Cruciferous LAA Predicts Stroke in Otherwise-Low-Risk AF Patients

February 26, 2013

TOKYO — Adding cruciferous vegetables like cauliflower to one's diet would generally be considered a change for better health. But a left-atrial appendage (LAA) shape reminiscent of a cauliflower at computed tomography (CT) could be a cautionary sign, suggests a study of patients awaiting catheter ablation of nonvalvular atrial fibrillation (AF) [1].

Cauliflower: A main lobe <4 cm long without any forked lobes [Source: Dr Takehiro Kimura]

A "cauliflower" shape, one of four LAA morphologies at CT imaging with evocative names--the others were "cactus," "windsock," and "chicken wing"--independently added to stroke risk stratification in patients with a low risk as measured by CHADS2 score. That was true even after adjustment for the more discriminating CHA2DS2-VASc score.

LAA morphology information could potentially help decide the often-difficult question of whether such a patient should receive oral anticoagulation, conclude the authors, led by Dr Takehiro Kimura (Keio University School of Medicine, Tokyo, Japan).

The LAA is considered to be overwhelmingly the most common source of thrombi in cardiogenic stroke, and new procedures for closing the structure off from the rest of the left atrium, as an alternative or supplement to oral anticoagulation, are an active research area. For example, final results of the multicenter randomized PREVAIL trial of the percutaneously delivered Watchman (AtriTech/Boston Scientific) LAA closure device compared with long-term warfarin for stroke prevention are scheduled for presentation at the March American College of Cardiology 2013 Scientific Sessions.

Kimura et al's look at cardiac CT images from 80 candidates for AF ablation at one center from 2008 to 2011, 30 of whom had a history of stroke, was retrospective, with all the associated limitations, they acknowledge. Still, their findings suggest that "LAA anatomy might predict strokes in patients who cannot be discriminated by the CHADS2 score, left atrial size, or LAA flow velocity," they conclude in a report published online February 4, 2013 in Heart Rhythm.


Cactus: A main lobe <4 cm long and >2 lobes of >1 cm [Source: Dr Takehiro Kimura]

CT images of the LAA from the 30 patients with prior stroke were compared with those from the cohort's remaining 50 patients, who had been matched for age and sex. Their mean CHADS2 scores were not significantly different, at 0.8 and 0.6, respectively; nor did they differ in body-mass index, LVEF, left-atrial size, LAA flow, or levels of brain-type natriuretic peptide (BNP).

Volumes of the LAA were measured, and LAA were categorized by shape and internal dimensions. A cauliflower-shaped LAA, defined as "having a main lobe of less than 4 cm long without any forked lobes," was an independent predictor of a history of stroke.

Patients who had experienced stroke had higher CHA2DS2-VASc scores (p=0.015), but the stroke odds ratio (OR) for a cauliflower LAA was still 3.55 (95% CI 1.24–9.05, p=0.017) after adjustment for CHA2DS2-VASc.

That, said Kimura to heartwire by email, means that "it didn't matter what the CHA2DS2

Chicken wing: A main lobe >4 cm long with a folded angle <100 degrees [Source: Dr Takehiro Kimura]

The LAA morphology categories were developed as part of the evaluation for LAA closure procedures, Kimura observed. In theory, such visual LAA shape categorization could help stratify for cardiogenic stroke risk in nonvalvular-AF patients otherwise at low stroke risk, he said. To help with that, Kimura and his colleagues added LAA dimensions to the shape criteria already in place.

"Categorization based on the morphology might be enough; however, it relies on a subjective decision," he said. "By measuring the volume and length of LAAs, we avoid the risk of misclassification. This is the unique point of our paper."

Cardiac MRI might possibly be used instead of CT, he said, or perhaps three-dimensional echocardiography. But transthoracic echo wouldn't work, and "the evaluation of LAA morphology by transesophageal echo needs further investigation."

LAA Shapes "Worth Understanding"

Windsock: A main lobe >4 cm long with a folded angle >100 degrees [Source: Dr Takehiro Kimura]

According to Dr Sumit Verma (Heart Rhythm Center, Pensacola, FL) [2], "It is worth understanding the Classification of the LAA shapes. This is likely to impact the management of patients with AF in the near future." The LAA shape classification system, he writes in an accompanying editorial, "may help in determining the prescription of anticoagulants in the high-risk morphologies with low CHADS2 scores."

He cautions that further research is needed before the classifications could be applied to anticoagulation management. "Also, as we select patients for LAA occlusion using the Watchman device [AtriTech] or suture ligation using the LARIAT device [SentreHeart], consideration may be given to the LAA anatomy as a marker for stroke risk, as we weigh the risk-to-benefit ratio of those procedures."

Kimura et al report that they have no conflicts of interest. The Verma editorial had no disclosures.