Purpose: It is well known that the presence of a low flow state may yield to discordant grading of aortic stenosis (AS) severity: i.e. small aortic valve area (AVA<1.0 cm2) but low mean gradient (MG<40 mmHg). Furthermore, even in patients with normal transvalvular flow rate, the MG that theoretically corresponds to an AVA value of 1.0 cm2 is closer to 30-35 mmHg rather than to the 40 mmHg cut-off value proposed in the guidelines. It has thus been suggested to maintain the same cut-point for the MG but lower down the cut-off value of AVA for severe AS from 1.0 to 0.8 cm2. The objective of this study was to determine the optimal cut-point values of AVA and MG to predict outcomes in patients with AS. Methods: 928 consecutive patients (70±13 y.o., 59% men) with AS (54% severe) were included in this study. Receiver Operating Characteristic (ROC) curves were analyzed to determine the overall accuracy (area under the curve, AUC) and optimal cut-point values (i.e. providing the best balance between sensitivity and specificity) of MG, peak aortic jet velocity (Vpeak), and AVA to predict aortic valve replacement (AVR) and mortality at 1-year and 2-year follow-up. Results: During a mean follow-up of 5.6±3.8 years, 460 patients deceased, 538 underwent AVR, and 824 had a composite end-point of AVR or death. The AUCs and best cut-point values for prediction of AVR at 1- and 2-yr were for Vpeak: AUC=0.86, cut-point value 1-yr: 3.5 m/s, 2-yr: 3.4 m/s; for MG: AUC=0.86, cut-point value 1-yr: 31 mmHg, 2-yr: 26 mmHg; for AVA: AUC=0.83, cut-point value 1-yr: 1.01 cm2, 2-yr: 1.00 cm2. The analysis of the composite of AVR or death provided similar results: AUCs were slightly lower and cut-point values were the same as for the AVR endpoint. The ROC analysis of mortality however revealed that AVA but not MG (or Vpeak) was a significant predictor for this endpoint: 1-yr mortality: AUC: 0.65, cut-point value: 1.01 cm2; 2-yr mortality: AUC: 0.58, cut-point value: 0.98 cm2. Conclusion: The optimal cut-point values of AVA identified in this study to predict outcomes are very similar to those proposed in the guidelines, whereas those of MG and Vpeak are much lower than the guidelines criteria. Hence, as opposed to what was previously suggested, these findings would rather support the statu quo for the AVA cut-point value (1.0 cm2) but revision or cautious interpretation of the cut-points appears recommended for the gradient or velocity.