CLINICAL RESEARCH STUDY

Prior History of Falls and Risk of Outcomes in Atrial Fibrillation: The Loire Valley Atrial Fibrillation Project Amitava Banerjee, MPH, DPhil,a Nicolas Clementy, MD,b Ken Haguenoer, MD,b Laurent Fauchier, MD, PhD,b,1 Gregory Y.H. Lip, MDa,1 a University of Birmingham Centre for Cardiovascular Sciences, City Hospital, Birmingham, UK; bService de Cardiologie, Pôle Coeur Thorax Vasculaire, Centre Hospitalier, Universitaire Trousseau et Faculté de Médecine, Université François Rabelais, Tours, France.

ABSTRACT BACKGROUND: Patients with nonvalvular atrial fibrillation are often denied oral anticoagulation due to falls risk. The latter is variably defined, and existing studies have not compared the associated risk of bleeding with other cardiovascular events. There are no data about outcomes in individuals with nonvalvular atrial fibrillation with a prior history of (actual) falls, rather than being “at risk of falls.” Our objective was to evaluate the risk of cardiovascular outcomes associated with prior history of falls in patients with atrial fibrillation in a contemporary “real world” cohort. METHODS: Patients with nonvalvular atrial fibrillation in a 4-hospital institution between 2000 and 2010 were included. Stroke/thromboembolism event rates were calculated according to prior history of falls. Risk factors were investigated by Cox regression. RESULTS: Among 7156 atrial fibrillation patients, prior history of falls/trauma was uncommon (n ¼ 76; 1.1%). Compared with patients without history of falls, those patients were older and less likely to be on oral anticoagulation; they also had higher risk scores for stroke/thromboembolism but not for bleeding. Compared with no prior history of falls, rates of stroke/thromboembolism (P ¼ .01) and all-cause mortality (P < .0001) were significantly higher in patients with previous falls. In multivariable analyses, prior history of falls was independently associated with stroke/thromboembolism (hazard ratio [HR] 5.19; 95% confidence interval [CI], 2.1-12.6; P < .0001), major bleeding (HR 3.32 [1.23-8.91]; P ¼ .02), and all-cause mortality (HR 3.69; 95% CI, 1.52-8.95; P ¼ .04), but not hemorrhagic stroke (HR 4.20; 95% CI, 0.58-30.48; P ¼ .16) in patients on oral anticoagulation. CONCLUSION: In this large “real world” atrial fibrillation cohort, prior history of falls was uncommon but independently increased risk of stroke/thromboembolism, bleeding, and mortality, but not hemorrhagic stroke in the presence of anticoagulation. Prior history of (actual) falls may be a more clinically useful risk prognosticator than “being at risk of falls.” Ó 2014 Elsevier Inc. All rights reserved.  The American Journal of Medicine (2014) 127, 972-978 KEYWORDS: Atrial fibrillation; Bleeding; Falls; Stroke; Thromboembolism

The global burden of nonvalvular atrial fibrillation and ischemic stroke/thromboembolism is unquestionable.1-3 Funding: See last page of article. Conflicts of Interest: See last page of article. Authorship: See last page of article. Requests for reprints should be addressed to Gregory Y.H. Lip, University of Birmingham Centre for Cardiovascular Sciences, City Hospital, Birmingham B18 7QH, UK. E-mail address: [email protected] 1 Joint senior authors. 0002-9343/$ -see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjmed.2014.05.035

Oral anticoagulation, most commonly with the vitamin K antagonists (VKA, eg, warfarin) but also with non-VKA oral anticoagulants,4,5 confers a well-established prognostic benefit for prevention of ischemic stroke/thromboembolism in the setting of atrial fibrillation. All oral anticoagulants also confer a risk of bleeding,6 but even in patients with high levels of comorbidity, the net clinical benefit is still in favor of oral anticoagulation.7,8 The same is probably true for non-VKA oral anticoagulants.5 Therefore, oral anticoagulation is recommended in all

Banerjee et al

Falls and Risk of Outcomes in Atrial Fibrillation

973

individuals with atrial fibrillation other than those who are at institution in an area of around 4000 km2, serving approx9 truly low risk of ischemic stroke/thromboembolism. Howimately 400,000 inhabitants. All patients diagnosed with AF or atrial flutter by the Cardiology Department between 2000 ever, despite improved clinical risk prediction tools for risk and 2010 were identified,28 excluding patients with valvular stratification of ischemic stroke/thromboembolism and 10,11 bleeding and consensus guideline recommendations, AF. Patients were followed from the first record of AF after January 1, 2000 (ie, index date) up to the latest levels of oral anticoagulation are still suboptimal in clinical data collection at the time of the practice, especially in the elderly.12 study (December 2010). TreatOne of the commonest reasons CLINICAL SIGNIFICANCE ment at discharge was obtained by for not giving oral anticoagulation screening hospitalization reports, is a perceived risk of falls.13-17  Patients with nonvalvular atrial fibrillaand information on comorbidities Falls are a significant cause of tion (NVAF) are often denied oral antiwas obtained from the computermorbidity and mortality, particucoagulation due to falls risk. ized coding system. larly in older populations, and  Among 7156 NVAF patients, prior history Prior history of falls was asincur high costs to individuals and certained from clinical history or health systems.18 Of note, the etiof falls/trauma was uncommon (1.1%), medical records. For each patient, ology of falls is multifactorial.19,20 but independently increased risk of the CHADS210 and CHA2DS2Atrial fibrillation (AF) itself is an stroke/thromboembolism, bleeding, and 21,22 independent predictor of falls. VASc11 scores were calculated. mortality, but not hemorrhagic stroke in The CHADS2 score was the sum In the setting of AF, the main the presence of anticoagulation. focus of the physician’s concern of points obtained after adding  Prior history of falls may be a more about falls is the increased pre1 point for congestive heart failclinically useful risk prognosticator than disposition to major bleeding if ure, hypertension, age 75 years, oral anticoagulation is initiand diabetes, and 2 points for “risk of falls.” ated.14,15 However, patients on previous stroke or thromboembolism.11 The CHA2DS2-VASc score oral anticoagulation at high risk of falls do not necessarily have a significantly increased risk of was the sum of points after adding 1 point for congestive major bleeds,23 suggesting that being at risk of falls is not a heart failure, hypertension, diabetes, vascular disease (including history of coronary, cerebrovascular or periphcontraindication to oral anticoagulation. Even studies that eral vascular disease), age 65-74 years, and female sex, and have shown a high rate of intracranial hemorrhage in AF 2 points for previous stroke or thromboembolism and age patients with high risk of falls suggest that there is still an 75 years.11 According to the 2 risk scores, patients with a overall benefit of oral anticoagulation due to prevention of 24 ischemic stroke. score of 0 on either schema were considered as “low risk,” 1 as “intermediate risk,” and 2 as “high risk” of stroke and Some studies have considered patients with AF at thromboembolism. increased risk of falls but with varying definitions and in The Hypertension, Abnormal renal and/or liver different subpopulations.21-27 A previous history of (actual) function, Stroke, Bleeding history or predisposition, Labile falls is probably the strongest risk factor for future falls, but International Normalized Ratio, Elderly [>65 years], the risk of cardiovascular events associated with previous Drugs [antiplatelet drugs or nonsteroidal anti-inflammatory history of falls has not been compared with risk of major drugs]/alcohol excess concomitantly (HAS-BLED) score is bleeding in AF patients. a validated scoring system for bleeding risk stratification In the first contemporary study of its kind, our objective in AF patients.29 For each patient, the HAS-BLED score was to evaluate the risk of bleeding and cardiovascular outcomes associated with a prior history of (actual) fallse was also calculated as the sum of the points obtained after rather than “being at risk of falls” per seein a large, “realadding 1 point for the presence of each individual factor). world” cohort of individuals with AF. We tested the Patients with a HAS-BLED score of 0-2 were deemed to hypothesis that prior history of falls would have an impact have “low” bleeding risk and those with a HAS-BLED score on ischemic stroke/thromboembolism, bleeding, and morof 3 were classified as “high” bleeding risk. tality in patients with AF. During follow-up, information on outcomes of thromboembolism, stroke (ischemic or hemorrhagic), major bleeding, and all-cause mortality were recorded. Major bleeding was defined as bleeding with a reduction in the METHODS hemoglobin level of at least 2 g per liter, or with transfusion Study Population of at least 1 unit of blood, or symptomatic bleeding in a critical area or organ (eg, intracranial, intraspinal, intraocThe methods of the Loire Valley Atrial Fibrillation Project, ular, retroperitoneal, intra-articular or pericardial, or intrawhich is based at the Centre Hospitalier Régional et Unimuscular with compartment syndrome) or bleeding that versitaire in Tours (France), have been reported previcauses death. All bleeding data were identified with the ously.28 The institution includes 4 hospitals covering all diagnosis coded in a subsequent hospitalization during medical and surgical specialties and is the only public

