suggest that catheter ablation may improve cognition in older adults with atrial fibrillation. Observational studies have associated atrial fibrillation with reduced brain volume, lower cognitive function, and greater risk of dementia.7778 Whether restoration of sinus rhythm can attenuate these risks is an area of active investigation.79 For example, 96 patients with atrial fibrillation refractory to at least one anti-arrhythmic drug were randomized to continued medical therapy or ablation.8081 Although patients randomized to ablation experienced more postoperative cognitive dysfunction (14% v 2%; P=0.03), likely related to anesthesia and silent cerebral emboli, this effect fully resolved by one year (0% v 2%; P≥0.05). Furthermore, 14% of patients randomized to ablation derived cognitive improvement by one year, compared with 0% in the continued medical therapy arm (P=0.007).
As above, the mean age of patients included was 70 in EAST-AFNET4 and 68 in CABANA. Thus, fit and functional patients older than 65 should be offered early rhythm control, with catheter ablation preferred given its demonstrated superiority in maintaining sinus rhythm and improving quality of life (fig 6). However, deciding between ablation and anti-arrhythmic drugs ought to be individualized for multimorbid/frail older adults. Anti-arrhythmic drugs have extensive interactions with other drugs and pharmacokinetic interactions with impaired hepatic and/or renal function. These risks must be carefully balanced against short term risks of anesthesia and procedural complications, with treatment tailored to individual circumstances, particularly among multimorbid/frail patients. Although patients with more comorbidities and shorter life expectancy might be anticipated to derive less survival and quality of life benefit from early rhythm control, this is not necessarily true (for example, for heart failure). Multimorbidity makes rhythm control challenging in older adults with atrial fibrillation, but, for some, reducing the burden of atrial fibrillation may have a greater impact. At the end of life, rhythm control is rarely an important consideration, other than palliating symptoms.
Summary of recommendations for rate and rhythm control in older adults with atrial fibrillation (AF). AAD=anti-arrhythmic drug; LVEF=left ventricular ejection fraction
Heart failure and atrial fibrillation
Management of atrial fibrillation in heart failure requires special consideration because a bidirectional relation exists between atrial fibrillation and heart failure, wherein heart failure increases the risk of incident atrial fibrillation and atrial fibrillation worsens heart failure outcomes. Atrial fibrillation and heart failure often coexist; among Medicare beneficiaries aged ≥65 years, 29% of patients with heart failure are estimated to have atrial fibrillation and 23% of patients with atrial fibrillation are estimated to have heart failure.2982 Coexistence also leads to compounded symptom burden, worsened quality of life, and mortality, emphasizing the importance of considering prognosis when making treatment decisions.83 Treatment recommendations in this setting are primarily extrapolated from data in younger populations. Presence of coexisting heart failure with reduced ejection fraction (HFrEF) favors early rhythm control to maintain sinus rhythm, and catheter ablation should be considered over long term anti-arrhythmic therapy.3183 The CASTLE atrial fibrillation trial randomized patients (median age 64; ~73% with hypertension, ~30% with diabetes, and ~30% with myocardial infarction) with paroxysmal or persistent atrial fibrillation, heart failure with left ventricular ejection fraction <35%, New York Heart Association class II to IV, and an implantable cardiac device or cardiac resynchronization therapy device to ablation or standard medical therapy.84 Ablation reduced the risk of all cause death by 47% and admission to hospital for heart failure by 44%. In a meta-analysis of 11 RCTs comparing ablation and medical therapy, ablation was associated with lower risks of all cause death and readmission to hospital and greater improvement in left ventricular ejection fraction.85 Although the primary endpoint was not reduced with ablation in the intention-to-treat analysis of the entire CABANA population, it was significantly reduced among patients with heart failure (9.0% v 12.3%; hazard ratio 0.64, 95% CI 0.41 to 0.99), including a significant reduction in all cause mortality (6.1% v 9.3%; 0.57, 0.33 to 0.96)).86 In HFrEF, non-dihydropyridine calcium channel blockers (diltiazem and verapamil) and dronedarone are contraindicated owing to their negative inotropic effect, and flecainide and sotalol are contraindicated owing to their pro-arrhythmic effects.29 Another option to reduce polypharmacy and alleviate refractory symptoms is pacemaker implantation with or without atrioventricular junction ablation on the basis of RCT data showing that this strategy can reduce exacerbations of heart failure and improve quality of life.
Balancing risks and benefits of oral anticoagulants for thromboprophylaxis
Oral anticoagulants reduce atrial fibrillation related strokes but at the expense of increased bleeding.87 Aging influences multiple facets of the risk-benefit assessment, or “net clinical benefit,” of anticoagulation.88 Stroke risk increases with advancing age, but so too does anticoagulation related bleeding.8990919293 Relative to the population prevalence of atrial fibrillation, foundational RCTs of anticoagulants for stroke prevention under-enrolled people aged 80 years and over (fig 7).94 In addition, as people age and reach their full life expectancy, the potential benefit of stroke prevention through anticoagulation may be tempered by the competing risk of death and disability from non-stroke related causes.9596 Current consensus guidelines consider all patients aged 75 years and older as being at high risk for atrial fibrillation related stroke and recommend anticoagulation.27283072 However, the recently released ESC guidelines explicitly state that, “Not enough evidence is available for OAC in elderly patients, frail polypharmacy patients, those with cognitive impairment/dementia….”27. Importantly, the 2023 ACC/AHA/ACCP/HRS guidelines explicitly recommend against using bleeding risk prediction scores (for example, HAS-BLED, HEMORR2HAGES, ATRIA) because they poorly discriminate between people who develop bleeding and those who do not, and instead suggest mitigating reversible risk factors for bleeding.72 Guidelines generally recommend direct oral anticoagulants (DOACs) over warfarin because of similar efficacy, generally lower bleeding rates, fewer drug-drug interactions, and less need for monitoring.27283072 Of note, clinicians should avoid the inclination to substitute aspirin for anticoagulation as a compromise, as aspirin use in atrial fibrillation is now considered a class III harm on the basis of studies showing that it confers equivalent bleeding risk but is inferior for stroke prevention.97 The 2023 update to the American Geriatrics Society Beers Criteria for potentially inappropriate medication use in older adults also recommends DOACs over warfarin when starting anticoagulation and goes a step further in recommending apixaban over other DOACs owing to a lower incidence of bleeding.98 Comparing among DOACs remains challenging, and the results of several ongoing head-to-head RCTs will be very informative (for example, NCT03266783, NCT04642430, NCT03129555, NCT03129490). For the subset of fit and functional patients with atrial fibrillation, little hesitation about recommending anticoagulants is warranted. Recent studies have helped to apply these guidelines to the nuanced needs of multimorbid/frail older adults.
Age distribution of participants in clinical trials for stroke prevention in patients with atrial fibrillation compared with US prevalent population. Adapted from Lau et al.94 This figure is eligible to be reused through the CC-BY license
Several studies challenge the paradigm that anticoagulants are of net benefit for all older adults.99 Observational studies have raised concerns about potential underuse of anticoagulants, estimating that more than 40% of patients considered by guidelines to be eligible for anticoagulation are not receiving it.100101 However, deeper examination finds that anticoagulants may not be appropriate for all of these patients; a recent cross sectional study found that nearly one third of nursing home residents with atrial fibrillation and advanced dementia remained on anticoagulation in the last six months of life, and those with the most severe markers of dementia and at highest bleeding risk were counterintuitively more likely to be treated with anticoagulation.102 Frailty and geriatric syndromes are prevalent among older adults and associated with a reduced likelihood of prescription of anticoagulant.4 A study that modeled the net clinical benefit of anticoagulants over the lifespan found that the benefits of anticoagulants decreases substantially owing to the competing risk of death from non-atrial fibrillation causes.103 These observations argue for a more nuanced approach to the application of current guidelines that accounts for advancing age and comorbidity burden.
Anticoagulation in frail or multimorbid older adults
Results of recent studies give guidance for anticoagulation therapy in multimorbid/frail older adults. Reduced dose anticoagulation has been investigated. In the Edoxaban Low-Dose for Elder Care Atrial Fibrillation Patients (ELDERCARE-AF) trial, 984 Japanese patients 80 years of age or older with atrial fibrillation who were considered inappropriate candidates for standard dose anticoagulants—because of severe chronic kidney disease, concurrent drugs with high bleeding risk, low body weight, or history of significant bleeding—were randomized to a reduced dose of edoxaban versus placebo.104 Reduced dose edoxaban was superior to placebo in preventing stroke or systemic embolism (annualized rate 2.3% in the edoxaban group versus 6.7% in the placebo group; hazard ratio 0.34, 95% CI 0.19 to 0.61) without increasing bleeding risk (hazard ratio 1.87, 0.90 to 3.89) or overall mortality (0.97, 0.69 to 1.36). One target trial emulation analysis of Medicare data investigated whether the presence of frailty affects the effectiveness and safety of DOACs versus warfarin.2 It found that apixaban reduced the relative risk of a composite endpoint of death, ischemic stroke, or major bleeding by one third compared with warfarin across all frailty subgroups, whereas dabigatran and rivaroxaban were associated with lower event rates only among non-frail patients. In addition to recommending apixaban on the basis of the strength of observational data showing lower bleeding risk, the Beers criteria also recommended against switching older patients maintained on warfarin to DOACs; this recommendation is supported by the open label, randomized superiority FRAIL-AF trial conducted in 1330 older, frail patients, which was halted for futility after an interim analysis showed that switching from warfarin to DOAC was associated with more bleeding complications (17.8 v 10.5 per 100 patient years after complete follow-up; hazard ratio 1.69, 95% CI 1.23 to 2.32) without a reduction in thromboembolic complications (hazard ratio 1.26, 0.60 to 2.61).104
Chronic kidney disease
Older adults with atrial fibrillation and chronic kidney disease (CKD) merit special discussion because renal disease is associated with a higher incidence of atrial fibrillation and higher risk of stroke and bleeding events.105106107 Anticoagulation with DOACs for patients with mild or moderate CKD (estimated glomerular filtration rate 30-59 mL/min or stage 3a and 3b) may be beneficial.108 For patients with atrial fibrillation and end stage kidney disease, no RCT has shown net benefit from anticoagulation.109110111 The use of anticoagulants in dialysis is associated with high bleeding rates and mortality. To provide better evidence, several trials are under way examining anticoagulation versus placebo in patients with end stage kidney disease.112 Whether DOACs or warfarin are favored in end stage kidney disease is also unclear, although early trial data suggest that factor Xa inhibitors may be acceptable alternatives to warfarin.112113114 Guidelines support either warfarin or DOACs in patients with mild or moderate CKD, with a weak recommendation for warfarin or apixaban in those with severe CKD.72115
Anticoagulant associated hemorrhage
Several interventions can reduce the risk of anticoagulant associated hemorrhage (fig 8). One intervention that can reduce the bleeding risk associated with anticoagulants is de-prescribing antiplatelet agents, which confer a 1.5-fold to twofold increased bleeding risk without reducing thrombosis.116117118 Expert consensus guidance on concurrent antiplatelet and anticoagulant medications recommends avoiding aspirin for primary prevention of cardiovascular disease, avoiding “triple therapy” (dual antiplatelet therapy plus anticoagulation) except very short duration in high risk clinical circumstances (for example, recent percutaneous coronary intervention), anticoagulation monotherapy for long term treatment of patients with an indication for antiplatelet therapy and anticoagulation (for example, stable ischemic heart disease and six to 12 months after acute coronary syndrome or percutaneous coronary intervention), and anticoagulation monotherapy for patients with cerebrovascular disease without carotid stenting.119 Reversible risk factors for bleeding, including hypertension and non-steroidal anti-inflammatory drugs, should be reduced and proton pump inhibitors considered to prevent gastrointestinal bleeding in patients taking two or more antithrombotic agents.120
Modifiable risk factors for bleeding in older adults on anticoagulants. NSAID=non-steroidal anti-inflammatory drug
In summary, each facet of the decision to start and continue anticoagulation to prevent atrial fibrillation related stroke among older adults requires careful, individualized decision making (fig 9). For patients with longer life expectancies, anticoagulants have the greatest potential effect, and the benefit diminishes as people age. DOACs are favored over warfarin for most patients, including those with frailty and multimorbidity, but decision making must incorporate cost/insurance, the patient’s preferences, and patient related factors (such as once a day dosing and drug interactions). Providers should endeavor to stop anticoagulation in patients at the end of life, who likely do not have sufficient life expectancy to accrue the benefits and can suffer harm.121
Summary of recommendations for anticoagulation in older adults with atrial fibrillation. DOAC=direct oral anticoagulant
Left atrial appendage closure
Transcatheter endovascular left atrial appendage closure (LAAC) was approved by the US Food and Drug Administration (FDA) in 2015 as an alternative to anticoagulation for stroke prevention in atrial fibrillation. The rationale behind LAAC stems from the observations that most left atrial thrombi in atrial fibrillation form in the left atrial appendage, an outpouching of the left atrium that is prone to blood stasis, and the risk of embolic complications after cardioversion is very low in the absence of left atrial appendage thrombus.122123124 Compared with warfarin, sealing off the left atrial appendage is as effective at preventing thromboembolic stroke and avoids bleeding complications. Data are less robust comparing LAAC with DOACs, which are generally more effective and safer than warfarin, supporting caution while additional studies are completed.
The safety and efficacy of LAAC was tested in two non-inferiority RCTs that compared LAAC and warfarin, the standard of care at the time when the trials were completed.122125 Initially, the trial results were mixed. Whereas the PROTECT AF (WATCHMAN left atrial appendage system for embolic PROTECTion in 707 patients with Atrial Fibrillation) trial randomizing patients to LAAC versus warfarin showed non-inferiority for the primary outcome of stroke, systemic embolism, or cardiovascular death (rate ratio 0.60, 95% CI 0.41 to 1.05), the PREVAIL (Prospective Randomized Evaluation of the Watchman LAA Closure Device In Patients With Atrial Fibrillation Versus Long Term Warfarin Therapy) trial, which enrolled 407 older and more multimorbid patients, failed to show non-inferiority. These concerns were mitigated in the five year patient level meta-analysis that showed comparable rates of the primary outcome (2.8/100 person years in the LAAC group versus 3.4/100 person years in the warfarin group). Additionally, patients randomized to LAAC had lower rates of the secondary outcomes of hemorrhagic strokes and disabling strokes.126 Importantly, LAAC was tested against warfarin, which is waning in preference to drugs such as apixaban, which is more effective at stroke prevention and causes less bleeding. Early evidence suggests that LAAC is non-inferior to DOACs, and ongoing RCTs will add to current evidence.5960127 Following FDA approval of the LAAC devices, such studies have been difficult to complete; for example, the ASAP-TOO trial testing LAAC versus no anticoagulation was terminated early owing to slow recruitment.53128
As with any procedure, complications are a key consideration for older adults. The PREVAIL trial reported a 4.2% complication rate (for example, cardiac tamponade, device embolization requiring retrieval, and procedure related major bleeding), and post-approval observational analyses show that rates are decreasing, now down to 2.2%.129 Post-approval observational studies indicate that adults aged >80 years have slightly higher rates of in-hospital adverse outcomes.5960 Looking beyond in-hospital outcomes, frail older adults have notably higher rates of adverse events following hospital discharge.128
A critical practice-evidence gap exists in selecting patients who will most likely benefit from LAAC. Although LAAC will most certainly reduce the risk of thromboembolic complications compared with no oral anticoagulant, patients who are least likely to receive oral anticoagulants owing to concerns about major bleeding were most likely excluded from the pre-approval trials. Patients with geriatric comorbidities (for example, dementia, frailty) are less likely to receive oral anticoagulants, but they are more likely to sustain serious procedural complications and more likely to bleed from lifelong post-LAAC antiplatelet therapy with aspirin.82
Use of LAAC is approved by the Centers for Medicare and Medicaid Services (CMS) and recommended by practice guidelines only for patients with contraindications to long term oral anticoagulants.5354 Similarly, the 2023 ACC/AHA/ACCP/HRS guideline for the diagnosis and management of atrial fibrillation provides a moderate recommendation for LAAC when patients have a contraindication to long term anticoagulation owing to a non-reversible cause and a weak recommendation when patients have a high risk of major bleeding.72 Interestingly, CMS also mandates documented use of patient decision aids for shared decision making, but no specific decision aid is suggested or provided.130
Older patients were well represented in the RCTs that support LAAC, so fit and functional patients older than 65 should be offered LAAC in accordance with the guideline recommendations (fig 10). For multimorbid/frail older adults, the decision should be individualized considering risk tolerance, periprocedural complications, and the long term benefit of thromboprophylaxis without anticoagulant use. Patients at the end of life are not appropriate candidates for LAAC.
Summary of recommendations for left atrial appendage closure in older adults
Emerging treatments
Novel therapeutic strategies may affect the care of older adults with atrial fibrillation. For screening, the fundamental question of whether enhanced detection and treatment of atrial fibrillation identified via consumer devices reduces strokes is under evaluation in the HEARTLINE study randomizing participants to Apple Watch screening, with anticipated completion in 2025.131 Older adults with atrial fibrillation are a key demographic under study for RCTs of novel anticoagulants targeting different coagulation cascade proteins (factor XIa and XIIa) that are postulated to reduce risk of thrombosis with attenuated bleeding risks.121 Clinical trials are also assessing the optimal anticoagulation strategy after catheter ablation, including whether it can be discontinued, as well as the use of intermittent anticoagulation guided by continuous rhythm monitoring.132 Finally, although shared decision making is a laudable goal for older adults and is recommended by clinical guidelines, several trials are under way examining both the optimal format and whether this improves clinical outcomes.133134
Guidelines
Clinicians have multiple guidance statements covering management of atrial fibrillation to choose from. During the editorial process for this review, the ESC guidelines were updated in August 2024.27 These guidelines offer dedicated sections on anticoagulation management in older adults and those with cognitive dysfunction and are largely concordant with evidence and guidance we report. The 2024 ESC guidance newly emphasizes the lack of evidence to support anticoagulation in frail, multimorbid older adults, including those with dementia. The 2023 update to the ACC/AHA/ACCP/HRS guideline for the diagnosis and management of atrial fibrillation incorporates a brief discussion of shared decision making, but mainly to highlight lack of data on improvement in clinical outcomes.72 Guidance from the National Institute for Health and Care Excellence, from the UK and updated in 2021, calls for a “personalised package of care,” including psychological and social support, contact information, and educational information.30 Finally, the Canadian Cardiology Society guidance released in 2020 also promotes a multidisciplinary model of care for atrial fibrillation.28
Conclusion
Decades of research and clinical care dedicated to atrial fibrillation have dramatically reduced morbidity and mortality. Yet atrial fibrillation remains a quintessential disease of aging, for which exclusive use of a disease focused approach can lead one to miss the forest for the trees. We present a framework to apply existing evidence to the heterogeneous needs of older adults. In doing so, we also highlight the need for a dedicated effort to strengthen the evidence that informs individualized care for the growing population of older adults with atrial fibrillation. The increasing number of pragmatic clinical trials enrolling complex, frail participants and examining outcomes that are important to older adults such as cognition is heartening. Future clinical trials should avoid stringent exclusion criteria and enroll patients reflective of those we encounter in the clinic. Building on the increasing use of patient reported outcomes, primary and secondary outcomes in clinical trials should be grounded in patients’ priorities, emphasizing quality of life and function over narrow clinical events. Evidence based shared decision making approaches, which are most well developed for anticoagulation, should be expanded to include all decisions about management of atrial fibrillation. Recognizing that most patients with atrial fibrillation must also adhere to guidelines for multiple other burdensome chronic conditions, we should embrace a framework predicated on care directed to patients’ goals.
Questions for future research
Does population level screening for atrial fibrillation result in net benefit for older adults, particularly those who are multimorbid/frail?
What are the optimal priorities directed outcome measures for symptoms of atrial fibrillation and the optimal clinical implementation strategy?
Which multimorbid, frail older adults benefit from early rhythm control over rate control?
Is one direct oral anticoagulant superior to others for older adults?
Which older adults benefit from left atrial appendage occlusion rather than anticoagulation?
Glossary of abbreviations
ACC—American College of Cardiology
ACCP—American College of Clinical Pharmacy
AHA—American Heart Association
CI—confidence interval
CKD—chronic kidney disease
CMS—Centers for Medicare and Medicaid Services
DOAC—direct oral anticoagulant
EHRA—European Heart Rhythm Association
ESC—European Society of Cardiology
FDA—Food and Drug Administration
HFrEF—heart failure with reduced ejection fraction
HRS—Heart Rhythm Society
LAAC— left atrial appendage closure
RCT—randomized controlled trial
How patients were involved in creation of this article
A patient advocate and frequent research collaborator from the non-profit Arrhythmia Alliance (www.heartrhythmalliance.org) reviewed a draft of this review. They offered their perspective on aspects of atrial fibrillation that are of particular concern to older adults. In particular, they emphasized the need to distinguish between interventions for atrial fibrillation that treat symptoms (for example, ablation) and those that are preventive (for example, anticoagulation). They also highlighted ways to engage patients with scientific literature. Their suggested edits were incorporated into the final version where applicable.
Footnotes
Series explanation: State of the Art Reviews are commissioned on the basis of their relevance to academics and specialists in the US and internationally. For this reason they are written predominantly by US authors
Contributors: All authors contributed to the conception and analysis of the manuscript, drafted the manuscript, revised it critically for important intellectual content, and approved the final version to be published. SJS is the guarantor.
Competing interests: We have read and understood the BMJ policy on declaration of interests and declare the following interests: none.
Provenance and peer review: Commissioned; externally peer reviewed.