TY - JOUR
T1 - Duration of Diabetes Mellitus and Risk of Thromboembolism and Bleeding in Atrial Fibrillation
T2 - Nationwide Cohort Study
AU - Overvad, Thure Filskov
AU - Skjøth, Flemming
AU - Lip, Gregory Y H
AU - Lane, Deirdre
AU - Albertsen, Ida Ehlers
AU - Rasmussen, Lars Hvilsted
AU - Larsen, Torben Bjerregaard
N1 - © 2015 American Heart Association, Inc.
PY - 2015/8
Y1 - 2015/8
N2 - BACKGROUND AND PURPOSE: Guidelines advocate anticoagulant treatment to all patients with atrial fibrillation and concomitant diabetes mellitus. The potential refinement to thromboembolic risk stratification that may spring from subdividing diabetes mellitus is unexplored. The purpose was to investigate duration of diabetes mellitus as a predictor of thromboembolism and anticoagulant-related bleeding in patients with atrial fibrillation.METHODS: Using nationwide Danish registries, we identified all patients discharged from hospital with an incident diagnosis of atrial fibrillation from 2000 to 2011. Hazard ratios with 95% confidence intervals for thromboembolism and bleeding according to years of diabetes mellitus duration in categories (0-4, 5-9, 10-14, and ≥15) and as a continuous variable using cubic splines were calculated by Cox regression.RESULTS: The study population comprised 137 222 patients with atrial fibrillation, of which 12.4% had diabetes mellitus. Compared with patients without diabetes mellitus and after adjustment for anticoagulant treatment and CHA2DS2-VASc components (congestive heart failure, hypertension, age, previous stroke, vascular disease, and sex), the risk of thromboembolism was lowest in the 0 to 4 years duration category (hazard ratio, 1.11; 95% confidence interval, 1.03-1.20), and highest in the longest duration category of ≥15 years (hazard ratio, 1.48; 95% confidence interval, 1.29-1.70). When analyzed as a continuous variable, duration of diabetes mellitus was associated with risk of thromboembolism in a dose-response-dependent manner, but not with a higher risk of bleeding during anticoagulant treatment.CONCLUSIONS: In patients with atrial fibrillation, longer duration of diabetes mellitus was associated with a higher risk of thromboembolism, but not with a higher risk of anticoagulant-related bleeding. Considering the critical balance between preventing thromboembolism and avoiding bleeding, longer duration of diabetes mellitus may favor initiation of anticoagulant therapy.
AB - BACKGROUND AND PURPOSE: Guidelines advocate anticoagulant treatment to all patients with atrial fibrillation and concomitant diabetes mellitus. The potential refinement to thromboembolic risk stratification that may spring from subdividing diabetes mellitus is unexplored. The purpose was to investigate duration of diabetes mellitus as a predictor of thromboembolism and anticoagulant-related bleeding in patients with atrial fibrillation.METHODS: Using nationwide Danish registries, we identified all patients discharged from hospital with an incident diagnosis of atrial fibrillation from 2000 to 2011. Hazard ratios with 95% confidence intervals for thromboembolism and bleeding according to years of diabetes mellitus duration in categories (0-4, 5-9, 10-14, and ≥15) and as a continuous variable using cubic splines were calculated by Cox regression.RESULTS: The study population comprised 137 222 patients with atrial fibrillation, of which 12.4% had diabetes mellitus. Compared with patients without diabetes mellitus and after adjustment for anticoagulant treatment and CHA2DS2-VASc components (congestive heart failure, hypertension, age, previous stroke, vascular disease, and sex), the risk of thromboembolism was lowest in the 0 to 4 years duration category (hazard ratio, 1.11; 95% confidence interval, 1.03-1.20), and highest in the longest duration category of ≥15 years (hazard ratio, 1.48; 95% confidence interval, 1.29-1.70). When analyzed as a continuous variable, duration of diabetes mellitus was associated with risk of thromboembolism in a dose-response-dependent manner, but not with a higher risk of bleeding during anticoagulant treatment.CONCLUSIONS: In patients with atrial fibrillation, longer duration of diabetes mellitus was associated with a higher risk of thromboembolism, but not with a higher risk of anticoagulant-related bleeding. Considering the critical balance between preventing thromboembolism and avoiding bleeding, longer duration of diabetes mellitus may favor initiation of anticoagulant therapy.
KW - Aged
KW - Aged, 80 and over
KW - Atrial Fibrillation
KW - Cohort Studies
KW - Denmark
KW - Diabetes Mellitus
KW - Female
KW - Hemorrhage
KW - Humans
KW - Male
KW - Middle Aged
KW - Registries
KW - Risk Factors
KW - Thromboembolism
KW - Time Factors
U2 - 10.1161/STROKEAHA.115.009371
DO - 10.1161/STROKEAHA.115.009371
M3 - Article
C2 - 26152296
SN - 0039-2499
VL - 46
SP - 2168
EP - 2174
JO - Stroke
JF - Stroke
IS - 8
ER -
