ADVERTISEMENT
Immediate Improvement in Left Atrial Function After Transcatheter Aortic Valve Replacement on Doppler Echocardiography
Abstract: Background. More than 50% of embolic strokes occur >24 hours after transcatheter aortic valve replacement (TAVR) and are therefore not directly procedure related. We aimed to determine immediate changes in left atrial (LA) function after TAVR, that may alter short-term and long-term stroke risk after TAVR. Methods. Transesophageal and transthoracic echocardiography were performed before and immediately after TAVR to evaluate left atrial appendage (LAA) velocities and Doppler echocardiographic markers of LA function. Results. Eighty-five patients (mean age, 83.1 ± 7.5 years; 54% male) underwent TAVR. Patients in sinus rhythm (n = 57) during TAVR had immediate improvement of LA function, with an increase in A-wave velocity (92.3 ± 33.7 cm/s to 104.9 ± 34.6 cm/s), mitral inflow velocity time integral (VTI; 27.8 ± 3.6 to 29.6 ± 9.5), A-wave VTI (10.8 ± 4.1 to 12.8 ± 4.2), and lateral A´ velocity (8.8 ± 3.6 cm/s to 9.7 ± 3.6 cm/s; P<.05 for all), and a decrease in E/A ratio (1.2 ± 0.73 to 1.05 ± 0.59; P<.01) after TAVR. Low baseline LAA emptying velocities were common (53%), and velocities significantly increased after TAVR (mean change, +4.9 cm/s; P<.01). We identified several clinical and echocardiographic determinants of low LAA emptying velocity at baseline (<35 cm/s). Conclusions. There is immediate improvement in LA function and an increase in LAA velocities after TAVR. This improvement may benefit hemodynamics immediately after TAVR, but may also increase the short-term stroke risk (as recently shown in two independent studies) in patients with LAA thrombus or low emptying velocities. Therefore, evaluation of LA function and LAA thrombus may be warranted to identify patients at high risk for periprocedural stroke and guide the need for anticoagulation therapy.
J INVASIVE CARDIOL 2019;31(1):15-20. (Epub 2018 November 11).
Key words: Doppler echocardiography, left atrial function, stroke, transcatheter aortic valve replacement
Left atrial appendage (LAA) thrombus by preprocedural computer tomography or transesophageal echocardiogram (TEE) has been associated as a potential cause for periprocedural stroke in transcatheter aortic valve replacement (TAVR).1,2 Impairment of left atrial (LA) function occurs in various disease processes, including hypertension, hypertrophic cardiomyopathy, dilated cardiomyopathy, heart failure, atrial fibrillation, coronary artery disease, and, importantly, degenerative aortic stenosis (AS).3-9 The common denominator of these diseases is the development of left ventricular hypertrophy with systolic and (more commonly) diastolic dysfunction, which leads to increased LA pressure, LA remodeling, fibrosis, and frequently LA dysfunction.10
Relief of severe AS by surgical aortic valve replacement (SAVR) or TAVR has been shown to eventually improve LA function over time.11-15 The improvement of LA function after TAVR could be similar to the improvement after electrical cardioversion for atrial fibrillation, potentially leading to embolization of microthrombi (as seen with “sludge” or spontaneous echo contrast) or macrothrombi if present from the LAA. Two studies, including our experience, suggested that LAA thrombus may be an under-appreciated source of post-TAVR stroke. Changes in LA function have been shown to occur as early as 1 week and persist to at least 1 year after TAVR. Recent studies have shown improvement in LAA emptying velocities in the early period after TAVR,16,17 but have not evaluated both global LA and LAA contractile function.18,19 We therefore sought to determine the immediate effects of TAVR on LA and LAA function using Doppler echocardiography.
Methods
Patient population. The study sample included consecutive patients undergoing TAVR for severe valvular AS with TEE guidance between December 2014 and April 2015 at Cedars-Sinai Medical Center, a tertiary referral center in Los Angeles, California. These patients were evaluated by our heart team (comprising experienced cardiothoracic surgeons and interventional cardiologists) and were deemed to have severe AS that was inoperable or too high risk for SAVR. Exclusion criteria included the following: mitral stenosis (moderate or severe); prosthetic mitral valves; intraprocedure atrial pacing; change in atrial rhythm during the procedure; poor or inadequate intraprocedural TEE imaging quality; and hemodynamically significant intraprocedural adverse events (including cardiac arrest, left main coronary artery occlusion, and aortic rupture). A total of 85 patients were ultimately included in the analysis, with a focus on 57 patients who were in normal sinus rhythm at the time of TAVR.
Clinical information. Demographics and comorbidities were obtained for all patients from review of electronic medical records. Valves placed included: the Edwards Sapien XT valve (Edwards Lifesciences); (2) Edwards Sapien 3 valve (Edwards Lifesciences); (3) self-expandable CoreValve (Medtronic); (4) Direct Flow valve (Direct Flow Medical); and (5) Lotus valve (Boston Scientific), and were performed via transfemoral (n = 79), transapical (n = 1), transaortic (n = 3), or transsubclavian approach (n = 2) at the discretion of the operator. This study was approved by the Institutional Review Boards of Cedars-Sinai Medical Center with the need for individual consent waived.
Echocardiographic data. As part of patient routine clinical care, patients had a periprocedural TEE during TAVR by an experienced board-certified echocardiologist at the Cedars-Sinai non-invasive lab. TTEs were done before (the vast majority within 3 months prior to TAVR) and immediately after TAVR (within 24 hours).
Pre-TAVR TTEs were reviewed for LA area, left ventricular size, and presence of mitral regurgitation. Intraprocedural TEEs were reviewed for LAA emptying velocities (obtained at similar multiplane angles and pulse-wave depths before and after TAVR), presence of spontaneous echo-contrast, sludge (pre-clot), and thrombus. Pre-TAVR and post-TAVR TTEs were reviewed for mitral peak velocity of early filling (E), mitral peak velocity of late filling (A), early diastolic mitral annular velocity (e´), deceleration time (DT), mitral inflow velocity time integral (VTI), mitral A-wave VTI, lateral A´ velocity, aortic valve mean and peak gradients, left ventricular ejection fraction (LVEF), left ventricular outflow tract (LVOT), right ventricular systolic pressure (RVSP), and pulmonary vein flow dominance (systolic or diastolic).
Statistical analysis. All echo-derived variables before and after TAVR were compared using the paired t-test for continuous variables (expressed as mean ± standard deviation) and McNemar’s test for categorical variables (expressed as frequencies and percentages). Analyses were repeated in the following subgroups: LAA emptying velocity <35 cm/s, LVEF ≥50%, LVEF <50%, E/e´ ≥15, and E/e´ <15. Comparisons between those with LAA emptying velocity <35 cm/s and ≥35 cm/s were performed using the independent samples t-test for continuous variables (expressed as mean ± standard deviation) and Chi-square test for categorical variables. Multivariate logistic regression analysis was used to determine independent predictors of LAA emptying velocity <35 cm/s. A two-sided P-value of ≤.05 was used as the cut-off for statistical significance. Statistical analysis was performed using JMP version 12.0.1 (SAS).
Results
Patient characteristics are shown in Table 1. A total of 85 patients (mean age, 83.1 ± 7.5 years; 54.1% males) were included in this study. Of all patients, 72% had coronary artery disease, 85% had hypertension, 35% had diabetes, 34% had chronic kidney disease, and 35% had paroxysmal atrial fibrillation. A majority of patients (n = 57; 82%) were in sinus rhythm during TAVR. Mean overall LA area was 24 ± 7.3 cm2, left ventricular end-systolic diameter was 4.5 ± 0.7 cm, aortic valve area was 0.63 ± 0.16 cm2, aortic valve peak gradient was 82 ± 22 mm Hg, and aortic valve mean gradient was 48 ± 14 mm Hg. When comparing pre-TAVR and post-TAVR echocardiographic parameters in all patients, there was no significant change in LAA emptying velocity (38 ± 22 cm/s to 39 ± 19 cm/s; P=.60) and borderline 
decrease in RVSP (43 ± 16 mm Hg to 40 ± 15 mm Hg; P=.05). One-third (27%) had spontaneous LA echo contrast, 1.2% had LAA sludge (pre-clot), and 3.5% had LA thrombus. Comparisons of echocardiographic parameters before and after TAVR in those who were not in atrial fibrillation or flutter (n = 57) (Table 2) showed a significant increase in A-wave velocity (92 ± 34 cm/s to 105 ± 35 cm/s; P<.001) (Figure 1), mitral inflow velocity time integral (VTI; 28 ± 3.6 to 30 ± 9.5; P=.03), A-wave VTI (11 ± 4.1 to 13 ± 4.2; P<.001), lateral A´ velocity (8.8 ± 3.6 cm/s to 9.7 ± 3.6 cm/s; P=.03), LVOT-VTI (22 ± 5.1 to 25 ± 7.3; P=.02), and LVEF (60 ± 14% to 63 ± 14%; P=.03). There was a decrease in E/A ratio (1.2 ± 0.7 to 1.05 ± 0.6; P<.01) after TAVR.
In the subgroup with low LAA emptying velocities (n = 45), TAVR significantly increased LAA emptying velocities by a mean of 4.9 cm/s (23 ± 6.2 cm/s before TAVR to 28 ± 11 cm/s after TAVR; P<.01), increased A-wave velocity by a mean difference of +12.3 cm/s (85 ± 31 cm/s to 97 ± 34 cm/s; P=.01 [evaluated in 28 patients]), increased A-wave VTI by a mean difference of +2.8 (9.3 ± 4.5 to 12.1 ± 4.7; P<.001 [evaluated in 29 patients]), decreased E/A ratio by a mean difference of -0.3 (1.5 ± 0.8 to 1.2 ± 0.7; P<.01 [evaluated in 28 patients]), and increased LVOT-VTI by a mean difference of +2.0 (21 ± 5.5 to 23 ± 6.5; P=.02 [evaluated in 42 patients]). Further dividing this subgroup into those in normal sinus rhythm (n = 21), there was again a significant increase in A-wave velocity (78 ± 31 cm/s to 90 ± 36 cm/s; P=.04) and A-wave VTI (9.5 ± 4.4 to 12 ± 4.8; P=.02), with a decrease in E/A ratio (1.6 ± 0.9 to 1.4 ± 0.7; P=.03), and an increase in LVOT-VTI (21 ± 4.9 to 23 ± 5.6; P=.02).
In the preserved LVEF (≥50%) subgroup (n = 45) (Supplemental Table S1), A-wave velocity (96 ± 34 cm/s to 110 ± 36 cm/s; P<.001) and A-wave VTI (11 ± 3.8 to 13 ± 4.2; P<.001) increased significantly after TAVR, while E/A ratio decreased (1.1 ± 0.7 to 0.96 ± 0.6; P<.01), and E-wave deceleration time decreased (280 ± 120 ms to 240 ± 80 ms; P=.04). Otherwise, there were no significant changes in E-wave velocity, mitral VTI, lateral A´ velocity, LVOT-VTI, systolic pulmonary vein dominance, LVEF, or LAA emptying velocity. In the reduced LVEF (<50%) subgroup (n = 12), there was no statistically significant difference in any echocardiographic parameters except for LVEF, which increased significantly after TAVR (37 ± 7.3% to 49 ± 16%; P<.01).
In those with E/e´ ≥15, ie, Doppler evidence for increased left atrial pressure (n = 26) (Supplemental Table S2), there was a significant increase in A-wave velocity (100 ± 37 cm/s to 117 ± 36 cm/s; P<.001), 
mitral VTI (31 ± 9.7 to 34 ± 10; P=.04), and A-wave VTI (12 ± 4.2 to 14 ± 4.7; P<.01) after TAVR. There was also a decrease in E/A ratio (1.3 ± 0.8 to 1.1 ± 0.53; P<.01) after TAVR. Otherwise, there were no significant changes in echocardiographic parameters, including LVEF and LAA emptying velocity. In the E/e´ <15 subgroup (n = 30), there was only an increase in A-wave VTI (10 ± 4.0 to 12 ± 3.5; P=.02) after TAVR. Otherwise, there were no other significant changes in any of the echocardiographic parameters.
We compared those with LAA emptying velocity <35 cm/s vs ≥35 cm/s (Table 3) to determine predictors of low LAA emptying velocity (<35 cm/s). Those with LAA emptying velocity <35 cm/s were more likely than those with LAA emptying velocity ≥35 cm/to have history of coronary artery bypass surgery, history of paroxysmal atrial fibrillation, larger LA size, lower A-wave velocity, higher E-wave velocity, higher E/A 
ratio, lower lateral A´ velocity, and higher right ventricular systolic pressure (P<.05 for all). When adjusting for all univariate significant variables in a multivariate logistic regression model, history of paroxysmal atrial fibrillation was the strongest predictor (odds ratio, 6.8; 95% confidence interval, 1.55-29.6; P=.01) among others (Table 3).
Discussion
To the best of our knowledge, this is the first study demonstrating the immediate effects of TAVR on LA and LAA function. Our principal findings are the following: (1) there is an immediate improvement in LA function after TAVR; and (2) patients with low LAA velocities – a common finding in this elderly cohort – demonstrate a potentially clinically meaningful increase of LAA velocities. There are several implications for these findings. First, low LAA emptying velocities predispose to LAA thrombus or prethrombus, which could embolize with increasing LAA velocities and lead to stroke or other thromboembolic complications after TAVR independent of valve deployment (similar to increases in LAA velocities after electrical cardioversion of atrial fibrillation). LAA thrombus has been recently described as an under-appreciated source of stroke after TAVR. Second, immediate improvements of LA function after TAVR may improve cardiac output in patients with impaired diastolic or systolic left ventricular function, which may stabilize hemodynamics after TAVR, but also contribute to the oftentimes marked hypertensive response of patients after TAVR. Third, while improvement of LA function may increase the short-term risk of thromboembolism, it may reduce the long-term stroke risk. Although the TAVR field is moving away from periprocedural TEE guidance, in high-risk patients, TEE may identify LAA thrombus, which may warrant delay of TAVR and may also influence decisions on anticoagulant therapy before and after TAVR (in the case of the latter, when the postprocedural bleeding risk is deemed to be acceptable).
Previous studies have demonstrated improvement in LA contractile function after TAVR based on speckle tracking echocardiography. Spethmann et al studied the effects of TAVR on LA function.13,14 Both of these studies evaluated patients with severe AS undergoing TAVR and found improvement in atrial reservoir and conduit function 1 week and 1 year after TAVR. D’Ascenzi et al looked at change in atrial function at 40 days and 3 months post TAVR.15 They found improvement in both structure and function of the LA, as evidenced by decreased LA size and increased peak atrial longitudinal strain at both measured times. A separate study looked at similar parameters at 6 months post TAVR;12 in this study by D’Andrea et al, there was a significant reduction in LA volume index and increased LA longitudinal strain. These studies show consistent evidence of the reverse remodeling and cardiac dynamic changes that occur from relief of the increased afterload in severe AS.
Our study, which is based on more reproducible Doppler echocardiography, is the first to provide information on the immediate improvement of LA and LAA function. We demonstrate that the previously described improvements of LA function occur immediately after relief of AS post TAVR. In addition, there was a non-statistically significant trend for an increase in systolic pulmonary vein dominance after TAVR in most subgroups, likely reflecting an immediate reduction in LA pressure from improved LA transit. Interestingly, we found only a trend toward an improvement of the A-wave velocity in those with reduced LVEF, which may be a reflection of more advanced and partially irreversible LA cardiomyopathy or a spurious finding in this relatively small subgroup.
LAA emptying velocities observed in our cohort are fairly consistent with prior studies of patients with severe AS.16,17 We evaluated changes in LAA emptying velocities after TAVR, and found that there was a significant increase in those with already low (<35 cm/s) LAA velocities. We did not find any significant changes based on LVEF or diastolic dysfunction subgroups. This finding partially contrasts two previous studies that found significant increases in LAA emptying velocities immediately and 1 week after TAVR in these subgroups.16,17 The differences may be due to unknown intrinsic factors in the LAA, rather than inconsistencies in LA function improvement itself, given that the LAA and LA are anatomically and embryologically quite different, with poor functional correlation.18,19 Taken together, the cumulative evidence (including our study) suggests an overall increase in LAA emptying velocities after TAVR.
We found that history of paroxysmal atrial fibrillation was an independent predictor of low LAA emptying velocity (<35 cm/s). There were Doppler echocardiographic parameters predictive of low LAA emptying velocity, such as A-wave, E-wave, E/A ratio, and lateral A´ velocity, but these lost statistical significance when adjusting for atrial fibrillation, likely because LA dysfunction and atrial fibrillation go hand-in-hand.
Our findings are early evidence for a previously under-appreciated potential source of post-TAVR thromboembolic event, which arguably is its most devastating complication.20-22 Patients with severe AS are known to have low LAA emptying velocities and hence increased risk for stroke.23,24 In addition, our group and others have recently found LAA thrombus to be a potential source for periprocedural stroke.1,2 Our study suggests the immediate improvement in LA function after TAVR may be one explanation for a possible correlation between LAA thrombus and post-TAVR stroke. However, our findings are preliminary and should trigger additional larger studies, adequately powered to correlate changes in LA function with immediate stroke risk after TAVR. Our findings suggest that a preprocedural TEE may be indicated to assess the LA and LAA, at least in higher-risk (and older) patients studied here, as it can provide useful clinically important information that likely alters clinical management (ie, anticoagulation regimen). Further study of the predictors of low LAA emptying velocities may be useful to determine who is likely to benefit from TEE before or during TAVR.
Study limitations. The relatively small sample size and a single but high-volume institution limit our study results. However, the sample size is similar to prior studies on this topic and all data were meticulously collected in a prospective manner, under similar procedural conditions. The small number of strokes prohibits drawing any causal conclusions between LAA and LA function and incidence of thromboembolic complications. However, our preliminary findings inform the TAVR field and help to design larger studies evaluating LAA as a source for thromboembolism during or after TAVR.
Conclusion
In this elderly, high-surgical risk cohort, there was immediate improvement in LA function after TAVR based on Doppler echocardiography. In addition, severe AS was associated with low LAA emptying velocities, and low velocities increased after TAVR. Recovery of LA function in the setting of LAA thrombus or low emptying velocities may increase risk for thromboembolic events after TAVR; therefore, assessment of the LAA may influence periprocedural management prior to TAVR in such high-risk patients.
References
1. Chalian A, Phan D, Rader F, AlJilaihawi H, Makkar R, Siegel RJ. Left atrial appendage: a new potential source of stroke following transcatheter aortic valve replacement. Heart Circ. 2017;1:004.
2. Palmer S, Child N, de Belder MA, Muir DF, Williams P. Left atrial appendage thrombus in transcatheter aortic valve replacement: incidence, clinical impact, and the role of cardiac computed tomography. JACC Cardiovasc Interv. 2017;10:176-184.
3. O’Connor K, Magne J, Rosca M, Piérard LA, Lancellotti P. Impact of aortic valve stenosis on left atrial phasic function. Am J Cardiol. 2010;106:1157-1162.
4. Badran HM, Faheem N, Elnoamany MF, Kenawy A, Yacoub M. Characterization of left atrial mechanics in hypertrophic cardiomyopathy and essential hypertension using vector velocity imaging. Echocardiography. 2015;32:1527-1538. Epub 2015 Jan 20.
5. Triposkiadis F, Moyssakis I, Hadjinikolaou L, et al. Left atrial systolic function is depressed in idiopathic and preserved in ischemic dilated cardiomyopathy. Eur J Clin Invest. 1999;29:905-912.
6. Yu CM, Fang F, Zhang Q, et al. Improvement of atrial function and atrial reverse remodeling after cardiac resynchronization therapy for heart failure. J Am Coll Cardiol. 2007;50:778-785.
7. Teo SG, Yang H, Chai P, Yeo TC. Impact of left ventricular diastolic dysfunction on left atrial volume and function: a volumetric analysis. Eur J Echocardiogr. 2010;11:38-43. Epub 2009 Oct 13.
8. Choi JI, Park SM, Park JS, et al. Changes in left atrial structure and function after catheter ablation and electrical cardioversion for atrial fibrillation. Circ J. 2008;72:2051-2057.
9. Yu CM, Fung JW, Zhang Q, et al. Tissue Doppler echocardiographic evidence of atrial mechanical dysfunction in coronary artery disease. Int J Cardiol. 2005;105:178-185.
10. Elahi MM, Chuang A, Ewing MJ, Choi CH, Grant PW, Matata BM. One problem two issues! Left ventricular systolic and diastolic dysfunction in aortic stenosis. Ann Transl Med. 2014;2:10.
11. Lisi M, Henein MY, Cameli M, et al. Severity of aortic stenosis predicts early post-operative normalization of left atrial size and function detected by myocardial strain. Int J Cardiol. 2013;167:1450-1455. Epub 2012 May 4.
12. D’Andrea A, Padalino R, Cocchia R, et al. Effects of transcatheter aortic valve implantation on left ventricular and left atrial morphology and function. Echocardiography. 2015;32:928-936. Epub 2014 Oct 17.
13. Spethmann S, Dreger H, Baldenhofer G, et al. Short-term effects of transcatheter aortic valve implantation on left atrial mechanics and left ventricular diastolic function. J Am Soc Echocardiogr. 2013;26:64-71.e2. Epub 2012 Nov 8.
14. Spethmann S, Baldenhofer G, Dreger H, et al. Recovery of left ventricular and left atrial mechanics in various entities of aortic stenosis 12 months after TAVI. Eur Heart J Cardiovasc Imaging. 2014;15:389-398. Epub 2013 Sep 12.
15. D’Ascenzi F, Cameli M, Henein M, et al. Left atrial remodelling in patients undergoing transcatheter aortic valve implantation: a speckle-tracking prospective, longitudinal study. Int J Cardiovasc Imaging. 2013;29:1717-1724. Epub 2013 Jul 14.
16. Aslan S, Gul M, Cakmak HA, et al. Short-term effects of transcatheter aortic valve implantation on left atrial appendage function. Cardiol J. 2015;22:527.
17. Ando T, Holmes AA, Slovut DP, Taub CC. Impact of transcatheter aortic valve implantation on left atrial appendage flow velocities. J Clin Ultrasound. 2016;44:375-382. Epub 2016 Feb 15.
18. Agmon Y, Khandheria BK, Meissner I, et al. Are left atrial appendage flow velocities adequate surrogates of global left atrial function? A population-based transthoracic and transesophageal echocardiographic study. J Am Soc Echocardiogr. 2002;15:433-440.
19. Blume GG, Mcleod CJ, Barnes ME, et al. Left atrial function: physiology, assessment, and clinical implications. Eur J Echocardiogr. 2011;12:421-430. Epub 2011 May 12.
20. Leon MB, Smith CR, Mack M, et al; PARTNER Trial Investigators. Transcatheter aortic-valve implantation for aortic stenosis in patients who cannot undergo surgery. N Engl J Med. 2010;363:1597-1607. Epub 2010 Sep 22.
21. Adams DH, Popma JJ, Reardon MJ, et al; U.S. CoreValve Clinical Investigators. Transcatheter aortic-valve replacement with a self-expanding prosthesis. N Engl J Med. 2014;370:1790-1798. Epub 2014 Mar 29.
22. Hassell ME, Nijveldt R, Roos YB, et al. Silent cerebral infarcts associated with cardiac disease and procedures. Nat Rev Cardiol. 2013;10:696-706. Epub 2013 Oct 29.
23. Petty GW, Khandheria BK, Whisnant JP, Sicks JD, O’Fallon WM, Wiebers DO. Predictors of cerebrovascular events and death among patients with valvular heart disease: a population-based study. Stroke. 2000;31:2628-2635.
24. Goldman ME, Pearce LA, Hart RG, et al. Pathophysiologic correlates of thromboembolism in nonvalvular atrial fibrillation: I. Reduced flow velocity in the left atrial appendage (The Stroke Prevention in Atrial Fibrillation [SPAF-III] study). J Am Soc Echocardiogr. 1999;12:1080-1087.
From the 1Department of Medicine, Cedars-Sinai Medical Center, Los Angeles, California; 2Kaiser Permanente, Panorama City Medical Center, Panorama City, California; 3Heart Institute, Cedars-Sinai Medical Center, Los Angeles, California.
Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Makkar reports grant support from Abbott and Edwards Lifesciences; employment income from Cedars-Sinai Medical Center; consultant income from Cordis and Medtronic. The remaining authors report no conflicts of interest regarding the content herein.
Manuscript submitted August 6, 2018, final version accepted August 19, 2018.
Address for correspondence: Florian Rader, MD, MSc, 127 S. San Vicente Blvd, Suite A3408, Los Angeles, CA 90048. Email: florian.rader@cshs.org
