By Dr. Geoffrey Modest
The CHA2DS2-VASc score is perhaps the best metric for predicting thromboembolic complications in patients with atrial fibrillation. This study assessed this tool for a variety of clinical outcomes in patients with heart failure, both with and without atrial fibrillation (see doi:10.1001/jama.2015.10725).
Details:
- Danish registry study of 42,987 patients (all >50yo, mean age 75) with incident heart failure (HF), not on anticoagulation, of whom 21.9% had concomitant atrial fibrillation (afib), from 2000-2012
- Assessed relation between CHA2DS2-VASc score and ischemic stroke, thromboembolism (TE) and death within 1 year of HF diagnosis
Results:
- Patients without afib, risks of ischemic stroke was 3.1% (n=977), TE was 9.9% (n=3178), and death was 21.8% (n=6956), with stratification by CHA2DS2-VASc score (max=10)
- Ischemic stroke: by CHA2DS2-VASc score of 1 through 6, the one year absolute risks were:
- With afib: 4.5%, 3.7%, 3.2%, 4.3%, 5.6%, 8.4%
- Without afib: 1.5%, 1.5%, 2.0%, 3.0%, 3.7%, 7%
- All-cause death:
- With afib: 19.8%, 19.5%, 26.1%, 35.1%, 37.7%, 45.5%
- Without afib: 7.6%, 8.3%, 17.8%, 25.6%, 27.9%, 35.0%
- At CHA2DS2-VASc score>=4, absolute risk of TE was high regardless of presence of afib (e.g. for score of 4, 9.7% and 8.2% for those without and with afib)
- Ischemic stroke: by CHA2DS2-VASc score of 1 through 6, the one year absolute risks were:
- The negative predictive value for ischemic stroke at 1 year post HF diagnosis was 92% (91-93%) in those with afib and 91% (88-95%) in those without afib
So, this study found that those with HF and without afib are at high risk of ischemic stroke, TE and death; the CHA2DS2-VASc score was helpful in stratifying these patients and had a moderately high negative predictive value as determined by 1 year post HF diagnosis; and those with CHA2DS2-VASc score >=4 had high absolute risk of TE (and even higher in those without afib than those with afib, though it seems that they only excluded those on anticoagulation prior to the HF diagnosis). On subgroup analysis, there was no association between female sex and increased risk of ischemic stroke, in patients both with and without afib (actually, of the individual components of the CHA2DS2-VASc score as noted below, female sex was somewhat protective in the group without afib and was not associated with ischemic stroke in those with afib. So, there seems to be differences depending on the individual components of the CHA2DS2-VASc score.)
In general, in patients with afib, a stroke risk of >1%/yr is typically used as the cutpoint in identifying benefit from anticoagulation (i.e., tends to outweigh risks); in this Danish study the risk of ischemic stroke in those without afib was approx 1.5%/yr with CHA2DS2-VASc score>1. However, it is important to comment that it is not clear what the cutpoint should be in those without afib, though there are other studies showing that those with HF without afib are at increased risk of stroke and TE, and that these clinical events are decreased with warfarin therapy.
One clear concern is that this study does not have data on the LV ejection fraction (EF). Are the ones with terrible EFs the ones who get TE? And, does the CHA2DS2-VASc score, which it seems would correlate mostly with vascular risk, just pick out those with ischemic cardiomyopathy/low EF (i.e., are those with low EFs, who are more likely to have embolic events because of LV clots and/or stasis, being identified by the CHA2DS2-VASc score, and really just the EF itself is important??). There are some studies in the literature which suggest that those with definite HF (recent decompensation requiring hospitalization), that HF itself was a significant independent risk factor for stroke/systemic embolism irrespective of LV systolic function, with overall rate of stroke being 1.5-2.4%/year — perhaps related to the finding in those with HF without afib that there are higher levels of pro-coagulants and pro-inflammatory factors such as elevated b-thromboglobulin, thrombin-antithrombin III complexes, and D-dimers (see Clin Ther. 2014; 36: 1135-44). Other studies have also found the CHA2DS2-VASc score predicted clinical events even in patients without HF: in a 4.1 years study of 20,970 patients who were discharged with a diagnosis of acute coronary syndrome without known afib in a Canadian registry, 453 (2.2%) had a stroke or TIA with an annual incidence >=1% in those with CHA2DS2-VASc score >=4 (e.g., see Heart 2014: 100: 1524-30).
Another concern is that those with HF and high CHA2DS2-VASc score but without afib on initial evaluation may actually have intermittent afib leading to the adverse clinical events. For example, identifying those with intermittent afib by an event monitor might find those at high risk for TE, allowing for targeted anticoagulant therapy.
So, bottom line: HF is a bad disease with 45-60% 5-yr mortality. This Danish study is an observational one, with a limited database (not have ejection fraction, or know if the patient smoked, or drank alcohol, or….). It seems to me that given the high incidence of HF and high mortality, there really should be a randomized control study using anticoagulation vs not in those with HF and no evident afib. And, perhaps as part of this study, it would also be useful to utilize event monitors to identify those with HF and intermittent afib to see if they might be the patients who really benefit from anticoagulant therapy.
Here is the CHA2DS2-VASc scoring system:
