Advertisement
Canadian Journal of Cardiology

The Introduction of Direct Oral Anticoagulants Has Not Resolved Treatment Gaps for Frail Patients With Nonvalvular Atrial Fibrillation

Open AccessPublished:December 20, 2021DOI:https://doi.org/10.1016/j.cjca.2021.09.021

      Abstract

      Background

      The extent to which the introduction of direct oral anticoagulants (DOACs) influenced treatment patterns in frail and nonfrail patients with nonvalvular atrial fibrillation (NVAF) is unclear.

      Methods

      This was a retrospective cohort study of all Albertans 20 years or older who were discharged from an emergency department or hospital with a new diagnosis of NVAF between April 1, 2009, and March 31, 2019. The Hospital Frailty Risk Score was used to define frailty and the CHA2DS2-VASc and CHADS-65 scores were used to identify if anticoagulation was indicated.

      Results

      Among 75,796 patients (median age, 75 years; 45% female) with a new diagnosis of NVAF, 17,143 (22.6%) were frail. Although guideline criteria for anticoagulation were more commonly met by frail patients than nonfrail patients (92.1% vs 74.2%, for CHA2DS2-VASc, and 96.8% vs 85.8% for CHADS-65; both P < 0.0001), frail patients were less likely to receive any anticoagulant, even after those with contraindications to anticoagulation were excluded (adjusted odds ratio, 0.61; 95% confidence interval, 0.58-0.64). After DOACs became available, anticoagulant prescribing for patients with guideline indications increased more in nonfrail patients (from 42.4% to 68.2%) than in frail patients (from 29.0% to 52.2%) and frail patients were less likely to receive a DOAC than warfarin (adjusted odds ratio, 0.66; 95% confidence interval, 0.54-0.81).

      Conclusions

      Although they stand to potentially derive greater benefits from anticoagulation, frail patients were less likely to receive an anticoagulant and, if anticoagulated, they were more likely to receive warfarin than a DOAC. The introduction of DOACs has increased anticoagulation rates but not resolved treatment gaps for frail patients with NVAF.

      Résumé

      Contexte

      On ne sait pas exactement dans quelle mesure l’introduction des anticoagulants oraux directs (AOD) a influé sur les modalités de traitement des patients fragiles ou non fragiles atteints de fibrillation auriculaire non valvulaire (FANV).

      Méthodologie

      Nous présentons ici une étude de cohorte rétrospective portant sur tous les Albertains âgés de 20 ans ou plus ayant reçu leur congé des urgences ou d’un hôpital entre le 1er avril 2009 et le 31 mars 2019 après avoir reçu un nouveau diagnostic de FANV. Le score HFRS (Hospital Frailty Risk Score) a été utilisé pour définir la fragilité et les scores CHA2DS2-VASc et CHADS-65 ont servi à déterminer si l’anticoagulation était indiquée.

      Résultats

      Parmi 75 796 patients (âge médian de 75 ans; 45 % de femmes) chez qui un nouveau diagnostic de FANV avait été posé, 17 143 (22,6 %) étaient fragiles. Bien que les critères de recommandation des anticoagulants étaient plus fréquemment satisfaits chez les patients fragiles que chez les patients non fragiles (92,1 % vs 74,2 %, selon le score CHA2DS2-VASc, et 96,8 % vs 85,8 % selon le score CHADS-65; p < 0,0001 dans les deux cas), les patients fragiles étaient moins susceptibles de recevoir un anticoagulant, même après l’exclusion de ceux chez qui l’anticoagulation était contre-indiquée (rapport de cotes corrigé : 0,61; Intervalle de confiance [IC] à 95 % : 0,58-0,64). Après l’introduction des AOD, l’augmentation de la prescription d’anticoagulants aux patients répondant aux critères de recommandation était plus importante chez les patients non fragiles (passant de 42,4 % à 68,2 %) que chez les patients fragiles (passant de 29,0 % à 52,2 %), et ces derniers étaient moins susceptibles de recevoir des AOD que de la warfarine (rapport de cotes corrigé : 0,66; IC à 95 % : 0,54-0,81).

      Conclusions

      Bien qu’ils puissent potentiellement tirer de plus grands bienfaits des anticoagulants, les patients fragiles étaient moins susceptibles de recevoir un anticoagulant et, le cas échéant, ils étaient plus susceptibles de recevoir de la warfarine que des AOD. L’introduction des AOD a augmenté les taux d’anticoagulation, sans pour autant combler certaines lacunes dans le traitement des patients fragiles atteints de FANV.
      Atrial fibrillation (AF) is the most common cardiac arrhythmia
      • Go A.S.
      • Hylek E.M.
      • Phillips K.A.
      • et al.
      Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study.
      and patients with nonvalvular AF (NVAF) have a twofold increased risk of mortality and a fivefold increased risk of stroke, although their absolute risk depends on additional risk factors.
      • Hughes M.
      • Lip G.Y.
      Stroke and thromboembolism in atrial fibrillation: a systematic review of stroke risk factors, risk stratification schema and cost effectiveness data.
      Although these risks are mitigated by anticoagulation,
      • Andrade J.G.
      • Aguilar M.
      • Atzema C.
      • et al.
      The 2020 Canadian Cardiovascular Society/Canadian Heart Rhythm Society comprehensive guidelines for the management of atrial fibrillation.
      NVAF patients are often undertreated.
      • Ogilvie I.M.
      • Newton N.
      • Welner S.A.
      • Cowell W.
      • Lip G.Y.H.
      Underuse of oral anticoagulants in atrial fibrillation: a systematic review.
      Frailty is a multidimensional syndrome of poor physiological reserve leading to increased vulnerability to stressors, resulting in dependency and poor health outcomes,
      • Morley J.E.
      • Vellas B.
      • van Kan G.A.
      • et al.
      Frailty consensus: a call to action.
      ,
      • Villacampa-Fernandez P.
      • Navarro-Pardo E.
      • Tarın J.J.
      • Cano A.
      Frailty and multimorbidity: two related yet different concepts.
      and it is increasingly recognized that frailty is more relevant for personalizing care than chronological age.
      • Clegg A.
      • Young J.
      • Iliffe S.
      • Rikkert M.O.
      • Rockwood K.
      Frailty in elderly people.
      A recent systematic review showed that the prevalence of frailty in NVAF patients ranged from 4.4% to 75.4%, depending on the criteria used to assess frailty and the age of patients enrolled.
      • Villani E.R.
      • Tummolo A.M.
      • Palmer K.
      • et al.
      Frailty and atrial fibrillation: a systematic review.
      Another recent meta-analysis
      • Wilkinson C.
      • Todd O.
      • Clegg A.
      • Gale C.P.
      • Hall M.
      Management of atrial fibrillation for older people with frailty: a systematic review and meta-analysis.
      highlighted a paucity of randomized trial evidence on the safety and efficacy of anticoagulation in frail patients. However, it is known that the risks of stroke, death, and major bleeding are all increased with frailty, and recent post hoc analyses of randomized trials have suggested that the presence of frailty or multimorbidity does not alter the relative efficacy or safety of warfarin or direct oral anticoagulants (DOACs).
      • Alexander K.P.
      • Brouwer M.A.
      • Mulder H.
      • et al.
      Outcomes of apixaban versus warfarin in patients with atrial fibrillation and multi-morbidity: insights from the ARISTOTLE trial.
      ,
      • Wilkinson C.
      • Wu J.
      • Searle S.D.
      • et al.
      Clinical outcomes in patients with atrial fibrillation and frailty: insights from the ENGAGE AF-TIMI 48 trial.
      Recent observational studies have also shown that, in frail NVAF patients, event rates were similar for users of warfarin, dabigatran, or rivaroxaban, and were possibly lower in apixaban users (although the possibility of residual confounding with any observational study precludes definitive statements about relative efficacy).
      • Zhang J.
      • Wang X.
      • Liu X.
      • et al.
      Comparative effectiveness and safety of direct acting oral anticoagulants in nonvalvular atrial fibrillation for stroke prevention: a systematic review and meta-analysis.
      ,
      • Kim D.H.
      • Pawar A.
      • Gagne J.J.
      • et al.
      Frailty and clinical outcomes of direct oral anticoagulants versus warfarin in older adults with atrial fibrillation: a cohort study.
      Although a recent study from 384 English general practices reported that frail patients with AF were more than twice as likely to be prescribed anticoagulants as nonfrail patients,
      • Wilkinson C.
      • Clegg A.
      • Todd O.
      • et al.
      Atrial fibrillation and oral anticoagulation in older people with frailty: a nationwide primary care electronic health records cohort study.
      an analysis from the United States Veterans Administration

      Sanghai SR, Liu W, Wang W, et al. Prevalence of frailty and associations with oral anticoagulant prescribing in atrial fibrillation [e-pub ahead of print]. J Gen Intern Med doi:10.1007/s11606-021-06834-1, accessed December 2, 2021.

      showed that frail veterans were 34% less likely to be anticoagulated than nonfrail patients but were 75% more likely to be prescribed a DOAC than warfarin. However, the Systematic Assessment of Geriatric Elements in Atrial Fibrillation investigators
      • Mailhot T.
      • McManus D.D.
      • Waring M.E.
      • et al.
      Frailty, cognitive impairment, and anticoagulation among older adults with nonvalvular atrial fibrillation.
      reported that frailty was not associated with anticoagulation rates.
      Because it is unclear whether the introduction of DOACs decreased treatment gaps in Canadians with NVAF (particularly those who are frail), we designed this study to examine anticoagulation treatment patterns in frail vs nonfrail NVAF patients and the effect of the introduction of DOACs.

      Methods

      We conducted a retrospective population-based study using the following linked administrative databases in Alberta, Canada: (1) the Discharge Abstract Database, which records the most responsible and up to 24 other diagnoses for all acute care hospitalizations anywhere in the province; (2) the National Ambulatory Care Reporting System, which captures all visits to any emergency department (ED) and includes up to 10 diagnoses per visit; (3) the Registry File, which includes demographic and geographic information for all residents; (4) the Pharmacy Information Network, which captures drug dispensations from outpatient pharmacies in the province; (5) the Provincial Laboratory Database that includes international normalized ratio (INR) results and dates of tests; and (6) the Alberta Vital Statistics Database, which records all deaths in the province. This study was approved by the University of Alberta Research Ethics Board (Pro00010852) with a waiver for individual subject consent because we used deidentified data.
      We identified all Albertans 20 years old or older who were discharged from an ED or hospital with a new (using a 5-year lookback period to exclude prevalent cases) diagnosis of NVAF between April 1, 2009, and March 31, 2019, using international classification of diseases (ICD) 10th revision code I48 in any diagnosis field, excluding patients with valvular disease (ICD codes I05, I06, I08.0, I08.1, I085.2, I08.3, I34, and I3 for mitral or aortic valve disease; ICD codes I07, I08.1, I08.2, I08.8, I08.9, I36, and I37 for tricuspid or pulmonary valvular disease; Canadian Classification of Health Interventions procedure codes 1.HS.80, 1.HS.90, 1.HT.80, 1.HT.89, 1.HT.90, 1.HU.80, 1.HU.90, 1.HV.80, and 1.HV.90 for valve surgery and procedures). This NVAF case definition has been previously validated.
      • Hawkins N.M.
      • Daniele P.R.
      • Humphries K.
      • et al.
      Empirical insights when defining the population burden of atrial fibrillation and oral anticoagulation uptake using administrative data.
      We used the Hospital Frailty Risk Score (HFRS) to define frailty.
      • Gilbert T.
      • Neuburger J.
      • Kraindler J.
      • et al.
      Development and validation of a Hospital Frailty Risk Score focusing on older people in acute care settings using electronic hospital records: an observational study.
      The HFRS determined according to the presence/absence of 109 different ICD 10th revision codes during each patient’s index event and any other hospitalizations in the previous 2 years and includes diagnoses such as decubitus ulcers, falls, osteoporosis, compression fractures, dementia, urinary incontinence, and electrolyte disorders. Any patients with a HFRS score less than 5 are defined as not frail, a score between 5 and 15 is deemed intermediate risk of frailty, and any patients with a score greater than 15 are classified as high risk of frailty. Although use of administrative data to identify frailty leads to lower prevalence estimates than gold standard face-to-face measures like comprehensive geriatric assessment or the Clinical Frailty Scale,
      • McAlister F.A.
      • Lin M.
      • Bakal J.A.
      Prevalence and postdischarge outcomes associated with frailty in medical inpatients: impact of different frailty definitions.
      ,
      • Kim D.H.
      • Patorno E.
      • Pawar A.
      • Lee H.
      • Schneeweiss S.
      • Glynn R.J.
      Measuring frailty in administrative claims data: comparative performance of four claims-based frailty measures in the U.S. Medicare data.
      the HFRS has been previously validated using Canadian and other administrative data.
      • McAlister F.A.
      • van Walraven C.
      External validation of the Hospital Frailty Risk Score and comparison to the Hospital-patient One-year Mortality Risk Score to predict outcomes in elderly hospitalized patients: a retrospective cohort study.
      ,
      • Hollinghurst J.
      • Housley G.
      • Watkins A.
      • Clegg A.
      • Gilbert T.
      • Conroy S.P.
      A comparison of two national frailty scoring systems.
      We also calculated the Elixhauser comorbidity score.
      All of the anticoagulant prescription analyses were evaluated in light of NVAF stroke risk scores advocated in Canadian guidelines (the Congestive Heart Failure, Hypertension, Age ≥ 75 Years [doubled], Diabetes Mellitus, Stroke [doubled], Vascular Disease, Age 65 to 74 Years, Sex Category [Female] [CHA2DS2-VASc] and Congestive Heart Failure, Hypertension, Age ≥ 65 Years, Diabetes, Stroke/Transient Ischemic Attack [CHADS-65] scores)
      • Andrade J.G.
      • Aguilar M.
      • Atzema C.
      • et al.
      The 2020 Canadian Cardiovascular Society/Canadian Heart Rhythm Society comprehensive guidelines for the management of atrial fibrillation.
      derived from at least 5 years of hospitalization, ED and physician claims data, and after excluding patients with contraindications to anticoagulation (intracranial hemorrhage at any time or major bleeding within 90 days before index date). We examined oral anticoagulants (OAC, including warfarin and DOACs) or low-molecular-weight heparin (LMWH) outpatient prescriptions within 120 days of first NVAF diagnosis in frail vs nonfrail patients, stratifying patients by guideline-based indications for anticoagulation (CHA2DS2-VASc of 2 or more for men or 3 or greater for women) and excluding those already taking oral anticoagulants (OACs) at baseline. Because the Canadian guidelines were modified in 2014 to endorse a CHADS-65 score of 1 or more as warranting anticoagulation, in a sensitivity analysis we used CHADS-65 ≥ 1 as the indication for anticoagulation. Logistic regression models (on the basis of clinically important factors and backward stepwise selection) for the effect of frailty on OAC prescribing were adjusted for calendar year, age, sex, cancer, chronic kidney disease (CKD), end-stage liver disease, thrombocytopenia, anemia, history of alcohol abuse, or excessive falls on the basis of information in the chart for the index event or any other hospitalizations in the previous 5 years, as well as catheter ablation within 5 years.
      All analyses were carried out using SAS 9.4 (SAS Institute Inc, Cary, NC) and significance was defined according to P values < 0.05.
      Among 75,796 patients (median age, 75 years; 45% female) discharged from an ED or hospital with a new diagnosis of NVAF between April 1, 2009, and March 31, 2019, 17,143 (22.6%) were frail (17.9% intermediate risk and 4.7% high risk on the HFRS), including 24.0% of patients younger than 65 years (Table 1). Frail patients were more likely to be women and older and had higher comorbidity burdens (Table 1), although even nonfrail patients had a median Elixhauser comorbidity score of 2 (interquartile range, 1.0-3.0).
      Table 1Baseline characteristics at time of new diagnosis of nonvalvular atrial fibrillation, stratified according to Hospital Frailty Risk Score
      Hospital Frailty Risk Score
      High (> 15)Intermediate (5-15)Not Frail (< 5)All
      Distribution3563 (4.7)13,580 (17.9)58,653 (77.4)75,796 (100.0)
      Female sex1997 (56.0)6861 (50.5)25,057 (42.7)33,915 (44.7)
      Median age (IQR)83 (75-89)81 (71-87)73 (62-82)75 (64-84)
      Age groups, years
       Age 20-64360 (10.1)1893 (13.9)17,113 (29.2)19,366 (25.6)
       Age 65-74507 (14.2)2564 (18.9)14,128 (24.1)17,199 (22.7)
       Age 75 or older2696 (75.7)9123 (67.2)27,412 (46.7)39,231 (51.8)
      Heart failure1514 (42.5)4809 (35.4)11,904 (20.3)18,227 (24.0)
      Hypertension3063 (86.0)10,847 (79.9)38,001 (64.8)51,911 (68.5)
      Diabetes1296 (36.4)4604 (33.9)14,549 (24.8)20,449 (27.0)
      Stroke/TIA650 (18.2)1557 (11.5)3275 (5.6)5482 (7.2)
      Peripheral arterial disease575 (16.1)1617 (11.9)3466 (5.9)5658 (7.5)
      Coronary artery disease1405 (39.4)5039 (37.1)15,941 (27.2)22,385 (29.5)
      Excess alcohol329 (9.2)1080 (8.0)2286 (3.9)3695 (4.9)
      Previous myocardial infarction832 (23.4)2772 (20.4)7025 (12.0)10,629 (14.0)
      Anemia1760 (49.4)4731 (34.8)7895 (13.5)14,386 (19.0)
      Falls2084 (58.5)5142 (37.9)10,975 (18.7)18,201 (24.0)
      Liver254 (7.1)773 (5.7)1172 (2.0)2199 (2.9)
      Thrombocytopenia188 (5.3)554 (4.1)685 (1.2)1427 (1.9)
      Cancer760 (21.3)2991 (22.0)8468 (14.4)12,219 (16.1)
      Chronic kidney disease884 (24.8)2419 (17.8)3389 (5.8)6692 (8.8)
      Median Elixhauser comorbidities (IQR)
      The number of Elixhauser comorbidities can range from 0 to 31.
      5.0 (4.0-7.0)4.0 (3.0-6.0)2.0 (1.0-3.0)2.0 (1.0-4.0)
      CHA2DS2-VASc Score
       033 (0.9)284 (2.1)5277 (9.0)5594 (7.4)
       175 (2.1)628 (4.6)7169 (12.2)7872 (10.4)
       2 or greater3455 (97.0)12,668 (93.3)46,207 (78.8)62,330 (82.2)
       Median (IQR)5.0 (4.0-6.0)4.0 (3.0-5.0)3.0 (2.0-4.0)3.0 (2.0-5.0)
      CHADS-65 Score
       057 (1.6)491 (3.6)8307 (14.2)8855 (11.7)
       1 or greater3506 (98.4)13,089 (96.4)50,346 (85.8)66,941 (88.3)
      Cranial hemorrhage (90 days prior)93 (2.6)220 (1.6)322 (0.5)635 (0.8)
      Major bleeding (90 days prior)62 (1.7)211 (1.6)303 (0.5)576 (0.8)
      Anticoagulation (120 days prior)566 (15.9)1991 (14.7)7334 (12.5)9891 (13.0)
      Data are presented as n (%) except where otherwise noted.
      CHADS-65, Congestive Heart Failure, Hypertension, Age ≥ 65 Years, Diabetes, Stroke/Transient Ischemic Attack; CHA2DS2-VASc, Congestive Heart Failure, Hypertension, Age ≥ 75 Years [doubled], Diabetes Mellitus, Stroke [doubled], Vascular Disease, Age 65 to 74 Years, Sex Category [Female]; IQR, interquartile range; TIA, transient ischaemic attack.
      The number of Elixhauser comorbidities can range from 0 to 31.
      Most NVAF patients met guideline criteria for anticoagulation (92.1% of frail patients vs 74.2% of nonfrail patients; P < 0.0001 for CHA2DS2-VASc and 96.8% vs 85.8%; P < .0001 for CHADS-65). However, 1.5% of these patients had a contraindication to anticoagulation, ranging from 1.0% of nonfrail patients to 4.1% of those with HFRS scores in the highest frailty risk group. After excluding those with contraindications, 50.1% of patients eligible for anticoagulation on the basis of their CHA2DS2-VASc score and 49.4% of those eligible on the basis of their CHADS-65 scores were using anticoagulants within 120 days of their NVAF diagnosis (Table 2, Fig. 1). Furthermore, although the proportion of individuals who should be anticoagulated increased as their frailty scores increased, the proportion using anticoagulants was lower in frail patients (Table 2) and the adjusted odds ratio for anticoagulant use in frail patients was 0.61 (95% confidence interval, 0.58-0.64; Table 3) compared with nonfrail patients, even after excluding those with contraindications.
      Table 2Anticoagulant use within 120 days of diagnosis in patients at high, intermediate, and low risk of frailty in whom anticoagulation is indicated, after excluding those with contraindications to anticoagulation
      Anticoagulation defined as being indicated if CHA2DS2-VASc score 2 or more for men, 3 or more for women, or CHADS-65 of 1 or more, and contraindications defined as intracranial hemorrhage at any time in past or major bleeding within 90 days before index date.
      Hospital Frailty Risk Score
      High (> 15)Intermediate (5-15)Not Frail (< 5)All
      Sample excluding those with contraindications3418 (4.6)13,174 (17.6)58,058 (77.8)74,650 (100.0)
      Anticoagulant indicated on the basis of CHA2DS2-VASc score
      Anticoagulant prescribed896 (27.4)4569 (38.0)23,731 (55.2)29,196 (50.1)
       Warfarin507 (15.5)2626 (21.8)13,070 (30.4)16,203 (27.8)
       Low molecular weight heparin114 (3.5)371 (3.1)2167 (5.0)2652 (4.6)
       Any DOAC358 (10.9)2005 (16.7)11,703 (27.2)14,066 (24.1)
       Apixaban169 (5.2)939 (7.8)4254 (9.9)5362 (9.2)
       Dabigatrin53 (1.6)332 (2.8)2281 (5.3)2666 (4.6)
       Rivaroxaban146 (4.5)770 (6.4)5475 (12.7)6391 (11.0)
      Anticoagulant indicated on the basis of CHADS-65 score
      Anticoagulant prescribed914 (27.2)4755 (37.4)26,890 (54.0)32,559 (49.4)
       Warfarin513 (15.3)2708 (21.3)14,397 (28.9)17,618 (26.8)
       Low molecular weight heparin117 (3.5)409 (3.2)2580 (5.2)3106 (4.7)
       Any DOAC368 (10.9)2090 (16.4)13,580 (27.3)16,038 (24.4)
       Apixaban174 (5.2)967 (7.6)4806 (9.7)5947 (9.0)
       Dabigatrin54 (1.6)347 (2.7)2631 (5.3)3032 (4.6)
       Rivaroxaban151 (4.5)815 (6.4)6492 (13.0)7458 (11.3)
      Data are presented as n (%).
      CHADS-65, Congestive Heart Failure, Hypertension, Age ≥ 65 Years, Diabetes, Stroke/Transient Ischemic Attack; CHA2DS2-VAScCongestive Heart Failure, Hypertension, Age ≥ 75 Years [doubled], Diabetes Mellitus, Stroke [doubled], Vascular Disease, Age 65 to 74 Years, Sex Category [Female]; DOAC, direct oral anticoagulant.
      Anticoagulation defined as being indicated if CHA2DS2-VASc score 2 or more for men, 3 or more for women, or CHADS-65 of 1 or more, and contraindications defined as intracranial hemorrhage at any time in past or major bleeding within 90 days before index date.
      Figure thumbnail gr1
      Figure 1Trends over time in anticoagulant prescription rate in frail and nonfrail patients who should have received an anticoagulant on the basis of stroke risk score and lack of contraindications. CHADS-65, Congestive Heart Failure, Hypertension, Age ≥ 65 Years, Diabetes, Stroke/Transient Ischemic Attack; CHA2DS2-VAScCongestive Heart Failure, Hypertension, Age ≥ 75 Years [doubled], Diabetes Mellitus, Stroke [doubled], Vascular Disease, Age 65 to 74 Years, Sex Category [Female]; DOAC, direct oral anticoagulant.
      Table 3Factors associated with anticoagulant use in patients in whom anticoagulation is indicated, after excluding patients with contraindications (intracranial hemorrhage at any time in the past or major bleed in the previous 90 days)
      VariableIndicated on the basis of CHA2DS2-VASc scoreIndicated on the basis of CHADS-65 score
      Adjusted OR95% CIAdjusted OR95% CI
      Frailty0.610.58-0.640.620.59-0.65
      Female sex1.020.98-1.061.010.97-1.05
      Age < 65 (referent 65-75)0.810.76-0.870.750.70-0.79
      Age 75 or older0.980.93-1.031.061.01-1.11
      Anemia0.770.73-0.820.770.73-0.81
      Thrombocytopenia0.850.72-1.000.830.70-0.97
      Cancer0.810.76-0.850.800.76-0.84
      End-stage liver disease0.690.59-0.800.700.61-0.80
      Excess alcohol0.790.71-0.890.730.66-0.81
      Falls0.750.71-0.790.760.72-0.79
      Chronic kidney disease0.860.80-0.920.870.81-0.94
      Catheter ablation0.400.21-0.760.520.32-0.85
      Models were adjusted for calendar year.
      CHADS-65, Congestive Heart Failure, Hypertension, Age ≥ 65 Years, Diabetes, Stroke/Transient Ischemic Attack; CHA2DS2-VAScCongestive Heart Failure, Hypertension, Age ≥ 75 Years [doubled], Diabetes Mellitus, Stroke [doubled], Vascular Disease, Age 65 to 74 Years, Sex Category [Female]; CI, confidence interval; OR, odds ratio.
      However, 33.2% of patients with CHA2DS2-VASc scores below treatment thresholds and 23.6% with CHADS-65 scores of 0 were prescribed anticoagulants, with a higher proportion of nonfrail patients receiving an anticoagulant when apparently not indicated than frail patients (34.1% vs 22.9%; P < 0.0001 for CHA2DS2-VASc and 24.0% vs 17.4%; P = 0.0007 for CHADS-65). However, 27.0% of patients prescribed anticoagulants despite having risk scores below treatment thresholds had an attempted cardioversion within 90 days (defined as a cardioversion procedure code or dispensations for either amiodarone or sotalol; Supplemental Table S1).
      Over the years we studied, anticoagulant prescription rates increased in frail (from 29.0% to 52.2% for those with CHA2DS2-VASc indications and 28.3% to 50.6% for CHADS-65) and nonfrail (from 42.4% to 68.2% for CHA2DS2-VASc and 41.2% to 66.7% for CHADS-65) patients meeting guideline indications for anticoagulation (Fig. 1). Similar increases were seen in DOAC prescribing between 2011 (when they first came on formulary in Alberta) to 2018: from 20.5% to 82.7% of OAC prescriptions in nonfrail patients and from 13.5% to 80.1% of OAC prescriptions in frail patients (Fig. 1). In the multivariable analyses (Table 3), other factors that decreased the likelihood of receiving any anticoagulant when it was indicated (in addition to frailty) were anemia, cancer, end stage liver disease, history of alcohol abuse, falls, CKD, or being younger than 65 years.
      For patients who were prescribed anticoagulation, frail patients were less likely to receive a DOAC compared with warfarin (adjusted odds ratio, 0.66; 95% confidence interval, 0.54-0.81; Table 4) and, even when prescribed, lower doses of DOAC were more commonly used in frail patients (Supplemental Table S2). Anemia, end-stage liver disease, CKD, and being older than 75 years were also associated with decreased odds of receiving a DOAC (meaning more likely to receive warfarin), whereas being younger than 65 years and being female was associated with a higher probability of receiving a DOAC compared with warfarin (Table 4).
      Table 4Factors associated with direct oral anticoagulant use (rather than warfarin) in patients with nonvalvular atrial fibrillation diagnosed after 2011 and in whom anticoagulation was indicated, after excluding patients with contraindications (intracranial hemorrhage at any time in the past or major bleed in the previous 90 days)
      VariableIndicated on the basis of CHA2DS2-VASc scoreIndicated on the basis of CHADS-65 score
      Adjusted OR95% CIAdjusted OR95% CI
      Frail0.660.54-0.810.650.54-0.79
      Female sex1.101.02-1.181.091.02-1.16
      Age < 65 (referent 65-75)1.211.06-1.371.301.17-1.44
      Age 75 or older0.890.82-0.970.840.78-0.91
      Anemia0.790.72-0.870.780.71-0.85
      Thrombocytopenia1.180.87-1.611.170.87-1.58
      Cancer1.060.96-1.171.030.93-1.13
      End-stage liver disease0.630.45-0.870.640.47-0.87
      Excess alcohol0.950.77-1.170.890.74-1.08
      Falls0.970.89-1.050.950.88-1.03
      Chronic kidney disease0.360.31-0.410.350.31-0.40
      Catheter ablation2.440.39-15.267.161.47-34.95
      The models were restricted to 2011 and forward to account for drug availability (DOACs only came on formulary in 2011) and includes adjustment for calendar year and the interaction between age and frailty and between end-stage liver disease and frailty (no other interactions were statistically significant).
      CHADS-65, Congestive Heart Failure, Hypertension, Age ≥ 65 Years, Diabetes, Stroke/Transient Ischemic Attack; CHA2DS2-VASc, Congestive Heart Failure, Hypertension, Age ≥ 75 Years [doubled], Diabetes Mellitus, Stroke [doubled], Vascular Disease, Age 65 to 74 Years, Sex Category [Female]; CI, confidence interval; DOAC, direct oral anticoagulant; OR, odds ratio.

      Discussion

      We found that frailty was relatively common (23%) and that frail patients had higher comorbidity burdens (consistent with previous studies). However, even nonfrail patients with NVAF frequently had a number of comorbidities (median Elixhauser comorbidity score in nonfrail patients was 2). This highlights that frailty assessment might provide an additional perspective beyond just consideration of age and comorbidities in patients with NVAF. Indeed, we have previously shown that, in patients with heart failure, the HFRS is independently associated with prognosis beyond age and comorbidity score.
      • McAlister F.A.
      • Savu A.
      • Ezekowitz J.A.
      • Armstrong P.W.
      • Kaul P.
      The Hospital Frailty Risk Score in patients with heart failure is strongly associated with outcomes but less so with pharmacotherapy.
      Although more than three-quarters of these patients with newly diagnosed NVAF met guideline indications for anticoagulation, less than half received an anticoagulant within 120 days of their initial diagnosis and the anticoagulant prescription rate decreased as the frailty score increased. This is in line with reports from other studies.
      • Ogilvie I.M.
      • Newton N.
      • Welner S.A.
      • Cowell W.
      • Lip G.Y.H.
      Underuse of oral anticoagulants in atrial fibrillation: a systematic review.
      ,
      • Wilkinson C.
      • Clegg A.
      • Todd O.
      • et al.
      Atrial fibrillation and oral anticoagulation in older people with frailty: a nationwide primary care electronic health records cohort study.

      Sanghai SR, Liu W, Wang W, et al. Prevalence of frailty and associations with oral anticoagulant prescribing in atrial fibrillation [e-pub ahead of print]. J Gen Intern Med doi:10.1007/s11606-021-06834-1, accessed December 2, 2021.

      • Mailhot T.
      • McManus D.D.
      • Waring M.E.
      • et al.
      Frailty, cognitive impairment, and anticoagulation among older adults with nonvalvular atrial fibrillation.
      ,
      • O’Brien E.C.
      • Holmes D.N.
      • Ansell J.E.
      • et al.
      Physician practices regarding contraindications to oral anticoagulation in atrial fibrillation: findings from the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) registry.
      ,
      • Lefebvre M.C.D.
      • St-Onge M.
      • Glazer-Cavanagh M.
      • et al.
      The effect of bleeding risk and frailty status on anticoagulation patterns in octogenarians with atrial fibrillation: the FRAIL-AF study.
      Between 2009 and 2018, rates of OAC dispensations increased in frail and nonfrail patients, although the increases were smaller in the frail subgroup. Qualitative studies reveal that many of the barriers often cited by physicians against the use of anticoagulants in patients with NVAF are related to frailty (such as fall risk and adherence issues).
      • Bungard T.J.
      • Ghali W.A.
      • Teo K.K.
      • McAlister F.A.
      • Tsuyuki R.T.
      Why do patients with atrial fibrillation not receive warfarin?.
      ,
      • Bungard T.J.
      • Ghali W.A.
      • McAlister F.A.
      • et al.
      Physicians’ perceptions of the benefits and risks of warfarin for patients with nonvalvular atrial fibrillation.
      However, the benefits of anticoagulation still outweigh the risks even in the presence of these “barriers.”
      • Kim D.H.
      • Pawar A.
      • Gagne J.J.
      • et al.
      Frailty and clinical outcomes of direct oral anticoagulants versus warfarin in older adults with atrial fibrillation: a cohort study.
      ,
      • van Walraven C.
      • Hart R.G.
      • Connolly S.
      • et al.
      Effect of age on stroke prevention therapy in patients with atrial fibrillation. The Atrial Fibrillation Investigators.
      • Donzé J.
      • Clair C.
      • Hug B.
      • et al.
      Risk of falls and major bleeds in patients on oral anticoagulation therapy.
      • Man-Son-Hing M.
      • Laupacis A.
      Anticoagulant-related bleeding in older persons with atrial fibrillation: physicians’ fears often unfounded.
      • Man-Son-Hing M.
      • Nichol G.
      • Lau A.
      • Laupacis A.
      Choosing antithrombotic therapy for elderly patients with atrial fibrillation who are at risk for falls.
      It is also worth noting that patients are much more willing to accept anticoagulant bleeding risks as a trade-off for preventing strokes than are physicians.
      • Rusin G.
      • Konieczyńska M.
      • Bijak P.
      • et al.
      Bleeding tolerance among patients with atrial fibrillation on oral anticoagulation.
      ,
      • Man-Son-Hing M.
      • Laupacis A.
      • O’Connor A.
      • et al.
      Warfarin for atrial fibrillation. The patient’s perspective.
      In our study population, frailty was strongly associated with non-prescription of an anticoagulant even when indicated. Although results have been mixed in previous studies, with some finding frailty negatively associated with anticoagulant prescriptions

      Sanghai SR, Liu W, Wang W, et al. Prevalence of frailty and associations with oral anticoagulant prescribing in atrial fibrillation [e-pub ahead of print]. J Gen Intern Med doi:10.1007/s11606-021-06834-1, accessed December 2, 2021.

      ,
      • O’Brien E.C.
      • Holmes D.N.
      • Ansell J.E.
      • et al.
      Physician practices regarding contraindications to oral anticoagulation in atrial fibrillation: findings from the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) registry.
      ,
      • Lefebvre M.C.D.
      • St-Onge M.
      • Glazer-Cavanagh M.
      • et al.
      The effect of bleeding risk and frailty status on anticoagulation patterns in octogenarians with atrial fibrillation: the FRAIL-AF study.
      and others not,
      • Wilkinson C.
      • Clegg A.
      • Todd O.
      • et al.
      Atrial fibrillation and oral anticoagulation in older people with frailty: a nationwide primary care electronic health records cohort study.
      ,
      • Mailhot T.
      • McManus D.D.
      • Waring M.E.
      • et al.
      Frailty, cognitive impairment, and anticoagulation among older adults with nonvalvular atrial fibrillation.
      this is not surprising because of the lack of consensus on how to assess frailty and the marked variability in frailty prevalence when different tools are used.
      • Wilkinson C.
      • Todd O.
      • Clegg A.
      • Gale C.P.
      • Hall M.
      Management of atrial fibrillation for older people with frailty: a systematic review and meta-analysis.
      Although the gold standard for assessment of frailty is a Comprehensive Geriatric Assessment of physical and mental function, this is unfortunately not available in administrative databases. However, the HFRS that we used has been validated in multiple populations
      • McAlister F.A.
      • van Walraven C.
      External validation of the Hospital Frailty Risk Score and comparison to the Hospital-patient One-year Mortality Risk Score to predict outcomes in elderly hospitalized patients: a retrospective cohort study.
      ,
      • Hollinghurst J.
      • Housley G.
      • Watkins A.
      • Clegg A.
      • Gilbert T.
      • Conroy S.P.
      A comparison of two national frailty scoring systems.
      and of the 35 frailty scales tested in the English Longitudinal Study of Aging (ELSA), the HFRS came closest to approximating the results of comprehensive geriatric assessments (the gold standard).
      • Aguayo G.A.
      • Donneau A.F.
      • Vaillant M.T.
      • et al.
      Agreement between 35 published frailty scores in the general population.
      Although DOACs have a better safety profile
      • Ruff C.T.
      • Giugliano R.P.
      • Braunwald E.
      • et al.
      Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised trials.
      and fewer drug and food interactions compared with warfarin, frail patients were less likely to receive a DOAC compared with warfarin. Considering that DOACs were approved in Canada only shortly before our study period (dabigatran in 2008, rivaroxaban and apixaban in 2012), clinicians might have proceeded cautiously in these early years because of concerns that participants in clinical trials might not have been representative of the frail population and because of the lack of antidotes for reversal of DOACs. Our finding that patients younger than 65 years were more likely to receive a DOAC possibly reflects the fact that DOACs are generally covered by younger patients’ work-related drug benefits whereas for patients older than 65 covered by the government-sponsored Blue Cross program, DOACs were only available via a special authorization process that required submission of documentation that the patient was unable to have their INRs monitored or had poorly controlled INRs with warfarin use).
      There are some limitations to our study. First, although our study provides data from a population-based contemporary cohort of NVAF patients in a single Canadian province, these results might not be generalizable to other countries where health care and drug benefits are not as widely available. Second, the data in this study were obtained from an administrative data set and therefore are subject to unmeasured confounders and potential misclassification bias, although we used previously validated case definitions for NVAF and all comorbidities, as well as the HFRS. Third, we relied on medication dispensations to define therapy use and do not have information on whether patients actually consumed the medications they were dispensed or whether they split pills or consumed them on alternate days (which might be more likely in frailer patients, perhaps accounting for some of the apparent underdosing of DOACs in frail patients that we observed). Fourth, we might have overestimated the proportion of patients taking anticoagulants without a guideline indication because we only used data for 5 years before the index date to define comorbidities and thus might have undercaptured some stroke risk factors. Further, approximately half of the patients prescribed anticoagulants despite having CHA2DS2-VASc scores below treatment thresholds had an attempt at cardioversion (either electrically or with antiarrhythmic medications) within 90 days and thus use of anticoagulants in those situations would not have been inappropriate (which is why we avoided use of that term in our work). Fifth, we could not calculate Hypertension, Abnormal Renal/Liver Function, Stroke, Bleeding History or Predisposition, Labile INR, Elderly (> 65 Years), Drugs/Alcohol Concomitantly (HAS-BLED) scores for our cohort because some elements necessary to calculate the HAS-BLED score (lability of INR results, over-the-counter nonsteroidal anti-inflammatory drug use, systolic blood pressure > 160 mm Hg, and drinking more than 8 units of alcohol per week) are not available in the administrative data sets we used. However, we did examine the factors associated with underuse of anticoagulants and use of DOACs compared with warfarin in our Table 3, Table 4 and included those elements of the HAS-BLED score (such as CKD or liver disease) for which we have validated case definitions in Canadian administrative databases.
      In conclusion, we found that even after excluding those with contraindications and focusing only on those with guideline-recommended indications for anticoagulation, frail patients were less likely to receive an anticoagulant and when anticoagulated they were less likely to receive a DOAC compared with warfarin. These are important findings because the proportion of individuals with NVAF who are frail is increasing over time and there is a widespread assumption that the treatment gap in frail patients will have improved since introduction of DOACs—our data demonstrate this is a misconception. Because frail patients have higher event rates and thus stand to potentially derive greater benefits from anticoagulation than their nonfrail counterparts, we encourage clinicians to evaluate absolute risks and benefits when making prescribing decisions. We also encourage educators and those responsible for the dissemination of guidelines to focus attention on these treatment gaps in frail NVAF patients. Finally, we encourage clinical trialists to include more frail patients in future randomized trials on NVAF to establish a more robust evidence base in this area.

      Data access and responsibility

      Dr Dover conducted the statistical analyses and takes responsibility for the integrity and accuracy of the data. Dr McAlister had full access to the analyses.

      Data availability statement

      The data underlying this article was provided by the Government of Alberta under the terms of a research agreement and the authors are not custodians of the data. Inquiries regarding access to the data for research projects with ethics approval can be made to [email protected] .

      Acknowledgements

      This study is on the basis of data provided by the Customer Relationship Management and Data Access Unit at Alberta Health. The interpretation and conclusions are those of the researchers and do not represent the views of the Government of Alberta nor Alberta Health.

      Funding Sources

      The study was funded by operating grants from University Hospital Foundation , Servier Canada , and the Heart and Stroke Foundation . The funders had no input into the design, analysis, or interpretation of the data or the decision to submit for publication. Dr Kaul is a CIHR Chair in Sex and Gender Science and is supported by the Heart & Stroke Chair in Cardiovascular Research. Dr McAlister is supported by an Alberta Health Services Chair in Cardiovascular Outcomes Research.

      Disclosures

      Dr Sandhu holds research grants from Servier and Pfizer/BMS but neither organization had any input into this study concept, design, conduct, analysis, or write-up. The remaining authors have no conflicts of interest to disclose .

      Supplementary Material

      References

        • Go A.S.
        • Hylek E.M.
        • Phillips K.A.
        • et al.
        Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study.
        JAMA. 2001; 285: 2370-2375
        • Hughes M.
        • Lip G.Y.
        Stroke and thromboembolism in atrial fibrillation: a systematic review of stroke risk factors, risk stratification schema and cost effectiveness data.
        Thromb Haemost. 2008; 99: 295-304
        • Andrade J.G.
        • Aguilar M.
        • Atzema C.
        • et al.
        The 2020 Canadian Cardiovascular Society/Canadian Heart Rhythm Society comprehensive guidelines for the management of atrial fibrillation.
        Can J Cardiol. 2020; 36: 1847-1948
        • Ogilvie I.M.
        • Newton N.
        • Welner S.A.
        • Cowell W.
        • Lip G.Y.H.
        Underuse of oral anticoagulants in atrial fibrillation: a systematic review.
        Am J Med. 2010; 123: 638-645
        • Morley J.E.
        • Vellas B.
        • van Kan G.A.
        • et al.
        Frailty consensus: a call to action.
        J Am Med Dir Assoc. 2013; 14: 392-397
        • Villacampa-Fernandez P.
        • Navarro-Pardo E.
        • Tarın J.J.
        • Cano A.
        Frailty and multimorbidity: two related yet different concepts.
        Maturitas. 2017; 95: 31-35
        • Clegg A.
        • Young J.
        • Iliffe S.
        • Rikkert M.O.
        • Rockwood K.
        Frailty in elderly people.
        Lancet. 2013; 381: 752-762
        • Villani E.R.
        • Tummolo A.M.
        • Palmer K.
        • et al.
        Frailty and atrial fibrillation: a systematic review.
        Eur J Intern Med. 2018; 56: 33-38
        • Wilkinson C.
        • Todd O.
        • Clegg A.
        • Gale C.P.
        • Hall M.
        Management of atrial fibrillation for older people with frailty: a systematic review and meta-analysis.
        Age Ageing. 2019; 48: 196-203
        • Alexander K.P.
        • Brouwer M.A.
        • Mulder H.
        • et al.
        Outcomes of apixaban versus warfarin in patients with atrial fibrillation and multi-morbidity: insights from the ARISTOTLE trial.
        Am Heart J. 2019; 208: 123-131
        • Wilkinson C.
        • Wu J.
        • Searle S.D.
        • et al.
        Clinical outcomes in patients with atrial fibrillation and frailty: insights from the ENGAGE AF-TIMI 48 trial.
        BMC Med. 2020; 18: 401
        • Zhang J.
        • Wang X.
        • Liu X.
        • et al.
        Comparative effectiveness and safety of direct acting oral anticoagulants in nonvalvular atrial fibrillation for stroke prevention: a systematic review and meta-analysis.
        Eur J Epidemiol. 2021; 36: 793-812
        • Kim D.H.
        • Pawar A.
        • Gagne J.J.
        • et al.
        Frailty and clinical outcomes of direct oral anticoagulants versus warfarin in older adults with atrial fibrillation: a cohort study.
        Ann Intern Med. 2021; 174: 1214-1223
        • Wilkinson C.
        • Clegg A.
        • Todd O.
        • et al.
        Atrial fibrillation and oral anticoagulation in older people with frailty: a nationwide primary care electronic health records cohort study.
        Age Ageing. 2021; 50: 772-779
      1. Sanghai SR, Liu W, Wang W, et al. Prevalence of frailty and associations with oral anticoagulant prescribing in atrial fibrillation [e-pub ahead of print]. J Gen Intern Med doi:10.1007/s11606-021-06834-1, accessed December 2, 2021.

