Direct Oral Anticoagulant Use in Older Adults with Atrial Fibrillation: Challenges and Solutions

AF is the most common arrhythmia in clinical practice, with a large preponderance in the older (>75 years) adult population. Stroke is the most feared complication of AF, with huge corresponding morbidity and mortality. Anticoagulation is the mainstay for stroke prevention in AF, but it is commonly underutilised in clinical practice due to the fear of intracerebral bleeding. Direct oral anticoagulants (DOACs) have now been used for a decade in clinical practice and have been found to reduce major bleeds. The advantages of DOACs over vitamin K antagonist (VKA) use includes no need for intermittent international normalised ratio (INR) monitoring, fewer drug interactions and a reduction in intracerebral haemorrhage (ICH) in general. The disadvantages include having to take multiple doses per day and a lack of a universal antidote, as opposed to VKA. In this review, we describe the need for anticoagulation in older patients with AF, review the data for VKA use, discuss the advantages of DOACs and compare the data for individual DOACs.

Burden of AF in Older Patients

Age is one of the most important pathogenic factors for AF. The incidence of AF rises with increasing age, especially after the age of 65 years. Different populations have different prevalence of AF; however, it has been well documented in all populations that the prevalence increases with increasing age. The prevalence is 0.5% among populations aged ≤50 years, while it is 10–17% among populations aged ≥80 years.¹ Approximately 70% of patients with AF are aged >75 years.² AF contributes to increased stroke risk and cardiovascular (CV) mortality, so screening and management of AF in older adults is extremely important.

Risk of Stroke with AF

AF is one of the well-known risk factors for stroke.³ The annual risk of stroke in patients with AF is around 1–20%.⁴ The risk of stroke is increased by a factor of 5 in patients with AF. For patients with AF, there are several risk stratification scores that aid in decision-making for initiating thromboprophylaxis. One of them is the CHA₂DS₂-VASc score, which is the most widely used scoring system because of its ease in identifying low-risk AF patients. The annual risk of stroke per 100 patient years increases with the number of points on the CHA₂DS₂-VASc score: 0.5% (0 points), 1.5% (1 point), 2.5% (2 points), 5% (3 points), 6% (4 points) and 7% (5–6 points; Figure 1).⁵ Contemporary guidelines indicate that thromboprophylaxis is needed if the total score is ≥2 in men and ≥3 in women.^6,7

Stroke Risk Reduction with Anticoagulation: The Evidence

Before the advent of DOACS, VKAs were the conventional oral anticoagulant (OAC) therapy for patients with AF at risk of stroke. The dose is adjusted to maintain an INR of 2–3. VKAs reduce the incidence of stroke by 66% and death by 25% compared with no anticoagulant therapy.^8–11 In contrast, antiplatelet therapy reduces the risk of stroke by only 22%.¹¹ In patients aged ≥75 years, treatment with warfarin leads to a 52% reduction in the incidence of fatal and disabling stroke, including arterial embolism and ICH.⁹ These benefits are also seen in individuals aged > 80 years when compared with aspirin after 1 year of follow-up.¹¹ Despite the strong evidence in favour of VKAs for stroke prevention, there has been a high rate of under-prescription due to well-known drawbacks, such as unpredictable pharmacodynamics and pharmacokinetics, narrow therapeutic window, several drug–drug and food–drug interactions, and the need for frequent laboratory monitoring and dose adjustments. However, the most important concern is the risk of bleeding, especially ICH, especially in individuals aged >80 years.

Evidence of Direct Oral Anticoagulants Benefits and Safety in Older Patients

Recent guidelines recommend DOACs (including the direct thrombin inhibitor, dabigatran, and the direct factor Xa inhibitors, rivaroxaban, apixaban and edoxaban) over VKAs in patients with non-valvular AF (AF in the absence of mechanical prosthetic heart valves and in the absence of moderate to severe mitral stenosis).^6,7,12 This is due to the multitude of benefits of DOACs over VKAs, such as an improved efficacy/safety ratio, predictable effect without any need for monitoring, and fewer food and drug interactions. Furthermore, Phase III DOAC trials have demonstrated an approximately 12% decrease in mortality compared with warfarin and a 20% decrease in stroke or systemic embolisation.¹³ The most impressive benefit of DOACs is a 50% reduction in ICH, although they are associated with an increased incidence of gastrointestinal bleeding.^13,14 Although there are no clinical trials that specifically assess the efficacy and safety of DOACs in older people, there were a considerable number of patients aged >75 years enrolled in the studies favouring their use. Moreover, the absolute risk reduction would be higher in older patients than in younger patients, as they are at an increased overall risk. Thus, there would be a favourable balance between the risks and benefits of treatment in older patients, as demonstrated in younger individuals (Supplementary Tables 1 and 2).^15–17

