Discussion
This is the first study to show that the LAA ostium reduces in size after AF ablation, and that this change is markedly asymmetrical, with the minor axis length shortening significantly more than the major axis length. In addition, we found that AF patients with HFrEF had a significantly greater LAA ostium areas than non-HFrEF patients. We also shown that LAA ostium size was positively correlated with both LA and LAA volume, and that patients with a “cauliflower” morphology of the LAA had significantly larger LAA ostia area than patients with other morphologies. Taking a functional perspective, the LAEF, as well as total and passive LA strains decreased as the LAA ostium size increased. LVEF increased significantly after ablation but changes in LVEF after ablation did not correlation with changes in LAA ostium area after ablation.
Of all the comorbidities analyzed, only HFrEF correlated significantly with LAA ostium size. This can be hypothesized to be due to stretching of the LAA ostium over time secondary to volume overload seen in heart failure(21). Both LA volume(22) and LAA volume(23) increases in patients with heart failure, and both LA enlargement(24) and LAA enlargement(25) is correlated with increased stroke risk. Another consideration originates from findings of the CASTLE-AF trial, which showed the superiority of ablation vs. medical therapy among AF patients with comorbid heart failure (22). The trial showed that the hazard ratio for stroke in these patients was 0.46 in favor of catheter ablation vs. medical therapy. This supports our hypothesis that by reducing the LAA ostial size, catheter ablation is also mitigating stroke risk in this patient group. Future investigation should further elucidate the underlying mechanisms for superiority of ablation when compared to medical therapy in the heart failure population.
When correlating the LAA ostium area to other morphological metrics of the LA and LAA, we found that the “cauliflower” LAA morphology was associated with the largest ostial area. Previous studies have found that “single-lobed”, “cauliflower” LAA morphology and larger LAA volume was associated with higher rates of stroke (26, 27). This finding would be supported by the results of our investigation, as a larger LAA ostium may confer a greater risk for thromboemboli generated in the LAA of passing into systemic circulation (28). In addition, others have reported that non- “chicken-wing” LAA morphologies were associated with higher rates of stroke compared to the “chicken-wing” morphology (29). However, contradictory findings have also been reported, stating that stroke rates were similar across patients with different LAA morphologies (30). Though most studies including ours used a two-expert consensus in labeling LAA morphology, the relatively subjective nature of this categorization demands for future revamping and improvement.
In addition to anatomical metrics, various LA functional metrics were also found to be significantly associated with LAA ostium area. Specifically, the LAA ostium area was negatively correlated with LAEF, total LA strain and passive LA strain. While LAEF is more commonly used as a surrogate for LA function (31), the three components of LA strain may show a clearer picture and more thoroughly breakdown the role of the LAA during the normal cardiac cycle and in AF. The total, passive, and active components of LA strain correspond to the reservoir, conduit and pumping functions of the LA, respectively (32). Our finding that the LAA ostium area only correlates with total and passive LA strain may imply that the LAA is mostly involved in the reservoir and conduit roles of the atrium, but not necessarily in its pumping function. This is supported by previous investigations that have characterized the LAA’s role as an overflow and decompression chamber for the LA (33). It is also unsurprising that the LAA does not contribute significantly to the LA pumping function, given its peripheral location, thin walls and relatively weak contractility (34).
Finally, our study is the first to characterize changes in the LAA ostium over time after AF ablation. Our hypothesis for the underlying mechanism includes an ablation-induced chronic remodeling of the left atrial substrate, which also includes changes to the LAA. The termination of AF and restoration of sinus rhythm may allow atrial myocytes to have more regular contractions and higher contractility(35). Additionally, although there was a near-significant negative correlation of pre-ablation LVEF and LAA ostium area, there was no direct correlation found between post-ablation change in LVEF and change in LAA ostium area. Furthermore, post-ablation change in LAEF also did not correlate with change in LAA ostium area. These findings suggest that the reduction in LAA ostium area seen after AF ablation is not a simple, direct result of functional improvement of the LA or LV. Rather, this reduction may be due to a more unique remodeling of the LAA in response to specifically catheter ablation therapy. Interestingly, previous studies have also found that a LAA ostium size <3.5 cm2 was associated with a reduced risk of stroke(36). Although we did not find a direct correlation of LAA ostium area with stroke in our study cohort, 29.7% of our patients did experience a significant reduction of LAA ostia from >3.5 cm2 to <3.5 cm2. A prospective, longitudinal study is needed to track stroke incidence in this group.
Finally, the reduction in LAA ostium area observed in our patient cohort was markedly asymmetrical, with significantly more shrinkage in the minor axis length than the major axis length of the centroid. Without a larger and more comprehensive study, the clinical implications of this finding remain unclear. Although with the rising popularity of LAA occlusion devices such as the Watchman™, it may be beneficial to take these post-ablation changes in the ostia into account when selecting the appropriate device size. It will also be intriguing to explore the underlying histology of LA myocytes surrounding the LAA ostium to determine whether distinct features are present in cells that align in the minor axis length direction. Lastly, it may be worth investigating patients whose LAA ostium does not shrink in size after ablation, as this cohort may exhibit an “ablation-refractory” phenotype and thus suffer from higher stroke risk despite ablation. For this patient cohort, LAA occlusion should be considered in addition to standard anticoagulation therapy to mitigate stroke risk (37).
Our findings bear important and compelling clinical implications that should be further explored in future investigations. Growing knowledge of the LAA and the LAA ostium is crucial to our overall understanding of AF pathophysiology and substrate stratification in AF management.