Background: Rheumatic mitral stenosis is one of the important health problems, especially in Thailand. Thromboembolic complications from this disease remain the major problem in these patients. These complications are usually related to the left atrial thrombus formation.

Objective: To determine the parameters that can predict the presence of left atrial thrombus in these patients.

Material and Method: Two hundred and sixty Thai patients with mitral stenosis from Siriraj Hospital were prospectively recruited in the study. The baseline clinical characteristics of these patients which were related to thrombus formation were properly collected. All patients underwent the transthoracic and transesophageal echocardiography with the standard technique to detect the mitral valve area, mitral valve score, left atrial size, LV ejection fraction, right ventricular systolic pressure, spontaneous echo contrast (SEC) and associated valvular lesions. Transesophageal echocardiography was used as the gold standard for evaluation of left atrial thrombus. These parameters were analyzed to demonstrate association with the presence of left atrial thrombus by univariate and multiple logistic regression analysis. Equation with score for prediction of left atrial thrombus was also purposed.

Results: There were 77 men and 183 women in the present study. Left atrial thrombus was detected in 26% and previous thromboembolic complications occurred in 16.5%. Atrial fibrillation presented in 52.5%. From univariate analysis, atrial fibrillation (81.2% vs 42.9%, p<0.001), male sex (37.7% vs 25.6%, p<0.050), left atrial enlargement by electrocardiogram (45.5% vs 15.4%, p<0.001), left atrial diameter (7.26 vs 6.97 cm, p <0.030), left atrial volume (132.7 vs 113.5 cm2, p<0.001), EF by Teich method (58.9 vs 62.5%, p<0.011), EF by MOD-bp (60.7 vs 64.6%, p<0.005), tricuspid regurgitation (46.4% vs 28.8%, p<0.008), mitral regurgitation (10.1 vs 3.1%, p<0.022), and RVSP (57.0 vs 49.7 mmHg, p<0.005) can predict the presence of left atrial thrombus formation with statistical significance. Using multiple logistic regression model, only atrial fibrillation (OR 5.95, 95% CI 1.21-29.3, p<0.02) and RVSP (OR 1.02, 95% CI 1.01-1.04, p<0.04) were independent predictors. The authors proposed score for predicting probability of left atrial thrombus formation that equal to -3.61 + 1.79 AF + 0.03 RVSP with AUC of 0.764. The best cut-off point for this score was - 1.49, which gave a sensitivity of 91%, specificity of 56%, PPV of 48%, and NPV of 90%.

Conclusion: Prevalence of thromboembolic complications and thrombus formation is high in patients with mitral stenosis in Thailand. From the present study, the predictors for left atrial thrombus formation in the patients with mitral stenosis were atrial fibrillation and RVSP. The model for predicting left trial thrombus
formation was also proposed with high sensitivity and NPV.

Keywords: Echocardiographic Parameters, Left Atrial Thrombus Formation, Predication, Mitral Stenosis