The bleeding profile of DOACs is more favourable than that of warfarin, particularly for life-threatening haemorrhagic events and ICH.^13,18 All the DOAC trials have shown non-inferiority or even superiority for the overall incidence of bleeding.^6,7 However, as described below, there is an increased risk of gastrointestinal (GI) bleeding, especially with dabigatran and rivaroxaban.

A meta-analysis of four pivotal trials of DOACs also compared treatment outcomes in patients of differing age groups.¹⁹ A subanalysis of patients aged ≥75 years showed that all four DOACs have a similar efficacy in preventing stroke and systemic embolism (SE) compared with warfarin. There is an overall highly significant reduction in RR favouring DOACs (HR 0.70; 95% CI [0.61–0.80]), with no significant heterogeneity between trials (p for heterogeneity=0.83). However, the results from the individual trials warrant further analysis. There were significantly fewer major bleeding events in the ARISTOTLE and ENGAGE AF-TIMI 48 trials in older patients with apixaban 5/2.5 mg twice daily (RR 0.67, 95% CI [0.55–0.82]) and edoxaban 60/30 mg once daily (RR 0.82, 95% CI [0.69–0.97]) compared with warfarin. In contrast, neither dabigatran (RR 1.10, 95% CI [0.94–1.29]) nor rivaroxaban (RR 1.14, 95% [0.97–1.37]) reduced bleeding in older patients compared with warfarin. However, each of the four trials showed significantly lower ICH rates with DOACs versus warfarin in patients aged ≥75 years (>50% reduction with dabigatran, apixaban and edoxaban, and 20% with rivaroxaban).^20–23 Although, in these trials, only 20% of patients were aged >80 years, and only 5% were aged >85 years, the results were consistent with patients aged >75 years.

A more recent patient-level meta-analysis of these four randomised controlled trials has explored the relationship between age as a continuous variable and outcomes of stroke/SE, major bleeding and mortality.²⁴ In this meta-analysis, patients randomised to lower doses of dabigatran 110 mg and edoxaban 30/15 mg were excluded. There was no significant treatment interaction across the age range of 50–100 years for the efficacy outcomes of stroke/SE or mortality. The HRs for DOACs versus warfarin were close to 1.0 in younger patients, and tended to decrease as the age increased. However, for the safety endpoint of major bleeding, the reverse was seen. The HR of DOACs versus warfarin increased by 10.2% (95% CI [1.3–19.9]) per 10-year increase in age. Thus, the benefit of reduction in bleeding with the DOACs versus warfarin was present only in younger patients (p for interaction=0.02). The relationship between age and major bleeding varies significantly between different DOACs; therefore, pooling results across trials may obscure the actual true relationships between age, bleeding and specific DOACs. For example, there was an increased risk of GI bleeding with dabigatran and rivaroxaban in individuals aged >75 years; however, it was not seen with apixaban and edoxaban.

In a large nationwide database from Norway of older AF patients aged >75 years (n=30,401) taking oral anticoagulation, DOAC use was associated with similar stroke/SE risk, while there was lower bleeding risk.²⁵ The use of both a standard and a reduced dose produced equivalent results. A few studies have evaluated anticoagulation in very elderly patients.

A prospective analysis of seven stroke cohorts compared the safety and effectiveness of oral anticoagulation in patients with recent stroke stratified by age.²⁶ Out of 5,984 patients enrolled in the study, 1,380 (23%) were aged ≥85 years and 3,688 (62%) received a DOAC. On follow-up, the HR for the primary composite outcome (recurrent stroke, ICH and all-cause death) was lower with DOACs compared with VKAs (HR 0.74; 95% CI [0.63–0.86]). Interestingly, the beneficial effects of DOACs did not differ between patients aged ≥85 and <85 years (HR 0.65 versus 0.79, respectively; p for interaction=0.129). Separate analyses on recurrent stroke, ICH and death were performed, and the results were consistent with the primary analysis showing that the benefits of DOACs were irrespective of age. A similar net clinical benefit with DOACs was seen in patients aged ≥85 years (+1.73 to +2.66) and <85 years (+1.90 to +3.36 events/100 patient-years).

