By Michael H. Crawford, MD

Professor of Medicine, Chief of Clinical Cardiology, University of California, San Francisco

Dr. Crawford reports no financial relationships relevant to this field of study.

SYNOPSIS: An investigation of patients in a national database with atrial fibrillation, no comorbidities, and not on aspirin or anticoagulants showed that the clinical features that make up the CHA2DS2-VASc score change over time and can increase a patient’s risk for stroke, which could affect therapy decisions. Thus, the CHA2DS2-VASc score should be reassessed periodically and appropriate therapeutic actions taken promptly.

SOURCES: Chao TF, Lip GYH, Liu CJ, et al. Relationship of aging and incident comorbidities to stroke risk in patients with atrial fibrillation.
J Am Coll Cardiol 2018;71:122-132.

Gage BF. Stroke prediction rules in atrial fibrillation. J Am Coll Cardiol 2018;71:133-134.

The authors of previous studies of stroke risk in patients with atrial fibrillation (AF) have assessed the risk factors imbedded in the CHA2DS2-VASc score at baseline and then observed the patients during years of follow up. However, as these patients age, comorbidities that affect their risk often change.

Investigators from Taiwan hypothesized that changes in the CHA2DS2-VASc score would carry greater predictive value than the baseline score. From the Taiwan National Health Insurance database, 31,039 patients with AF who were not on antiplatelet or anticoagulant drugs and did not exhibit any features in the CHA2DS2-VASc scheme (except age and sex) were identified from 1996-2009. In this group, 4,103 experienced a stroke during follow-up. The follow-up CHA2DS2-VASc score was the highest measured before the occurrence of stroke, mortality, or the end of 2009. The difference between the baseline and follow-up CHA2DS2-VASc score (delta) and the slope of the score change were calculated.

During follow-up, the mean age increased from 64 to 68 years and the delta CHA2DS2-VASc was 1.02. The baseline, follow-up, and delta CHA2DS2-VASc scores were higher in those who experienced a stroke. About 52% of patients acquired a new comorbidity that affected the CHA2DS2-VASc score, most commonly hypertension (37%), heart failure (27%), and diabetes (13%). CHA2DS2-VASc remained unchanged in only 41% of patients. The delta CHA2DS2-VASc score predicted stroke (hazard ratio, 1.52; 95% confidence interval, 1.48-1.56; P < 0.001) and performed better than baseline or follow-up CHA2DS2-VASc. The area under the receiver operating curve (AUC) was 0.74 for delta CHA2DS2-VASc, compared to 0.58 for baseline and 0.73 for follow-up. Also, the slope of the delta CHA2DS2-VASc was higher in those with stroke compared to those without (0.58 vs. 0.42; P < 0.001).

The authors concluded that the CHA2DS2-VASc score is not static over time and most patients with AF develop one or more new comorbidities that affect the CHA2DS2-VASc score in addition to aging. This increment in the CHA2DS2-VASc score also was highly predictive of stroke.


In some ways, the results of this study are obvious. If one acquires risk factors for stroke over time, the risk of stroke increases. Since this has not been investigated or proven, this study is of interest. The main finding of the study is that the AUC of the follow-up CHA2DS2-VASc score is significantly higher than that of the baseline score (0.74 vs. 0.58). However, considering that at baseline all patients included in this study exhibited no comorbidities and registered CHA2DS2-VASc scores of 0 for men and 1 for women, this, too, is unsurprising. One would expect that as patients with AF age they would pick up comorbidities that affect the CHA2DS2-VASc score, as most subjects in this study did. What is perhaps the most interesting finding in this study is that soon after an increase in the CHA2DS2-VASc score, the risk of stroke is higher than later. The editorialist for this article suggested that this is because new comorbidities may not be well controlled early after diagnosis. He noted that the risk of stroke is high early after a transient ischemic attack but decreases over time. The same may be true for newly diagnosed hypertension, heart failure, or diabetes.

There are limitation to this study. It was an insurance database study, and all diagnoses were made by the patient’s own physicians. However, the authors noted that the accuracy of this approach has been validated previously in the Taiwan database. There are no laboratory data available, so the influence of biomarkers cannot be ascertained. Although they assessed the value of changes in the CHA2DS2-VASc score, the authors explicitly stated that they are not proposing a new measure of delta CHA2DS2-VASc. The authors only showed these data to bolster their point that changes in the CHA2DS2-VASc over time must be considered and therapeutic changes made as appropriate to reduce the rising risk of stroke.