Best Hospitals, U.S. News and World Report, National Ranked in 13 specialties - 2014-15

Duke University Hospital: consistently ranked among the nation's best.

Duke Perspectives

Cause-Specific Mortality Among Patients Taking Anticoagulation for AF

Although atrial fibrillation (AF) increases the risk of stroke, stroke is an uncommon cause of cardiovascular-related deaths among patients with AF taking an oral anticoagulant. Rather, new study findings demonstrate that sudden cardiac death and heart failure (HF) are the most common cardiovascular deaths in this population. In addition, a large proportion of deaths are not cardiovascular related.

Oral anticoagulation reduces the risk of stroke and stroke-related death in patients with AF, but the specific causes of death are poorly understood, said Abhinav Sharma, MD, of Duke, who presented findings of the study at the Scientific Sessions of the American Heart Association held November 12 to 16, 2016, in New Orleans, LA. Other Duke investigators involved in the study were Renato D. Lopes, MD, PhD, John H. Alexander, MD, and Christopher B. Granger, MD.

Sharma and colleagues sought to determine the causes of death and clinical predictors for cause-specific mortality among patients taking anticoagulation agents for AF who had at least 1 risk factor for stroke. To answer these questions, the investigators analyzed data on patients enrolled in the Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation (ARISTOTLE) trial.

Among the 18,201 patients enrolled in the ARISTOTLE trial, 1,272 participants died. All of these deaths were centrally adjudicated by an events committee and classified as cardiovascular or non-cardiovascular related. Cardiovascular-related deaths were further classified as related to HF, stroke (ischemic, hemorrhagic) or systemic embolism, bleeding, or sudden cardiac death. The median follow-up time was 1.8 years.

Sharma reported that 51% of deaths were cardiovascular, 3% were caused by bleeding, and 46% were non-cardiovascular related. The most common cause of cardiovascular death was sudden cardiac death (39%), followed by HF (26%) and stroke (16%).

Other causes were myocardial infarction (6%), unobserved (6%), and other (7%). Stroke-related deaths were not adjudicated, Sharma commented, so the underlying mechanisms of the sudden cardiac deaths were unknown.

Different demographics and clinical profiles were related to different causes of death. For example, sudden cardiac death was associated with the youngest age (72 years) and the lowest percentage of women (22%). Prior HF was the most common cause of HF-related death (59%), and prior stroke was the least common cause of stroke-related death (32%).

Multivariable models identified the predictors of cause-specific mortality (Table).

Table. Clinical Predictors of Cause-Specific Mortality

Cause of Death Multivariable Predictor HR P Value
Any cardiovascular death History of HF 2.0 < .001
Age (10-y increase) 1.3 < .001
Sudden cardiac death History of HF 2.1 < .001
Sex (male vs female) 2.2 < .001
HF History of HF 2.8 < .001
Hb level (1 g/dL decrease) 1.3 < .001
SBP (10 mm Hg increase) 0.8 < .001
Stroke/systemic embolism Prior stroke, systemic embolism, TIA 2.7 < .001
Age (10-y increase) 1.6 < .001
Apixaban over warfarin 0.6 .02
History of HF 1.6 .008

Hb = hemoglobin, HF = heart failure, HR = hazard ratio, SBP = systolic blood pressure, TIA = transient ischemic attack.

History of HF and age were the main risk factors for cardiovascular mortality. A history of HF was a strong predictor of cause of death and was associated with every type of cardiovascular death studied. Prior stroke, systemic embolism, and transient ischemic attack were strong predictors of death caused by stroke or systemic embolism.

The researchers also evaluated stroke mortality and found that apixaban reduced the risk of death due to stroke and systemic embolism by 40%. This reduction may have been because of a reduction in deaths due to hemorrhagic stroke.

The mortality rates for hemorrhagic stroke were 0.10% for study patients taking apixaban and 0.24% for those taking warfarin (hazard ratio [HR] 0.42; 95% confidence interval [CI], 0.24-0.75; P = .003). In comparison, the mortality rates for ischemic or uncertain stroke were 0.12% for study patients taking apixaban and 0.15% for those taking warfarin (HR 0.84; 95% CI, 0.47-1.49; P = .55).

Overall, Sharma concluded that these findings suggest that strategies to improve outcomes for patients being treated with anticoagulants for AF should include treatments to prevent deaths related to HF and other comorbidities.

Source: Sharma A, Hijazi Z, Andersson U, et al. Risk factors for cause-specific mortality in patients anticoagulated for atrial fibrillation: insights from the ARISTOTLE trial. Presented at: American Heart Association Scientific Sessions 2016; November 12-16, 2016; New Orleans, LA. Abstract 812.

More From This Cardiology & Heart Surgery Issue
Apo-A1 Is Well Tolerated and May Be Safe in Patients After Recent MI

Apo-A1 has the potential to reduce major adverse cardiovascular events.

Read more >

Format of Risk Communications Influences Perceived Risk and Willingness for Therapy

Communicating lifetime rather than short-term risk estimates has a greater influence on how patients perceive their risk of CVD.

Read more >


Duke Health is proud to provide you with complimentary access to the latest image quizzes, case challenges, practice management tips, and more for your specialty.
Confirm my free access