In a nationwide Taiwanese insurance database, oral anticoagulation with warfarin was effective in older patients aged >90 years (n=15,786) in reducing the risk of stroke (HR 0.69; 95% CI [0.49–0.96]) compared with no therapy.²⁷ The net clinical benefit (which takes into account the rate of ICH) was also in favour of warfarin. On the contrary, there was no benefit of antiplatelet therapy in this cohort, vindicating the Class III recommendations by contemporary guidelines. The use of DOACs led to further reduction in ICH (HR 0.32; 95% CI [0.10–0.97]) without any difference in stroke rates.

There are few studies suggesting variation among the effects of DOACs. A subanalysis of the ARISTOPHANES study compared DOACs with warfarin among very frail older patients with AF for stroke/SE and major bleeding.²⁸ There were 150,847 patients, and the patients were grouped into six cohorts (apixaban-warfarin, apixaban-rivaroxaban, apixaban-dabigatran, dabigatran-warfarin, dabigatran-rivaroxaban and rivaroxaban-warfarin). In this analysis, apixaban reduced major bleeding (both ICH and extracranial) more than warfarin, dabigatran and rivaroxaban (Figure 2). Although there were fewer ICHs with dabigatran and rivaroxaban compared with warfarin, the rates of GI bleeding were high. A network meta-analysis analysed the safety of DOACs with respect to major GI bleeding among older adults.²⁹ There was a 44% reduction in major GI bleeding with apixaban compared with rivaroxaban. Apixaban and edoxaban had the most favourable major GI bleeding safety profiles, while dabigatran and rivaroxaban had the least safety in this regard.

In a meta-analysis of six randomised controlled trials and six observational studies, Grymonprez et al. evaluated the effect of DOAC therapy in older AF patients aged >75 years.³⁰ DOACs were superior to VKA in reducing stroke/SE and all-cause mortality (HR 0.83 and HR 0.77, respectively). There was no difference in major bleeding compared with VKAs (HR 0.93; 95% CI [0.86–1.01]). DOACs were found to decrease ICH; however, they also increased the rates of GI bleeding. Apixaban again had the lowest risk of GI bleeding among DOACs compared with warfarin. Other meta-analyses of randomised controlled trials and non-randomised studies of older AF patients found a similar uniform risk–benefit ratio favouring DOACs compared with VKAs with respect to ischaemic stroke in older AF patients.^31–36 The effect on bleeding differed between studies, with a few suggesting preferences of apixaban and edoxaban for bleeding in the older adult population.^30,31,35,36 Similar findings are suggestive of a uniform class effect of DOACs for stroke/SE in AF/venous thromboembolism and differential effects for major bleeding have also been revealed in other studies in broad populations.^37–41 Apixaban use resulted in lower bleeding and specifically lower GI bleeding compared with other DOACs. The safety profile shown by apixaban also translated into lower mortality in some studies.³⁷

At this juncture, it is worthwhile to mention the ELDERCARE-AF trial. It was an interesting Phase III trial of anticoagulation in very elderly patients (>80 years) with AF with a CHADS₂ score >2.⁴² Patients were deemed unsuitable for the use of DOACs at standard doses due to renal dysfunction, recent bleeding, low bodyweight and concomitant non-steroidal anti-inflammatory agents (NSAIDs) or antiplatelet use. Patients were randomised to edoxaban 15 mg or a matching placebo. There were 984 patients with mean age of 86.6 + 4.2 years, with a mean estimated glomerular filtration rate (eGFR) of 36.3 ± 14.4 ml/min/1.73m². The primary efficacy endpoint (composite of stroke or SE) was significantly reduced by 66% at 15 months (2.3% per patient year versus 6.7% per patient year; HR 0.36; p<0.001). Deaths were also numerically lower in the edoxaban arm (9.9%/patient-year versus 10.2%/patient year; HR 0.97).

The edoxaban group also achieved lower rates of the secondary efficacy endpoint (annual rate of composite of stroke, SE and CV death; HR 0.72, 95% CI [0.50–1.03]), as well as the major adverse CV events rate (HR 0.70; 95% CI [0.49–1.01]), although both values did not attain statistical significance. These benefits were achieved at the cost of increased major bleeding by almost twofold (HR 1.87; p=0.09). With edoxaban, there were more GI bleeding events, but ICH rates were lower. A noteworthy point is that edoxaban did not increase all-cause death or CV death, despite a non-significant increase in bleeding. Another notable caveat of the study remains that it does not affirm superiority of edoxaban 15 mg over regular doses of edoxaban (60/30 or 30/15 mg) or VKA therapy.

Problems with Anticoagulants in Older Patients

Ageing is a complex process, and it is accompanied by numerous health problems, which are often coupled with reduced mobility, greater frailty and a greater tendency to fall.^43,44 Because of all these conditions, older patients with AF are often precluded from being anticoagulated.⁴⁵ Nonetheless, older patients with AF may have a good quality of life and be able to perform all activities of daily living, where excluding them from thromboprophylaxis may result in devastating stroke and severe disability. This is especially important, as the strokes associated with AF are more severe and result in greater disability.⁴⁶

Advanced age in itself should not be a contraindication to anticoagulation therapy, although it is increasingly being recognised as a vital factor for bleeding, especially in those aged >85 years.⁴⁷ In a real-world setting, 38–40% of AF patients at high risk of stroke were prescribed aspirin instead of anticoagulants, presuming it to be a safer option.⁴⁸ Pre-existing vascular disease, such as coronary artery disease, prior MI, unstable and stable angina, recent coronary artery bypass graft, peripheral arterial disease, and the coexistence of CV risk factors, such as hypertension and dyslipidaemia, were predictors of prescription of aspirin. Similarly, in the GARFIELD-AF registry, one-fifth of the patients were prescribed aspirin monotherapy. Older patients (aged >75 years) are more likely to be prescribed aspirin monotherapy (OR 1.24) compared with younger populations (aged <55 years, HR 1.01).⁴⁹

However, in the BAFTA trial, the efficacy of warfarin (INR 2–3) over aspirin 75 mg in an older adult population (aged >75 years) was clearly superior (1.8 versus 3.8% per year; RR 0.48; 95% CI [0.28–0.80]). Moreover, there was no difference in major bleeding rates between warfarin and aspirin (1.9 versus 2.0% risk per year).¹⁰

The AVERROES study attempted to assess the superiority of apixaban (5 mg twice daily) over aspirin (81–324 mg/day) for stroke prevention in 5,600 AF patients who were deemed unsuitable for VKA treatment.⁵⁰ The median duration of follow-up was 1.5 years, the mean age was approximately 70 years and the primary endpoint was a composite of stroke or SE. However, the study had to be prematurely terminated due to clear benefits of apixaban over aspirin. There was a 45% significant reduction in primary outcome events in those randomised to apixaban (1.6%/year versus 3.7%/year; 95% CI [0.32–0.62], p<0.001). Surprisingly, there was also no difference in major bleeds (1.4%/year versus 1.2%/year; HR 1.13; 95% CI [0.74–1.75]; p=0.57).

In a separate analysis of the older adult subpopulation from the study, apixaban use led to a greater reduction in primary efficacy endpoint compared with younger patients.⁵¹ The relative risk reductions with apixaban in those aged >75 years was 67% and those aged >85 years was 86%, respectively. Although the overall bleeding incidence was enhanced in the older adult subpopulation, the individual rates in both arms were comparable.

