I have seen those situations many times before. That is one reason I mentioned AVA by cath is better than echo, but not perfect. You also may get significant variability of thermodilution CO measurements affecting AVA calculation. If you look at the original Hakki paper from Circulation 1981, comparing HAkki equation to Gorlin, the correlation with Gorlin was strong (r 0.96) whether you use mean or peak-to-peak gradient. ~25% of patients had mean gradient that is higher than peak to peak, as in your case, which translates into 0.1-0.2 cm2 difference in AVA calculation between the 2 ways of doing Hakki but also with the Gorlin calculation (which can be higher or lower than either one, it does not better correlate with either one). I suggest the following: 1)IF the patient is in AF, take the beat with highest peak-to-peak and the beat with the highest mean, don't use the computer averages 2)One reason for mean >peak-to-peak is delay in the aortic pressure transducing more than LV pressure (eg, longer tubing on the aortic catheter). In those cases, make sure to shift the aortic waveform leftward on the software to make the LV upslope bisect the aortic dip. Mean gradient may end up being lower 3)If mean is truly higher than peak-to-peak, and AVA calculations are discrepant, make sure to measure indexed AVA. Remember that AVA2000 in men or 1200 in women establishes severe AS in this context

Thanks a lot for the comprehensive answer Dr.Hanna 👍👍. I really learn a lot from your videos and book