Predicting risk of cardioembolic stroke
ROCHESTER, MINNESOTA. About 50% of strokes occurring in atrial fibrillation (AF) patients are cardioembolic in nature and usually related to thrombus (blood clot) formation in the left atrial appendage (LAA). Such strokes are best prevented by anticoagulation with warfarin. NOTE: In my personal opinion, nattokinase would likely be an even better choice for preventing cardioembolic strokes. The other 50% of strokes are either thrombotic (most often involving rupture of atherosclerotic plaque) or hemorrhagic in nature. The optimum way of preventing thrombotic stroke is through the use of antiplatelet agents such as aspirin or clopidogrel. Thus, it is important to know if a patient is at increased risk of cardioembolic stroke and thus may need warfarin, or has no increased risk in which antiplatelet therapy may be a better choice.

The presence of thrombi in the LAA can be established with reasonable accuracy by doing a transesophageal echocardiography (TEE) in which the ultrasound probe is placed in the esophagus rather than on the chest as is the case in standard transthoracic echocardiography. TEE will show the presence of existing thrombi and will also detect spontaneous echocardiographic contrast (SEC) – a pattern of "smoke-like", slow-swirling, echodensities in the LAA or left atrium. SEC is believed to be the genesis of thrombi.

Electrophysiologists at the Mayo Clinic recently reported a study aimed at developing a scoring system which would predict the likelihood of finding thrombi in the LAA. Their study involved 110 patients with nonvalvular AF (not on warfarin) in whom thrombi had been detected in the LAA (cases) and 387 patients with nonvalvular AF (not on warfarin) in whom no thrombi had been detected during TEE (controls). Statistically significant differences between the two groups included the following:
The average CHADS2 score (a commonly used indicator of stroke risk) was also significantly higher among cases (mean = 2.8) than among controls (mean = 1.6). The incidence of LAA thrombi was not related to age, gender or the presence of hypertension. Thus, while age above 75 years is considered a risk factor in the CHADS2 score, there was no indication that advanced age is associated with an increased risk of LAA thrombi. Based on their findings, the Mayo researchers propose a new algorithm for predicting the presence of LAA thrombi and commensurate risk of cardioembolic stroke or TIA.

A prior stroke or TIA, permanent AF, diabetes, and AF duration longer than 48 hours would each be allocated 1 point, while the presence of SEC, congestive heart failure, and AF duration longer than one year would be allocated 2 points each. Therefore, LAA thrombi risk would be graded from 0 to 10 by this new system which has yet to receive a catchy acronym. A separate analysis showed the presence of SEC and congestive heart failure to be, by far, the most predictive of LAA thrombi (odds ratios of 9.68 and 5.12 respectively). The Mayo researchers point out that prior research has shown that if TEE detects no thrombi in the LAA the risk of a cardioembolic stroke is pretty close to zero.
Wysokinski, WE, et al. Predicting left atrial thrombi in atrial fibrillation. American Heart Journal, Vol. 159, April 2010, pp. 665-71