Thus, older patients with AF, if at increased risk of stroke, should be anticoagulated whenever possible, unless there are compelling contraindications. The special considerations for older patients include the following:

Haemorrhagic Risk

The most important complication of anticoagulation is bleeding. The risk of bleeding depends on the intensity of anticoagulation and intrinsic characteristics of the patient. The haemorrhagic risk can be calculated using various risk scores, such as HEMORR2HAGES, ATRIA-2, HAS-BLED, GARDFIELD-AF and ORBIT (Table 1).^52–56 The HAS-BLED score was developed from a real-world cohort of >3,000 AF patients on oral anticoagulation, and its use is encouraged in current guidelines.^6,12 A score >3 indicates high bleeding risk, while scores of 1–2 indicate moderate risk. The HAS-BLED score has outperformed other scores, such as ATRIA and HEMORR₂HAGES, in two systematic reviews.^57,58 Meanwhile, in other studies, it was at least non-inferior to other scores, justifying its widespread use based on guidelines.^59–61

Although initially developed in patients taking warfarin, recent studies have shown equivalent c-statistics in patients receiving DOAC therapy.⁵⁹ Similarly, there was no difference in diagnostic accuracy in older patients.⁵⁹ There was no loss in predictive accuracy, even when labile INR values were not available. However, those who have a high bleeding risk score are also automatically at high risk of ischaemic stroke and SE. Therefore, such patients should not be denied anticoagulation; rather, they should be monitored and followed more frequently. Careful attention should be directed towards modifiable risk factors for bleeding, such as uncontrolled hypertension, concomitant antiplatelet use, anaemia, excess alcohol consumption, impaired renal function, hepatic dysfunction, platelet count and NSAID use.⁶ A repeat HAS-BLED after control of such factors may enable anticoagulation in a fraction of such patients. In a nationwide cohort of AF patients, the HAS-BLED score at follow-up or ‘Delta’ HAS-BLED score was more predictive of bleeding than the baseline HAS-BLED score or the bleeding risk factors indicating the dynamic process of bleeding.⁶²

Adherence

There are several factors that lead to non-adherence in older patients, which causes therapeutic failure.⁶³ Since the half-life of DOACs is shorter, the impact of missing a dose would result in inadequate anticoagulation quickly. However, with VKAs, dose omission could lead to either under- or overanticoagulation if the INR is not properly monitored.

Falls

Falls are quite common among older patients, and thus, physicians tend to avoid using anticoagulation therapy for older patients.⁶⁴ However, it should be kept in mind that despite the risk of traumatic ICH from falls, there is a benefit of 5% annual risk reduction in ischaemic stroke, which is quite debilitating. Therefore, there is a lack of evidence of substantial harm of anticoagulation in AF patients who experience falls.⁶⁵

Chronic Kidney Disease

Chronic kidney disease (CKD) is more common in older patients. It is a risk factor for both thrombosis and bleeding. The pivotal trials of dabigatran, rivaroxaban and edoxaban excluded patients with eGFR <30 ml/min/1.73m², whereas in the pivotal trials of apixaban, the eGFR cut-off was 25 ml/min/1.73m².^6,66 DOACs are not approved for use in end-stage CKD (eGFR <15 ml/min/1.73m²) in the US, and in Europe, the dose needs to be reduced with eGFR <30 ml/min/1.73m² for all agents, except apixaban, which has separate criteria for dosing rather than relying on eGFR.⁶⁷

Among DOACs, apixaban has the least renal excretion and is approved by the Food and Drug Administration for use in patients with end-stage renal disease.⁶⁵ In a systematic review of 11 studies, it was found to be a reasonable alternative to warfarin in stage 4/5 CKD.⁶⁸ However, VKAs can be used in these patients. According to a meta-analysis of 11 cohort studies in AF patients with severe CKD, warfarin was associated with a lower risk of stroke, thromboembolism and mortality, with no appreciable increase in major bleeding.⁶⁹ Hence, for patients at high risk of bleeding, apixaban can be used, while those at low risk of bleeding can be placed on warfarin or other DOACs at modified doses according to their eGFR.

Polypharmacy and Drug–Drug Interactions

Older patients require multiple medicines due to age-associated comorbidities. This can lead to drug–drug interactions, enhanced adverse effects and worse clinical outcomes. This is especially relevant for antiplatelets and NSAIDs. Aspirin increases the risk of bleeding by twofold in patients receiving VKAs, while it is only increased by 1.3–1.6-fold in those who are taking DOACs.^70–73 Concomitant use of NSAIDs is also similarly associated with increased bleeding risk and, therefore, should be avoided. Meanwhile, drugs that inhibit or induce the cytochrome P450 enzyme can alter VKAs pharmacokinetics.⁴² Drugs that inhibit P-glycoprotein substrates can result in higher levels of dabigatran and edoxaban. Rivaroxaban and apixaban have a dual clearance mechanism (P-glycoprotein in the kidneys and cytochrome p450-3A4 in the liver), thus drugs that induce these two enzymes will result in low levels of rivaroxaban and apixaban.