        • Mailhot T.
        • McManus D.D.
        • Waring M.E.
        • et al.
        Frailty, cognitive impairment, and anticoagulation among older adults with nonvalvular atrial fibrillation.
        J Am Geriatr Soc. 2020; 68: 2778-2786
        • Hawkins N.M.
        • Daniele P.R.
        • Humphries K.
        • et al.
        Empirical insights when defining the population burden of atrial fibrillation and oral anticoagulation uptake using administrative data.
        Can J Cardiol. 2019; 35: 1412-1415
        • Gilbert T.
        • Neuburger J.
        • Kraindler J.
        • et al.
        Development and validation of a Hospital Frailty Risk Score focusing on older people in acute care settings using electronic hospital records: an observational study.
        Lancet. 2018; 391: 1775-1782
        • McAlister F.A.
        • Lin M.
        • Bakal J.A.
        Prevalence and postdischarge outcomes associated with frailty in medical inpatients: impact of different frailty definitions.
        J Hosp Med. 2019; 14: 407-410
        • Kim D.H.
        • Patorno E.
        • Pawar A.
        • Lee H.
        • Schneeweiss S.
        • Glynn R.J.
        Measuring frailty in administrative claims data: comparative performance of four claims-based frailty measures in the U.S. Medicare data.
        J Gerontol A Biol Sci Med Sci. 2020; 75: 1120-1125
        • McAlister F.A.
        • van Walraven C.
        External validation of the Hospital Frailty Risk Score and comparison to the Hospital-patient One-year Mortality Risk Score to predict outcomes in elderly hospitalized patients: a retrospective cohort study.
        BMJ Qual Saf. 2019; 28: 284-288
        • Hollinghurst J.
        • Housley G.
        • Watkins A.
        • Clegg A.
        • Gilbert T.
        • Conroy S.P.
        A comparison of two national frailty scoring systems.
        Age Ageing. 2021; 50: 1208-1214
        • McAlister F.A.
        • Savu A.
        • Ezekowitz J.A.
        • Armstrong P.W.
        • Kaul P.
        The Hospital Frailty Risk Score in patients with heart failure is strongly associated with outcomes but less so with pharmacotherapy.
        J Intern Med. 2020; 287: 322-332
        • O’Brien E.C.
        • Holmes D.N.
        • Ansell J.E.
        • et al.
        Physician practices regarding contraindications to oral anticoagulation in atrial fibrillation: findings from the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) registry.
        Am Heart J. 2014; 167: 601-609
        • Lefebvre M.C.D.
        • St-Onge M.
        • Glazer-Cavanagh M.
        • et al.
        The effect of bleeding risk and frailty status on anticoagulation patterns in octogenarians with atrial fibrillation: the FRAIL-AF study.
        Can J Cardiol. 2016; 32: 169-176
        • Bungard T.J.
        • Ghali W.A.
        • Teo K.K.
        • McAlister F.A.
        • Tsuyuki R.T.
        Why do patients with atrial fibrillation not receive warfarin?.
        Arch Intern Med. 2000; 160: 41-46
        • Bungard T.J.
        • Ghali W.A.
        • McAlister F.A.
        • et al.
        Physicians’ perceptions of the benefits and risks of warfarin for patients with nonvalvular atrial fibrillation.
        CMAJ. 2001; 165: 301-302
        • van Walraven C.
        • Hart R.G.
        • Connolly S.
        • et al.
        Effect of age on stroke prevention therapy in patients with atrial fibrillation. The Atrial Fibrillation Investigators.
        Stroke. 2009; 40: 1410-1416
        • Donzé J.
        • Clair C.
        • Hug B.
        • et al.
        Risk of falls and major bleeds in patients on oral anticoagulation therapy.
        Am J Med. 2012; 125: 773-778
        • Man-Son-Hing M.
        • Laupacis A.
        Anticoagulant-related bleeding in older persons with atrial fibrillation: physicians’ fears often unfounded.
        Arch Intern Med. 2003; 163: 1580-1586
        • Man-Son-Hing M.
        • Nichol G.
        • Lau A.
        • Laupacis A.
        Choosing antithrombotic therapy for elderly patients with atrial fibrillation who are at risk for falls.
        Arch Intern Med. 1999; 159: 677-685
        • Rusin G.
        • Konieczyńska M.
        • Bijak P.
        • et al.
        Bleeding tolerance among patients with atrial fibrillation on oral anticoagulation.
        Can J Cardiol. 2020; 36: 500-508
        • Man-Son-Hing M.
        • Laupacis A.
        • O’Connor A.
        • et al.
        Warfarin for atrial fibrillation. The patient’s perspective.
        Arch Intern Med. 1996; 156: 1841-1848
        • Aguayo G.A.
        • Donneau A.F.
        • Vaillant M.T.
        • et al.
        Agreement between 35 published frailty scores in the general population.
        Am J Epidemiol. 2017; 186: 420-434
        • Ruff C.T.
        • Giugliano R.P.
        • Braunwald E.
        • et al.
        Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised trials.
        Lancet. 2014; 383: 955-962