974

The American Journal of Medicine, Vol 127, No 10, October 2014

follow-upethus, we recorded all “hospitalizations with a bleed” as an additional criterion for major bleeding.

Statistical Analysis Risk factors were investigated by Cox regression. Baseline characteristics were determined based on prior history of falls. Differences were investigated using chi-squared test for categorical covariates and Kruskal-Wallis test for continuous covariates. Table 1

Event rates of ischemic stroke/thromboembolism, bleeding, and all-cause mortality were calculated for all patients by prior history of falls, stratifying by presence or absence of VKA therapy. Hemorrhagic strokes were excluded from analyses of “ischemic stroke” or “ischemic stroke/thromboembolism.” Event rates also were calculated by age and sex categories. The risk associated with prior history of falls was estimated in Cox proportional-hazard models. Both univariate and multivariate (including all the CHA2DS2-VASc risk factors) Cox regression models were

Characteristics of Patients with Atrial Fibrillation by Risk of Falls

n (%) Mean age (SD) Female Type of AF Paroxysmal Permanent Persistent Comorbidities Hypertension Diabetes Previous stroke Coronary artery disease Any vascular disease Heart failure Renal impairment Liver impairment Dyslipidemia Smoking Pacemaker/ICD Bleeding risk factors Previous bleeding Labile INR Anemia NSAIDs Drugs Cancer (active) Thrombocytopenia Antithrombotic agents Vitamin K antagonist Antiplatelets Any antithrombotic Other medical therapy ACE-I Beta-blocker Digoxin Diuretic Antiarrhythmic agent Calcium channel blocker CHADS2, Mean (SD) CHA2DS2-VASc, Mean (SD) HAS-BLED, Mean (SD)

Prior History of Falls n ¼ 76

No History of Falls n ¼ 7080

82.9 (8.9) 43 (56.6)

69.9 (15.1) 2661 (37.6)

48 (63.2) 26 (34.2) 2 (2.6)

4128 (58.3) 2578 (36.4) 374 (5.3)

40 19 12 17 21 24 12 0 11 5 13

(52.6) (25.0) (15.8) (22.4) (27.6) (31.6) (15.8) (0) (14.5) (6.6) (17.1)

2992 1089 582 2113 2339 3502 540 19 1352 912 1141

6 3 0 1 10 0 0

(7.9) (3.9) (0) (1.3) (13.2) (0) (0)

323 119 41 9 1225 119 6

17 (22.4) 33 (43.4) 44 (57.9) 21 16 10 22 12 4 2.7 4.4 2.07

(27.6) (21.0) (13.2) (28.9) (15.8) (5.3) (1.3) (1.6) (1.03)

P-Value

Age-adjusted P-Value

Prior history of falls and risk of outcomes in atrial fibrillation: the Loire Valley Atrial Fibrillation Project.

Patients with nonvalvular atrial fibrillation are often denied oral anticoagulation due to falls risk. The latter is variably defined, and existing st...
198KB Sizes 2 Downloads 3 Views