Giugliano et al. found that there was no effect of comorbidities and polypharmacy on the efficacy of DOACs compared with VKAs.¹⁹ However, there was an increase in major bleeding events as the number of co-medications increased to more than four with rivaroxaban (ROCKET-AF Trial) and more than nine with apixaban (ARISTOTLE trial). However, for edoxaban (ENGAGE AF-TIMI 48 trial), there was no increase in major bleeding, even with multiple co-medications. This is because edoxaban is not metabolised by the P450 mechanism, and the dose of edoxaban was reduced by 50% in patients taking strong P-glycoprotein inhibitors. Additionally, dynamic dose adjustment was allowed post-randomisation.¹⁹

Food–Drug Interactions

Food rich in vitamin K (green leafy vegetables) can reduce the anticoagulant effects of VKAs. Among the DOACs, rivaroxaban should be taken with food for optimal absorption.⁷

Frailty and Dementia

The prevalence of frailty increases with advancing age, and it is associated negatively with health outcomes.^74,75 Unfortunately, the incidence of frailty is higher in older non-valvular AF patients compared with matched controls.⁷⁶ This can be attributed not only to common risk factors between AF and frailty, but also bidirectional pathological mechanisms in which each condition promotes the other.⁷⁷ Frailty in AF is also associated with poor cognitive function.⁷⁶ Due to impaired cognitive function and frailty, older patients are less likely to receive anticoagulation.^43–45,63 The advantage of DOACs is that they have a simplified regimen and require no monitoring, but they have a short half-life, so missing even a single dose could result in suboptimal anticoagulation. In contrast, monitoring of INR for VKAs results in increased dependency on caregivers, which is not possible for many patients.

Frailty has been an emerging concept in the past two decades, and loosely refers to diminution of physiological reserves of the body, which preclude smooth functioning of routine activities.⁷⁸ In simple terms, it can be considered the premature acceleration of the process of senescence. Various indices have been proposed for the assessment of frailty. Fried’s frailty index was a simple score developed two decades ago. The presence of at least three variables out of these features – unintentional weight loss, self-reported exhaustion, weakness/poor grip strength, slow walking speed and low physical activity – indicates frailty.⁷⁹

Although a myriad of other scores are in vogue, only a handful of them were used in the studies on oral anticoagulation in older patients and are worth mentioning. The Hospital Frailty Risk Score was developed from >22,000 electronic hospital records of older patients using 109 ICD-10 codes, and a score >5 indicates frailty.⁸⁰ It is an easy-to-use, non-invasive and cost-effective method for frailty assessment. Modifying the Fried’s frailty index, Segel et al. developed a claims-based frailty indicator from analysis of 10 years of the Mediclaim database involving >4,400 patients and 21 variables.⁸¹ For Asian patients, a score involving five physical domains – weight loss, grip strength, walking speed, exhaustion and activity level – was developed by Makizako et al., and frailty is diagnosed with a score >3.⁴⁴ The Groningen Frailty Indicator uses multiple variables from physical, cognitive, social and psychological domains, and a score >3 indicates frailty (Table 2).⁸²

Data supporting the use of OAC in older frail patients are limited. In the ELDERCARE AF trial, a prespecified analysis of frail patients was performed.⁸³ All patients were categorised at randomisation as robust, prefrail or frail using the frailty score developed by Makizako et al., as described above. A total of 402 patients (41%) were considered frail, and among them, edoxaban 15 mg/day reduced stroke/SE by 65% (HR 0.35; 95% CI [0.14–0.87]) without any increase in major bleeds (HR 1.67; 95% CI [0.58–4.75]).⁵⁷

An analysis of frail patients (34%) in the ARISTOPHANES trial displayed similar beneficial effects of DOACs.²⁸ This study used the claims-based frailty score. Apixaban and rivaroxaban displayed lower rates of stroke, while apixaban and dabigatran showed lower rates of major bleeding compared with warfarin.

A retrospective analysis of a large Korean database of older and frail AF patients (n=83,635) revealed beneficial effects of OACs.⁸⁴ Frailty was defined based on the Hospital Frailty Risk Score. At a mean age of 78.5 years, there was a 22% reduction in net adverse clinical events, 9% reduction in stroke and 48% reduction in CV death with OAC use. Compared with warfarin, all DOACs had better outcomes with respect to major bleeding, stroke, CV death and net adverse clinical events.

Guideline Recommendation

Among older patients, DOACs provide large reductions in stroke and SE without increasing bleeding when compared with aspirin or no therapy. Furthermore, compared with warfarin, DOACs better reduce stroke/SE and death with less intracranial bleeding in older patients. According to guidelines, DOACs should be the preferred antithrombotic therapy in older patients.^6,7,12 The dose of all DOACs should be adjusted to eGFR, except for apixaban, which uses separate criteria (the ‘ABC’ rule). The 2023 European Heart Rhythm Association consensus document also states that the advantage of DOACs relative to VKAs are likely consistent in both frail and non-frail AF patients.

Choosing an Anticoagulant for an Older Patient

Patients who are eligible for standard-dose OACs should be considered for a DOAC instead of a VKA. Among the DOACs, apixaban 5/2.5 mg twice daily or edoxaban 60/30 mg once daily is preferable, as these two DOACs significantly reduced major bleeding, including GI bleeding, in older patients in the ARISTOTLE and ENGAGE AF-TIMI 48 trials, respectively. However, when the risk of bleeding is not high, dabigatran 110 mg twice daily or rivaroxaban 20/15 mg once daily may also be used. The aforementioned approach can also be applied to frail patients, although the data are less robust.

For patients treated with multiple drugs, attention to drug–drug interactions should be paid. In particular, co-administration of strong P-glycoprotein-interfering drugs (which interact with all DOACs) and strong cytochrome P-450 inhibitors (which affect the metabolism of rivaroxaban and apixaban) is very important. Edoxaban may be the preferred DOAC for patients taking strong P-glycoprotein inhibiting cardiac medications or strong cytochrome P-450 inhibitors, since it has the least interaction with these drugs. Figure 3 shows the advantages and disadvantages of choosing DOACs over VKAs in clinical practice.

Very Low-dose DOACs in Older Patients: A Way Out for Bleeding Risk?

It is evident from the DOAC and antiplatelet trials that bleeding is generally correlated to the dosage of drug. The ELDERCARE AF study discussed earlier successfully used a very low dose of edoxaban 15 mg once daily in older patients at high bleeding risk, and found superiority over placebo.^42,83 Can we extrapolate this data to other DOACs? How does the very low dose or ‘off-label’ dose of DOACs fare with a standard dose DOACs?

Lip et al. performed a real-world study of a Taiwanese database of >11,000 older AF patients undergoing anticoagulation with lower and standard DOAC doses.⁸⁵ Over a period of 7 years, they enrolled approximately 7,300 patients taking a very low-dose DOAC (edoxaban 15 mg, dabigatran 110/150, apixaban 2.5, rivaroxaban <10 mg – all given once daily) and 4,100 patients taking standard doses of DOAC. At 1-year follow-up, the risk of ischaemic stroke/SE and major bleeding were comparable between the DOAC groups. However, venous thromboembolism events were increased by 3.8-fold with very low-dose DOACs. Simultaneously, death and lower limb ischaemic events were increased by 21% and 57%, respectively, in the very low-dose DOAC arm. Similarly, concerns over efficacy of off-label use of 15 mg dose of rivaroxaban in Asian patients were shown in a Korean study of >2,200 older AF patients.⁸⁶ In the study, the net clinical benefit of rivaroxaban over warfarin was demonstrated with the 20 mg dose, but not the 15 mg dose. Therefore, the current strategy of using ultra-low-dose DOACs in older and frail patients to curtail bleeding cannot be advocated.

Switching from Vitamin K Antagonists to Direct Oral Anticoagulants in Older Adults: Is It the Right Time?

It is evident from the previous sections that older adults carry a high risk of bleeding with OAC use, and DOACs produce less bleeding than VKA in this population subset. Therefore, a pertinent question arises in patients who are already receiving VKA therapy, is it prudent to switch to a DOAC-based regimen? To answer this question, the FRAIL-AF study was undertaken. The hypothesis was whether switching from VKA to DOAC reduces bleeding events in frail older patients. The study patients were stable patients with AF receiving VKA therapy, aged ≥75 years, with Groningen frailty indicator scores ≥3 and eGFR ≥30 ml/min/1.73m². The patients were randomised to an intervention arm (switching from VKA to DOAC) or a control arm (continuing VKA) in 1:1 ratio.⁸⁷ The primary outcome was the occurrence of any major or clinically relevant non-major bleeding, while the secondary outcome was thrombo-embolic events and all-cause mortality. It was an investigator-initiated randomised, pragmatic, multicentre, open-label, superiority trial with 1-year follow-up. At study completion, both major and non-major clinically relevant bleeding were significantly higher in the intervention arm (HR 1.69; 95% CI [1.23–2.32]; p=0.00112) with no difference in thromboembolic events and all-cause mortality.

Whether the study represents a failure of DOACs or a specific DOAC agent is a matter of debate. As previously detailed, DOACs are heterogenous with respect to bleeding. Rivaroxaban was the predominant DOAC used in this study, and it is now known that among DOACs it has the highest tendency for bleeding. In a meta-analysis of AF patients with valvular heart disease, all DOACs prevented ischaemic stroke and SE compared with VKA. Interestingly, all DOACs also reduced major bleeding except rivaroxaban, which enhanced the risk.⁸⁸

The ARISTOPHANES study also revealed that rivaroxaban had higher rates of major bleeding compared with VKA, while apixaban and dabigatran had lower bleeding.⁸⁹ It was a large retrospective observational study of a Medicare claims database from the US consisting of >430,300 patients. DOAC–DOAC and DOAC–warfarin propensity matching was performed, and patients were divided into six cohorts. Compared with VKA, the rates of major bleeding were reduced by 40 and 29% with apixaban and dabigatran, respectively, while with rivaroxaban, it was higher (HR 1.06). Although all three DOACs reduced ICH compared with VKA, GI bleeding was increased with rivaroxaban, while apixaban reduced GI bleeding. Interestingly, apixaban reduced major bleeding compared with both dabigatran and rivaroxaban, while dabigatran reduced major bleeding compared with rivaroxaban. Therefore, rivaroxaban may not be the ideal DOAC to switch to in older adults. Nevertheless, the trial results emphasise that switching from a VKA to a DOAC should not be considered without a clear indication in frail older patients with AF.

Factor XI Inhibition: A Novel Approach to Reduce Bleeding

The idea to develop new agents with even better safety profiles than DOACs, and possibly preventing thrombosis without increasing the risk of bleeding, comes from knowledge of rare genetic disorders, specifically deficiency of the contact pathway coagulation protein, factor XI. This appears to be associated with reduced risk of thrombosis and only minor bleeding tendency. Therefore, factor XI inhibition appears to be a promising avenue for anticoagulation. Several molecules are in development, including monoclonal antibodies (abelacimab), antisense oligonucleotides (fesomersen) and small molecule inhibitors (milvexian).⁹⁰ The Phase II trials of anti-XI agents demonstrated the potential to be more effective than enoxaparin/DOACs with attenuated bleeding across variegated scenarios, such as AF, venous thromboembolism, post-stroke and post-MI.⁹¹ Larger Phase III trials of factor XI inhibitors are underway and will decide the fate of these molecules.

Conclusion

Patients with AF are at increased risk of stroke and SE, and this risk increases with advancing age. Older patients are often denied anticoagulation due to frailty and perceived risk of bleeding complications. However, evidence suggests that the use of anticoagulants, even in frail older patients, outweighs the risks. Therefore, in the absence of contraindications, all patients with AF who need anticoagulation should receive OACs. DOACs should be the first choice, as they are more effective and cause less bleeding, especially ICH, compared with VKAs. Although there are no head-to-head comparisons among the DOACs, multiple small studies and registries favour the use of apixaban and edoxaban over dabigatran and rivaroxaban, as they reduce all major bleeding complications, including GI bleeds, which are the most feared complication in older adults. However, if frail older patients are already receiving VKA therapy and are doing well, switching to DOACs without any clear indication should be discouraged. Finally, as older patients frequently take other medications, careful attention should be paid to drug interactions, and patients should be followed regularly.

Click here to view Supplementary Material.