Danish Registry Study Shows Potential of Stroke Risk Score in Heart Failure Patients

LONDON, England—In patients with heart failure with or without A-fib, the CHA2DS2-VASc score may help predict the risk of ischemic stroke, thromboembolism, and death, according to results from an observational study presented August 30, 2015, at the European Society of Cardiology Congress and simultaneously published in the Journal of the American Medical Association.  

Take Home: Danish Registry Study Shows Potential of Stroke Risk Score in Heart Failure Patients

Line Melgaard, MSc, of Aalborg University (Aalborg, Denmark), presented data from 3 Danish registries on 42,987 patients 50 years or older who had heart failure but were not on anticoagulant therapy. Of these, 9,395 (21.9%) had A-fib at baseline.

Over a median follow-up of 1.8 years, the absolute and cumulative risks of ischemic stroke increased along with number of comorbidities, as demonstrated by higher CHA2DS2-VASc scores. Attenuation of incidence rates after 5 years suggested that most stroke events occurred relatively shortly after diagnosis of heart failure. 

Furthermore, the absolute 1-year risk of thromboembolism in those with high CHA2DS2-VASc scores was increased in patients without vs with A-fib (P < .001 for interaction):

  • Score of 4—9.7% vs 8.2%  
  • Score of 5—11.9% vs 11.2%  
  • Score of 6—18.0% vs 14.9% 

The score showed modest predictive ability in patients with and without A-fib for the endpoints of ischemic stroke, thromboembolism, and death at 1- and 5-year follow-up (table 1).

Table 1. C Statistics for Predictive Ability of CHA2DS2-VASc

Additionally, at a cutoff value of 1, the CHA2DS2-VASc score yielded a moderately high negative predictive value (91.7%) for identifying patients without A-fib at low risk of stroke at 1 year.  

Utility for Patients in Sinus Rhythm

“The CHA2DS2-VASc score predicts stroke among heart failure patients without atrial fibrillation with comparable accuracy as in atrial fibrillation,” Ms. Melgaard said. “However, in this high mortality population, consideration of competing risks is highly relevant. The discriminatory performance was modest, but the clinical utility of the CHA2DS2-VASc score in patients with heart failure remains to be determined.”

Ms. Melgaard noted that her group was unable to distinguish between heart failure with preserved and reduced ejection fraction or estimate functional classification. Additionally, she said they cannot exclude the possibility that some patients may have had undiagnosed A-fib and cannot certain that the population studied is generalizable to more diverse heart failure populations.

Importantly, Ms. Melgaard said an additional sensitivity analysis to assess patients who developed A-fib after their heart failure diagnosis found no difference to the main analysis.

Sigrun Halvorsen, MD, of Oslo University Hospital (Oslo, Norway), commented that the study contributes to “an important field where there is a lack of knowledge” and asked how information obtained from CHA2DS2-VASc can inform the course of treatment in patients with heart failure without A-fib.

Ms. Melgaard responded that the score could potentially be used to identify patients who are not obvious candidates for anticoagulation therapy. The paper describes one such subset of heart failure patients as those with no A-fib and 2 or more components of the CHA2DS2-VASc score. 

Melgaard L, Gorst-Rasmussen A, Lane DA, et al. Assessment of the CHA2DS2-VASc score in predicting ischemic stroke, thromboembolism, and death in patients with heart failure with and without atrial fibrillation. JAMA. 2015;Epub ahead of print. 


  • Ms. Melgaard and Dr. Halvorsen report no relevant conflicts of interest. 

Related Stories: