Abstract
Atrial fibrillation is a growing epidemic in the developed world and current methods of evaluating AF patients and therapies are limited. In this thesis, we developed a novel set of tools to measure and simulate global and regional atrial function, both in healthy and diseased hearts. We used these tools to discover regional heterogeneity in healthy atrial function: some parts of the atrium move more than others, some are mostly passive while others are more active. We applied the same wall motion analysis to AF patients and found changes in global and regional mechanics: the heart enlarges, relies more on active contraction to move blood, and deteriorates in regions that were originally high-functioning and passive. Since AF develops alongside a series of other cardiac disease, we used pressure-volume loops to quantify changes in global mechanics. P-V loops revealed how ventricular dysfunction (congestive heart failure, left ventricular hypertrophy) and atrial dysfunction (type and duration of AF, any prior ablations) alter P-V loop shape and disrupt normal function.
To better understand why function changes due to AF, we built a novel finite element model of the left atrium and varied five factors that normally change due to AF: size, shape, pressure, fibrosis, and conduction. We found that pressure and size have the largest effect on function while shape and conduction have a smaller effect. We were able to recreate over 80% of the observed changes in atrial function between healthy adults and AF patients by simulating three of the five factors: size, pressure, and fibrosis. We discovered that atrial factors could not recreate the increased reliance on active contraction that was a hallmark of the wall motion analysis. We used a coupled circuit model of ventricular function to test the influence of changes in LV properties and discovered that impaired LV relaxation during early diastole recreated the increased reliance on active contraction.
We adapted this finite element model to simulate common ablation procedures and found that more aggressive procedures (wide area circumferential ablation) impaired atrial function more than the conservative procedures (pulmonary vein isolation). We also discovered that changes in global function depended on both the amount and location of ablation scar, and found that regional function and stress increased the most in areas where scar was applied. We used P-V loop analysis to measure acute changes in post-ablation function and used the wall motion analysis to measure long-term changes in global and regional function. We found that atrial function is acutely depressed, but recovers within one-month post-ablation. At six-months, the atrium shrinks in size and becomes more passive. These changes outweighed the effects of simulated scarring, and represented a return to healthier function following recovery of sinus rhythm.
