Legend: Figure 2: Association Between Pre-Procedural Global Longitudinal Strain (GLS) and Reverse Remodleing(depicted as change in End Diastolic Volume, EDV)
When the linear functional form for GLS is used as a predictor of change in EDV, the result is not significant (p=0.8196). The regression coefficient for linear │pre-GLS│ is -0.3008 (-2.9301, 2.3285; p=0.8196). Assuming the same regression coefficient and the same variability in the data, a sample size of 641 observations would be smallest sample size that would result in a significant p-value, as it reduces the standard error just enough to obtain a Wald confidence interval with zero as an upper bound. In that case, the result would be -0.3008 (-0.6016, 0.0000; p=0.05).
All but two patients demonstrated a change of at least -1 in severity post-clip with improvement to moderate(+2) MR or better post-clip. The two patients that failed to improve were mixed etiology.  Mean pre-clip MR severity rating of 3.6 improved to post-clip mean rating of 1.2.  Of those with mean gradients reported, 9 patients had mean gradients >7mmHg post clip and the post-clip average mean gradient was 4.55mm Hg. Also, while most patients underwent elective procedures, five were urgent and one was emergent.