IJCS | Volume 32, Nº5, September/October 2019

469 Linhares et al. Transesophageal echocardiography and CHA2DS2-Vasc score Int J Cardiovasc Sci. 2019;32(5):460-470 Original Article resonance imaging and computed tomography, classified 4 different types of LAA and correlated each type with risk of stroke. We evaluated 932 patients who were candidates to ablation of AF refractory to clinical treatment. Computed tomography was done to 499 patients and magnetic resonance imaging was done in 433 patients, and all were screened for history of stroke/ TIA. Each LAA morphology was categorized into one of the following four types: chicken wing, windsock, cauliflower and cactus. The results revealed that patients with the non-“chicken wing” morphology presented three times higher prevalence of stroke/TIA than those with chicken wing morphology (12% x 4%). Compared to the chicken wing morphology, those patients with cactus, windsock and cauliflower appendages were at risk increased 4, 5 and up to 8 times, respectively. 3 During our routine, it was found that the chicken wing type was easily identified by transesophageal echocardiography and although it was difficult to identify morphology alone in other morphological types proposed by Di Biase, they had multiple internal lobulations in common. Therefore, based on the studies with magnetic resonance imaging, considering the TEE imaging limitations, we propose the classification in non- thrombogenic, similar to the chicken wing morphology (smooth internal surface with just a large angle, similar to a chicken wing) and thrombogenic (irregular internal surface with several lobulations, characteristics that are present and common to the windsock, cactus and cauliflower types). In the analysis of these 2 groups classified as thrombogenic and non-thrombogenic, a higher incidence of the non-thrombogenic type was found, as in the study published by Di Biase et al., 3 . No correlation was found with echocardiographic data of presence of thrombus or spontaneous contrast, but there was an association with clinical history of stroke/TIA (25.2% in the thrombogenic group x 11.2% in the non- thrombogenic group, p = 0.005). Besides, it was clear that individuals with a CHA 2 DS 2 VASc score between 0 and 1, almost half had a thrombogenic morphological aspect. This finding underscores the importance of evaluating the thrombogenic risk profile and discussing early anticoagulation in this population (see table 3). We believe that these additional datamay contribute to the evaluation and conduct of patientswho are candidates to anticoagulant therapy, especially in those with CHA 2 DS 2 -Vasc = 1, when doubts as to treatment persist. Limitations The data analyzed in this study were taken from different echocardiographic device models (GE- Vingmed Ultrasound, Vivid 7 and S6, Horten, Norway and ESAOTE My Lab 70, Florence, Italy), but there is no warning on any international guideline on the variation in the reference values of the parameters acquired in our analysis, considering different pieces of equipment. 17,18 Unfortunately, the patients in our study did not perform any nuclear magnetic resonance imaging or computed tomography for comparison purposes. Therefore, based on the models proposed by these methods, we suggest a classification using TEE: non- thrombogenic (similar to chicken wing, as it has a smooth internal surface with prominent angulation) and thrombogenic (irregular internal surface with no well-defined angulation, as seen in cauliflower, cactus andwind sock shapes). Given our results consistent with MRI, we believe that TEE helps characterizing LAAs with a higher risk of stroke/TIA, but comparative studies are required to confirm our hypothesis. Conclusions There is an association between the morphodynamic abnormalities evaluated using TEE and CHA 2 DS 2 - VASc score. The higher the CHA 2 DS 2 -VASc score, the more frequent are the abnormal echocardiographic findings, especially in patients with AF. There was a higher occurrence of stroke/TIA in patients with LAA morphology considered thrombogenic by TEE, even in those with sinus rhythm. Author contributions Conception and design of the research: Linhares RR, Moreira DAR. Acquisition of data: Linhares RR, Peixoto LB, Cruz AP, Garcia LP. Analysis and interpretation of the data: Linhares RR, Moreira DAR, Peixoto LB, Cruz AP, Garcia LP. Statistical analysis: Linhares RR, Peixoto LB, Cruz AP, Garcia LP. Writing of the manuscript: Linhares RR, Moreira DAR, Barretto RBM, Le Bihan DCS, Silva CES. Critical revision of the manuscript for intellectual content: Linhares RR, Moreira DAR, Barretto RBM, Le Bihan DCS, Silva CES.