Use of Anticoagulation in Elderly Patients with Atrial Fibrillation Who Are at Risk for Falls

Candice L Garwood, PharmD BCPS; Tia L Corbett, PharmD


The Annals of Pharmacotherapy. 2008;42(4):523-532. 

In This Article

Choosing Antithrombotic Therapy in Patients with Risk for Falls

Often, elderly patients at high risk for falls are excluded from clinical trials, resulting in a lack of longitudinal data to quantify risks of anticoagulation. Few studies have appraised this issue. One retrospective study of 1245 Medicare beneficiaries quantified the incidence of ICH and the benefits of warfarin therapy in elderly patients at high risk of falls.[54] A subgroup analysis was conducted based on CHADS2 risk score. Patients with a high risk of falls suffered ICH more than twice as often as other study subjects. Those with ICH who were prescribed warfarin had a significantly higher 30-day mortality rate compared with patients not prescribed warfarin. Despite this finding, the investigation supported the use of warfarin in patients at moderate or high risk of stroke (CHADS2 ≥2), even with the odds of falling. In contrast, the benefit of warfarin use in those with a low risk of stroke (CHADS2 0-1) was uncertain. Therefore, the authors recommended use of aspirin or no therapy ( Table 4 ).[54]

Another examination of antithrombotic therapy in patients with AF at high risk for falls utilized a Markov decision analysis addressing this therapeutic dilemma. The decision model by Man-Son-Hing et al.[43] was designed using prospective cohort studies to determine antithrombotic drug efficacy in AF. The model was applied to patients aged 65 years or older at risk for falling to evaluate the choice of warfarin, aspirin, or no therapy. Stroke probability and outcomes were determined from the AFI data, which were reanalyzed including only subjects 65 years of age and older, and average stroke risk was found to be 6% per year. Probability and outcomes of SDH and ICH were estimated from population studies, randomized controlled trials, clinic cohort studies, and pooled case studies for all 3 options: warfarin, aspirin, or no therapy. Probability of major bleeding was gathered from AFI for each treatment strategy. Falls data were gathered from several cohort studies. Due to a lack of data for determination of the etiologic role of falls and their outcomes, the assumption was made that the probability and outcomes of ischemic stroke were not affected by the occurrence of falls. Utilities were assigned to the states of mild, moderate, and severe stroke based on data from Gage et al.[55]

Although some assumptions were made in constructing this decision model, statistics from the best available literature were applied to the model where possible. The authors found that quality-adjusted life expectancy was greatest for warfarin > aspirin > no treatment ( Table 5 ).[43] This finding remained consistent when variable rates of SDH were used and variable rates of falls were used. The same conclusions were made for patients over 75 years of age and for those closer to age 65 years, with average annual stroke risks of 8% and 2%, respectively. Persons with an average fall risk "must fall approximately 300 times in 1 year for warfarin to not be the optimal therapy." Only when annual stroke risk was less than 2% did aspirin become the preferred therapy. When stroke risk was less than 1.2% per year, no therapy was preferred. Data for risk of less than 2% and less than 1.2% were unavailable.

The findings of this Markov model study are in agreement with conclusions of the subgroup analysis of Medicare beneficiaries prone to falls.[54] The anticoagulant therapy utilities, with corresponding stroke rates, were quite comparable to the recommendations of the ACCP and ACC/AHA/ESC guidelines.[2,5] Additionally, another Markov model, which used a greater efficacy rate for aspirin and was published prior to the AFI trial, had similar findings.[56] While the Man-Son-Hing et al.[43] Markov model study predicted risk associated with fall-related SDH, it is limited in that it did not account for other significant injury and morbidity associated with falls. Also, it did not include clinically important factors, such as warfarin drug interactions or improper monitoring, that are involved in bleeding risk associated with anticoagulation-related management.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: