Abstract
Aims
The prognostic implication of left atrial (LA) dysfunction and left ventricular diastolic dysfunction (LVDD) in patients with coarctation of aorta (COA) is unknown. The purpose of this study was to determine whether LA dysfunction and LVDD were associated with mortality in COA patients.
Methods and results
This is a retrospective review of adults (age ≥18 years) with repaired COA that underwent transthoracic echocardiogram (2000–18). LVDD was determined using the 2016 guidelines for LV diastolic function assessment, and LA dysfunction was assessed using LA reservoir strain. Of 721 patients, LV diastolic function could be determined in 635 (88%); and 414 (65%) had no LVDD, while 146 (23%), 53 (8%), and 22 (4%) had Grade I/II/III LVDD, respectively. The mean LA reservoir strain was 39 ± 11%, and patients were divided into quartiles: top quartile (reference group), mild LA dysfunction, moderate LA dysfunction, and severe LA dysfunction. Grade III LVDD (but not Grades I and II) was associated with death/transplant. On the other hand, there was an incremental risk of death/transplant across LA strain quartiles: mild LA dysfunction [hazard ratio (HR) 1.16, 1.04–2.06], moderate LA dysfunction (HR 1.75, 1.27–3.58), and severe LA dysfunction (HR 3.49, 1.88–7.16). Of 86 patients with indeterminate diastolic function, there was a trend towards a lower 5-year transplant-free survival in patients with LA dysfunction vs. normal LA function (83% vs. 91%, P = 0.06).
Conclusion
LA dysfunction (but not LVDD) was associated with incremental risk of mortality and thus can be used for prognostication in all patients including those with indeterminate diastolic function.
Keywords: coarctation of aorta, left atrial dysfunction, left ventricular diastolic dysfunction
Introduction
Congenital heart disease is the most common cause of cardiovascular death in patients younger than 50 years of age, and most of the patients die from end-stage heart failure due to right or left ventricular systolic or diastolic dysfunction.1,2 Among patients with congenital heart disease, coarctation of aorta (COA) is a leading cause of left ventricular diastolic dysfunction (LVDD) and heart failure-related deaths.1–4 The cause LVDD in patients with COA is postulated to be due to LV pressure overload from hypertension, which in turn, leads to LV remodelling, fibrosis, and stiffness and the subsequent left atrial (LA) remodelling, atrial fibrillation, heart failure and death.5–8 LVDD typically begins in early childhood and can persist and progress even after successful COA repair.5–8 The early phase of LVDD is characterized by impaired myocardial relaxation, and as the disease progresses, there is a reduction in LV compliance and a compensatory increase in LV filling pressures.9 The LA modulates LV filling via its reservoir, conduit, and booster function and thus plays a central role in the onset and progression of heart failure symptoms due to LVDD.10,11
The diagnosis of LVDD requires an integration of multiple echocardiographic indices, and in clinical practice, determination of LVDD severity is based on guidelines proposed by the American Society of Echocardiography/European Association of Cardiovascular Imaging (ASE/EACVI).12 On the other hand, speckle tracking imaging of the LA provides a single robust metric for the diagnosis and monitoring of LA dysfunction.10,11,13,14 There are robust data demonstrating that LA strain and LV diastolic function indices can be used for monitoring of disease progression, response to therapy, and for prognostication in patients with acquired cardiovascular disease.15–20
Previous studies have described LA dysfunction and LVDD, and their relationship to symptoms in patients with COA.5–8,21 However, there are no systematic studies assessing whether LA strain and LV diastolic function indices can be used for prognostication in this population. The purpose of this study was to determine whether LA dysfunction and LVDD were associated with mortality in adults with COA. We hypothesized that LA and LV diastolic function indices will predict transplant-free survival in COA patients, and that LA dysfunction and LVDD will identify patients at high risk for mortality during follow-up. The rationale for this study is that demonstrating the prognostic role of LA and LV diastolic function indices based on data derived from the COA population (rather than data extrapolation from a different population) will strengthen the evidence for the use of these indices for clinical decision-making in this population.
Methods
Study population
We reviewed the Mayo Adult Congenital Heart Disease (MACHD) registry and identified adults (age ≥18 years) with repaired COA that underwent transthoracic echocardiogram between 1 January 2000 and 31 December 2018. From this cohort, we excluded patients with the following conditions: (i) atrial arrhythmias at the time of echocardiogram; (ii) ventricular pacing at the time of echocardiogram; (iii) significant mitral valve disease defined as a native mitral valve mean gradient >3 mmHg or >mild mitral regurgitation, severe mitral annular calcification based on qualitative assessment, or mitral valve prosthesis; (iv) incomplete data defined as inadequate images for LA strain analysis and/or incomplete diastolic function assessment defined as having <3 of the 4 primary LV diastolic function indices stipulated by the ASE/EACVI guidelines; and (v) patients with single ventricle palliation. Clinical variables and mortality were retrieved from the electronic health records. The Mayo Clinic Institutional Review Board approved the study.
Study objectives
The study objectives were (i) to determine whether conventional LVDD severity classification based on the ASE/EACVI guidelines was associated with death/transplant, and whether LVDD severity progressed unidirectionally during longitudinal follow-up. (ii) To determine whether LA dysfunction was associated with death/transplant and whether LA dysfunction severity progressed unidirectionally during longitudinal follow-up. Exploratory analysis was performed to determine whether LA strain analysis can be used for prognostication in patients with indeterminate diastolic function.
Assessment of left ventricular diastolic function
We selected the first comprehensive echocardiogram performed within the study period, and offline analyses of digital images were performed. The diagnosis of LVDD was based on the cut-off points for the four primary indices outlined in the ASE/EACVI guidelines (Supplementary data online, Figure S1)12: (i) septal tissue Doppler early velocity (e′) <7 cm/s or lateral e′ <10 cm/s; (ii) ratio of averaged mitral inflow pulsed wave Doppler early velocity to tissue Doppler early velocity (E/e′) >14 or septal E/e′ >15 or lateral E/e′ >13; (iii) LA volume index >34 mL/m2; and (iv) tricuspid regurgitation velocity >2.8 m/s. Based on the ASE/EACVI LVDD severity algorithm,12 we classified the patients into the following LV diastolic function subgroups: (i) normal diastolic function, (ii) Grade I diastolic dysfunction, (iii) Grade II diastolic dysfunction, (iv) Grade III diastolic dysfunction, and (v) indeterminate diastolic function.
Assessment of left atrial function
LA strain was assessed using speckle tracking strain imaging, obtained using Vivid E9 and E95 (General Electric Co, Fairfield, CT, USA) with M5S and M5Sc-D transducers (1.5–4.6 MHz) at frame rate of 40–80 Hz, and these images were exported (DICOM) and then analysed offline using TomTec (TomTec Imaging Systems, Unterschleissheim, Germany) as previously described.22 The offline assessment of LA reservoir strain, conduit strain, and booster strain was assessed as by two experienced research sonographers previously described (Supplementary data online, Figure S2). We chose LA reservoir strain as the primary metric of LA function based on previous data demonstrating superiority of LA reserve strain for prognostication when compared with LA conduit and booster strain.15 We defined LA dysfunction as LA reservoir strain <25%.21,23 In order to assess the relationship between LA dysfunction severity and mortality, we compared the risk of death/transplant across different quartiles of LA reservoir strain.
Progression of LA dysfunction and LVDD
A subgroup analysis was performed in patients who had two transthoracic echocardiograms at least 36 months apart without any surgical or transcatheter interventions between both echocardiograms. The purpose of this analysis was to determine whether LA dysfunction and LVDD progressed unidirectionally during follow-up and to identify the risk factors for unidirectional progression of LA dysfunction and LVDD.
Statistical analysis
Between-group comparisons were performed using Fisher’s exact test, unpaired t-test, Analysis of variance, and Wilcoxon rank-sum test as appropriate. Time-to-event analyses were performed using Cox regression and Kaplan–Meier method. First, we created a univariate model to identify the risk factors for death/transplant using clinical and haemodynamic variables that have to be clinical relevant in the COA population.3,4,8 The variables with P < 0.25 on univariate analysis were then entered into the multivariate model using stepwise forward selection with P < 0.1 required for a variable to remain in the model. Because some of the patients (n = 94) underwent surgical or transcatheter intervention during follow-up at different times, these interventions were modelled as time-dependent covariates in the proportional hazards regression models. A P-value <0.05 was considered statistically significant. All statistical analyses were performed with John’s Macintosh Project (JMP) and statistical analysis system software (versions 14.1 and 9.4, respectively; SAS Institute Inc., Cary, NC, USA).
Results
Diastolic function
Of 721 COA patients who met study inclusion criteria, LVDD severity grade could be determined in 635 (88%) patients while the other 86 (12%) were classified as indeterminate diastolic function (Supplementary data online, Figure S3). Of the 721 patients, 59 (8%) had LV ejection fraction <50%, and 46 (7%) had prior repair of ventricular septal defect. Table 1 shows the baseline clinical and echocardiographic characteristics of the 635 patients with LVDD severity grading. When compared with patients with normal diastolic function, those with LVDD were older, and more likely to have associated left heart obstructive lesion, comorbidities, LV hypertrophy, and LV systolic dysfunction.
Table 1.
Baseline characteristics based on LV diastolic function (n = 635)
| Normal DF (n = 414) | Grade I (n = 146) | Grade II (n = 53) | Grade III (n = 22) | P | |
|---|---|---|---|---|---|
| Age (years) | 27 (19–38) | 42 (29–54) | 44 (32–58) | 50 (33–61) | <0.001 |
| Male | 249 (60%) | 92 (63%) | 33 (62%) | 11 (50%) | 0.7 |
| Isolated COAa | 361 (87%) | 104 (71%) | 34 (64%) | 15 (68%) | <0.001 |
| Subaortic stenosis | 29 (8%) | 15 (10%) | 7 (13%) | 2 (9%) | 0.3 |
| Aortic valve diseaseb | 42 (10%) | 32 (22%) | 15 (28%) | 5 (23%) | <0.001 |
| Comorbidities | |||||
| Hypertension | 205 (50%) | 97 (66%) | 28 (53%) | 12 (55%) | 0.005 |
| Coronary artery disease | 11 (3%) | 13 (9%) | 10 (19%) | 3 (14%) | <0.001 |
| Diabetes | 9 (2%) | 13 (9%) | 9 (17%) | 3 (14%) | <0.001 |
| Atrial fibrillation | 11 (3%) | 13 (9%) | 5 (9%) | 8 (36%) | <0.001 |
| Medications | |||||
| Beta-blockers | 103 (25%) | 54 (37%) | 25 (47%) | 12 (55%) | <0.001 |
| Calcium channel blockers | 35 (9%) | 28 (19%) | 8 (15%) | 3 (14%) | 0.007 |
| RAAS antagonist | 91 (22%) | 53 (36%) | 26 (49%) | 10 (46%) | <0.001 |
| Loop diuretics | 22 (5%) | 11 (8%) | 18 (34%) | 8 (36%) | <0.001 |
| Laboratory data | |||||
| GFR (mL/min/1.73 m2) | 97 ± 19 | 90 ± 23 | 88 ± 16 | 81 ± 14 | 0.04 |
| NT-proBNP (pg/mL) | 76 (25––155) | 230 (68––352) | 477 (187––1213) | 899 (530––1953) | 0.008 |
| Diastolic function indices | |||||
| Mitral E velocity (m/s) | 1.0 ± 0.4 | 1.1 ± 0.4 | 1.3 ± 0.3 | 1.3 ± 0.4 | <0.001 |
| Mitral A velocity (m/s) | 0.6 ± 0.2 | 0.7 ± 0.3 | 0.9 ± 0.2 | 0.6 ± 0.3 | <0.001 |
| Mitral deceleration time (ms) | 185 ± 39 | 195 ± 45 | 204 ± 65 | 170 ± 39 | 0.003 |
| Septal e′ velocity (cm/s) | 11 ± 3 | 7 ± 2 | 5 ± 2 | 6 ± 3 | <0.011 |
| Lateral e′ velocity (cm/s) | 15 ± 4 | 9 ± 3 | 8 ± 3 | 7 ± 4 | <0.001 |
| Septal E/e′ | 9 ± 3 | 14 ± 3 | 21 ± 7 | 20 ± 6 | <0.001 |
| Lateral E/e′ | 7 ± 3 | 10 ± 4 | 18 ± 6 | 16 ± 5 | <0.001 |
| LA volume index (mL/m2) | 25 ± 6 | 30 ± 7 | 35 ± 9 | 59 ± 21 | <0.001 |
| TR velocity (m/s) | 2.4 ± 0.4 | 2.5 ± 0.4 | 2.8 ± 0.6 | 3.2 ± 0.7 | <0.001 |
| Strain indices | |||||
| LA reservoir strain (%) | 41 ± 11 | 34 ± 10 | 29 ± 13 | 17 ± 11 | <0.001 |
| LA conduit strain (%) | 26 ± 10 | 20 ± 11 | 16 ± 7 | 11 ± 8 | <0.001 |
| LA booster strain (%) | 13 ± 8 | 5 ± 9 | 13 ± 7 | 6 ± 5 | 0.009 |
| Other echo indices | |||||
| LV GLS (%) | 22 ± 3 | 21 ± 3 | 18 ± 3 | 17 ± 4 | <0.001 |
| LV ejection fraction (%) | 63 ± 6 | 63 ± 8 | 59 ± 12 | 63 ± 13 | 0.09 |
| LV mass index (g/m2) | 96 ± 26 | 106 ± 25 | 114 ± 38 | 125 ± 35 | <0.001 |
| Relative wall thickness | 0.39 ± 0.07 | 0.41 ± 0.08 | 0.41 ± 0.06 | 0.43 ± 0.06 | 0.02 |
| Aortic mean gradient(mmHg) | 8 (6–14) | 12 (7–27) | 18 (8–32) | 10 (6–18) | <0.001 |
| COA mean gradient (mmHg) | 14 (9–19) | 14 (9–20) | 13 (8–24) | 11 (4–15) | 0.6 |
A, late diastolic velocity; COA, coarctation of aorta; DF, diastolic function; E, early diastolic velocity; e′, mitral annular tissue Doppler early velocity; GFR: glomerular filtration rate; GLS, global longitudinal strain; LA, left atrium; LV, left ventricle; LA, left atrium; RAAS, renin–angiotensin–aldosterone system.
Patients with LA reservoir strain in the top quartile (LA reservoir strain >45%).
Aortic velocity >2 m/s and/or ≥mild aortic regurgitation.
Absence of other left heart obstructive lesions, such as aortic or subaortic stenosis.
Left atrial function (n = 635)
The mean LA reservoir strain was 39 ± 11%, and 83 (13%) patients had LA dysfunction (LA reservoir strain <25%). Based on the distribution of LA reservoir strain in our cohort, we divided the patients into four quartiles: (i) first quartile (reference group) defined as LA reservoir strain >45% and this represents the best LA function; (ii) second quartile (mild LA dysfunction) defined as LA reservoir strain 39 to 45%; (iii) third quartile (moderate LA dysfunction) defined as LA reservoir strain 29 to 38%; and (iv) fourth quartile (severe LA dysfunction) defined as <29%, and this represents the worst LA function. The baseline clinical and echocardiographic characteristics differed across the different LA function quartiles (Table 2). Compared with the patients with normal diastolic function, those with diastolic dysfunction were older, more likely to have left-sided valvular heart disease and left ventricular hypertrophy, and more likely to have comorbidities, such as hypertension, diabetes, and renal dysfunction.
Table 2.
Baseline characteristics based LA function (n = 635)
| No LAD (n = 159) | Mild LAD (n = 158) | Mod LAD (n = 159) | Severe LAD (n = 159) | P | |
|---|---|---|---|---|---|
| LA reservoir strain | >45% | 39–45% | 29–38% | <29% | |
| Age (years) | 29 (21–42) | 28 (19–42) | 31 (20–43) | 43 (27–54) | <0.001 |
| Male | 105 (66%) | 100 (63%) | 82 (52%) | 99 (62%) | 0.05 |
| Isolated COAa | 142 (89%) | 130 (82%) | 127 (80%) | 11 (72%) | <0.001 |
| Subaortic stenosis | 11 (7%) | 13 (8%) | 12 (8%) | 17 (11%) | 0.6 |
| Aortic valve diseaseb | 13 (8%) | 18 (16%) | 25 (16%) | 38 (24%) | <0.001 |
| Comorbidities | |||||
| Hypertension | 83 (52%) | 76 (48%) | 86 (54%) | 97 (61%) | 0.1 |
| Coronary artery disease | 2 (1%) | 7 (4%) | 11 (6%) | 17 (11%) | <0.001 |
| Diabetes | 2 (1%) | 10 (6%) | 3 (2%) | 19 (12%) | <0.001 |
| Atrial fibrillation | 4 (3%) | 18 (5%) | 6 (4%) | 19 (12%) | 0.002 |
| Medications | |||||
| Beta-blockers | 36 (24%) | 34 (22%) | 42 (26%) | 68 (42%) | <0.001 |
| Calcium channel blockers | 18 (11%) | 19 (12%) | 15 (9%) | 22 (14%) | 0.7 |
| RAAS antagonist | 91 (22%) | 53 (36%) | 26 (49%) | 10 (46%) | <0.001 |
| Loop diuretics | 28 (18%) | 37 (23%) | 32 (20%) | 28 (18%) | 0.05 |
| Laboratory data | |||||
| GFR (mL/min/1.73 m2) | 102 ± 22 | 94 ± 21 | 83 ± 16 | 76 ± 15 | <0.001 |
| NT-proBNP (pg/mL) | 81 (26–157) | 203 (62–344) | 504 (181–1183) | 704 (441–1998) | 0.01 |
| Diastolic function indices | |||||
| Mitral E velocity (m/s) | 1.0 ± 0.2 | 1.1 ± 0.4 | 1.1 ± 0.6 | 1.2 ± 0.4 | 0.004 |
| Mitral A velocity (m/s) | 0.6 ± 0.2 | 0.7 ± 0.3 | 0.7 ± 0.2 | 0.7 ± 0.3 | 0.1 |
| Mitral deceleration time (ms) | 192 ± 43 | 190 ± 40 | 194 ± 38 | 193 ± 52 | 0.8 |
| Septal e′ velocity (cm/s) | 11 ± 4 | 8 ± 2 | 8 ± 2 | 6 ± 3 | <0.011 |
| Lateral e′ velocity (cm/s) | 15 ± 4 | 13 ± 5 | 13 ± 4 | 9 ± 4 | <0.001 |
| Septal E/e′ | 9 ± 4 | 13 ± 3 | 11 ± 6 | 18 ± 6 | <0.001 |
| Lateral E/e′ | 7 ± 3 | 9 ± 4 | 10 ± 3 | 13 ± 7 | <0.001 |
| LA volume index (mL/m2) | 24 ± 7 | 26 ± 7 | 27 ± 8 | 38 ± 17 | <0.001 |
| TR velocity (m/s) | 2.4 ± 0.4 | 2.7 ± 0.3 | 2.8 ± 0.6 | 3.0 ± 0.7 | <0.001 |
| Other echo indices | |||||
| LV GLS (%) | −22 ± 3 | −22 ± 3 | −20 ± 3 | −18 ± 4 | <0.001 |
| LV ejection fraction (%) | 63 ± 6 | 63 ± 7 | 64 ± 5 | 62 ± 11 | 0.02 |
| LV mass index (g/m2) | 100 ± 28 | 99 ± 26 | 114 ± 38 | 116 ± 32 | <0.001 |
| Relative wall thickness | 0.39 ± 0.07 | 0.41 ± 0.09 | 0.42 ± 0.06 | 0.40 ± 0.02 | 0.1 |
| Aortic mean gradient (mmHg) | 8 (6–14) | 9 (6–19) | 10 (6–18) | 11 (7–25) | 0.03 |
| COA mean gradient (mmHg) | 14 (9–22) | 14 (9–20) | 12 (8–17) | 14 (9–20) | 0.2 |
A, late diastolic velocity; COA, coarctation of aorta; E, early diastolic velocity; e′, mitral annular tissue Doppler early velocity; GFR, glomerular filtration rate; GLS, global longitudinal strain; LA, left atrium; LV, left ventricle; LAD, left atrial dysfunction; LA, left atrium; RAAS, renin–angiotensin–aldosterone system.
Aortic velocity >2 m/s and/or ≥mild aortic regurgitation.
Absence of other left heart obstructive lesions, such as aortic or subaortic stenosis.
Outcomes (n = 635)
The median follow-up for these 635 patients was 7.9 (4.7–10.4) years, yielding a follow-up of 4891 patient-years. A total of 49 (8%) patients died during this period, and the causes of death were end-stage heart failure (n = 19), arrhythmic/sudden cardiac death (n = 7), stroke (n = 3), aortic dissection (n = 2), post-operative cardiac surgery (n = 3), sepsis/multisystem organ failure (n = 6), malignancy (n = 4), gastrointestinal bleeding (n = 3), and unknown/mixed (n = 2). The median age at the time of death was 54 (37–69) years. Four (0.6%) patients underwent heart transplant because of end-stage LV diastolic failure, and one of the patients who underwent heart transplant subsequently died during follow-up. Collectively, the endpoint of death/transplant occurred in 52 (8%) patients, yielding an event rate of 1.1% per year.
Overall, 221 (35%) patients had LVDD while 414 (65%) had normal diastolic function. The 10-year transplant-free survival was significantly lower in patients with LVDD when compared with patients with normal diastolic function (88% vs. 93%, log-rank P = 0.01), and was also different across the various LVDD severity grades (Figure 1). When compared with normal diastolic function, LVDD (all severity grades) was associated with a two-fold increase in death/transplant on univariate analysis [hazard ratio (HR) 2.32, 95% confidence interval (CI) 1.33–3.98, P = 0.003]. Table 3 and Supplementary data online, Table S1 show the univariate and multivariate models of the relationship between LVDD severity grades and mortality. On multivariate analysis, Grade III diastolic dysfunction (but not Grades I and II) was associated with mortality when compared with normal diastolic function (Table 3).
Figure 1.
Kaplan–Meier curves showing between-group comparison of transplant-free survival. Diastolic dysfunction was associated with lower transplant-free survival when compared with normal diastolic function (A/B). Similarly, left atrial (LA) dysfunction was associated with lower transplant-free survival when compared with normal LA function (C/D).
Table 3.
Cox model assessing prognostic relationship between diastolic dysfunction grade and mortality
| HR (95% CI) | P | |
|---|---|---|
| Univariate | ||
| Normal diastolic function | Reference | |
| Grade I diastolic dysfunction | 1.48 (0.71–2.93) | 0.3 |
| Grade II diastolic dysfunction | 3.01 (1.32–6.30) | 0.01 |
| Grade III diastolic dysfunction | 5.23 (2.19–11.2) | <0.001 |
| Multivariate | ||
| Normal diastolic function | Reference | |
| Grade I diastolic dysfunction | 0.93 (0.43–1.92) | 0.8 |
| Grade II diastolic dysfunction | 1.34 (0.47–4.02) | 0.2 |
| Grade III diastolic dysfunction | 1.91 (1.01–6.67) | 0.04 |
| Age, per 5 years increment | 1.16 (1.10–1.34) | <0.001 |
| Isolated COA | 0.87 (0.65–1.00) | 0.09 |
COA, coarctation of aorta; CI, confidence interval; HR, hazard ratio.
Multivariate model was created using stepwise forward selection of variables with P < 0.25 required for entry into the model and P < 0.1 required for a variable to remain in the model. The P values were derived from univariate analysis as shown in Supplementary data online, Table S1.
Overall, 72 (13%) patients had LA dysfunction while 473 (87%) had normal LA function. The 10-year transplant-free survival was significantly lower in patients with LA dysfunction when compared with patients with normal LA function (80% vs. 93%, log-rank P < 0.001) and was also different across quartiles of LA function subgroups Central illustration. When compared with normal LA function, patients with LA dysfunction had a three-fold increase in mortality on univariate analysis (HR 3.79, 95% CI 2.82–5.71, P < 0.001). The risk of mortality increased across the different quartiles of LA function after multivariate adjustment for confounder (Table 4).
Table 4.
Cox model assessing prognostic relationship between LA dysfunction severity and mortality
| HR (95% CI) | P | |
|---|---|---|
| Univariate | ||
| Reference group | Reference | |
| Mild LAD | 1.39 (1.12–2.71) | 0.008 |
| Moderate LAD | 2.18 (1.58–4.41) | <0.001 |
| Severe LAD | 4.16 (2.03–8.14) | <0.001 |
| Multivariate | ||
| Reference group | Reference | |
| Mild LAD | 1.16 (1.04–2.06) | 0.01 |
| Moderate LAD | 1.75 (1.27–3.58) | <0.001 |
| Severe LAD | 3.49 (1.88–7.16) | 0.001 |
| Age, per 5 years increment | 1.19 (1.08–1.37) | 0.002 |
| Isolated COA | 0.54 (0.29–1.02) | 0.07 |
The above LA function categorization was based on LA reservoir strain quartiles. (i) The first quartile (reference group) represents the best LA function, (ii) second quartile (mild LAD), (iii) third quartile (moderate LAD), and (iv) fourth quartile (severe LAD) represents the worst LAD function. Multivariate model was created using stepwise forward selection of variables with P < 0.25 required for entry into the model and P < 0.1 required for a variable to remain in the model. The P values were derived from univariate analysis as shown in Supplementary data online, Table S1.
COA, coarctation of aorta; CI, confidence interval; HR, hazard ratio; LA, left atrium; LAD, LA dysfunction.
Progression of LA dysfunction and LVDD
Of the 635 patients, 411 met the inclusion criteria for this subgroup analysis. At the time of baseline echocardiogram, 266 (65%) had normal LV diastolic function while 111 (27%), 31 (8%), and 3 (0.7%) had Grades I, II, and III LVDD, respectively. The median interval between the baseline and follow-up echocardiogram was 42 (38–47) months. The temporal change in LVDD function class was multi-directional with 27 (%) patients being re-classified to a higher severity class, 19 (5%) patients being re-classified to a lower severity class, and 6 (16%) patients being re-classified as indeterminate diastolic function (Figure 2).
Figure 2.
Flowchart showing left ventricular diastolic dysfunction (LVDD) progression. The blue arrow denotes patients who remained in the same LVDD severity grade at baseline and follow-up echocardiogram. The red and black arrows denote patients who were reclassified into a higher and lower LVDD severity grade, respectively. Altogether 66 patients were reclassified as indeterminate diastolic function class at the time of follow-up echocardiogram.
These 411 patients were classified based on LA function quartiles using the same cut-off points described above. At the time of follow-up echocardiogram, 375 (91%) remained in the same LA dysfunction class, while 36 (9%) patients were re-classified to a worse LA dysfunction category (Figure 3). The patients with LA dysfunction progression were older, had higher LV mass index and lower arterial compliance (Supplementary data online, Table S3). In contrast, the temporal change in LVDD severity was multi-directional, and there were no significant differences in demographic and haemodynamic characteristics between patients with progression vs. regression of LVDD.
Figure 3.
Flowchart showing left atrial dysfunction progression. The blue arrow denotes patients who remained in the same LA function quartile at baseline and follow-up echocardiogram, while the red arrows denote patients with LA dysfunction progression during follow-up. LA, left atrial.
Exploratory analyses
A total of 86 patients were classified as having indeterminate diastolic function (Supplementary data online, Figure S2). When compared with the study cohort (n = 635), those with indeterminate diastolic function (n = 86) were older, more likely to be females, and more likely to have associated left heart obstructive lesion and LA dysfunction (Supplementary data online, Table S2). Of the 86 patients, 27 (31%) had LA dysfunction. A diagnosis of LA dysfunction was associated with a lower 5-year transplant-free survival when compared with normal LA function (83% vs. 91%, P = 0.06), Figure 4C. Similarly, LA dysfunction was associated with a lower 5-year transplant among the 414 patients with normal diastolic function (86% vs. 98%, P < 0.001), and also among the 221 patients with LVDD (76% vs. 90%, P = 0.008), Figure 4A/B.
Figure 4.
Kaplan–Meier curves showing between-group comparison of transplant-free survival. LA dysfunction was associated with lower transplant-free survival among patients with normal diastolic function (A), diastolic dysfunction (B), and indeterminate diastolic function (C). LA, left atrial.
Additionally, we performed two other exploratory analyses to assess the impact of bicuspid aortic valve and aortic valve disease on outcomes. Of the 635 patients, 355 (56%) had bicuspid aortic valve. Compared with patients without bicuspid aortic valve, those with bicuspid aortic valve had similar LA reservoir strain 39 ± 7% vs. 38 ± 9%, P = 0.6, similar prevalence of LVDD (Grades I–III), 114 (34%) vs. 107 (36%), P = 0.1, and similar incidence of cardiovascular events, 0.9% vs. 1.1%, P = 0.2. Of the 635 patients, 94 (15%) had aortic valve disease defined as Aortic valve disease defined as aortic velocity >2 m/s and/or ≥mild aortic regurgitation. Compared with patients without aortic valve disease, those with aortic valve disease had similar LA reservoir strain 37 ± 10% vs. 39 ± 11%, P = 0.2, similar prevalence of LVDD (Grades I–III), 44 (39%) vs. 117 (34%), P = 0.08, and similar incidence of cardiovascular events, 1.0% vs. 1.1%, P = 0.5. There was no correlation between prior history of ventricular septal defect closure and LA function, LVDD, and cardiovascular events.
Discussion
Although previous studies have described LA dysfunction and LVDD in patients with COA, it is unknown whether LA and LV diastolic function indices can be used for prognostication in this population.5–8,21 The current study assessed the relationship between left heart diastolic function indices and transplant-free survival among 721 adults COA patients.
Left ventricular diastolic dysfunction
Of the 635 patients in whom LVDD severity could be determined, 146 (23%), 53 (8%), and 22 (4%) had Grades I, II, and III LVDD, respectively. However, only Grade III LVDD was associated with death/transplant, there was no difference in mortality between patients with Grades I and II LVDD when compared with those with normal LV diastolic function. These results suggest that the conventional LVDD severity classifications could not reliably identify COA patients at high risk for mortality. In contrast to our results, previous studies demonstrated the role of LVDD indices for prognostication in patients with acquired cardiovascular diseases.17 We speculate that the poor prognostic performance of the conventional LVDD severity classification in the current study may be related to differences in clinical and demographic characteristics of COA patients. For instance, the median age of our cohort was 32 years, while patients with LVDD due to acquired heart disease are typically older than 60 years of age.10,17,20 Since tissue Doppler indices are age-dependent with younger patients having higher e′ and lower E/e′ values,24 we speculate that the guideline criteria for abnormal tissue Doppler indices (septal <7 cm/s, lateral e′ <10 cm/s, and average E/e′ >14) may not be sensitive enough to detect LVDD in a young cohort, such as in the current study. Additionally, COA patients differ from patients with LVDD due to acquired heart disease, because they have a longer exposure to myocardial injury since LV pressure overload begins before birth and may persist and progress even after successful COA repair because of residual/recurrent obstruction or increased arterial stiffness.25,26
Left atrial dysfunction
Of the 635 patients in whom diastolic function category could be determined, 83 (13%) had LA dysfunction, and LA dysfunction diagnosis was associated with lower transplant-free survival. More importantly, we noted a consistent increase in mortality risk across the different quartiles of LA reservoir strain, ranging from 16% for mild LA dysfunction, 75% for moderate LA dysfunction, to more than two-fold increase in mortality risk for severe LA dysfunction when compared with a reference group of patients in the top quartile of LA strain (best LA function). Additionally, LA reservoir strain can be used for risk stratification in the subset of 86 patients with indeterminate diastolic function (12% of the cohort) and can also improve risk stratification in the patients with normal diastolic function and those with LVDD (Figure 4).
Our results are consistent with previous studies demonstrating a superior performance of LA strain (when compared with LVDD indices) for prognostication in patients with heart failure, atrial fibrillation, and valvular heart disease.10,13,15,16,20 The superior performance of LA strain imaging is likely due to the central role of the LA in modulation LV filling, pulmonary venous congestion, and heart failure symptoms in patients with left heart disease.11,27 LA strain imaging provides a comprehensive assessment of LA and LV function at different phases of the cardiac cycle, with LA reservoir strain assessing LA compliance, LA conduit strain assessing LV relaxation and chamber stiffness, and LA booster strain assessing intrinsic LA contractility and LV end-diastolic compliance.11
Longitudinal monitoring of left heart diastolic function
Longitudinal monitoring of disease progression using LVDD indices is often very challenging because these indices are affected by factors, such as volume status, heart rate, and LV afterload.10,11 As a result, it is sometimes unclear whether a progression in LVDD severity truly reflects disease progression vs. changes in loading condition. Although LA strain imaging is also affected by these haemodynamic factors, it is less load-dependent and has better correlation with haemodynamic and symptomatic deterioration.13,14 In our cohort, we observed a unidirectional progression of LA dysfunction, in contrast to the multi-directional trend of LVDD (Figure 3). The patients with LA dysfunction progression were older, had more LV hypertrophy and arterial stiffness, suggesting that perhaps interventions to reduce LV pressure overload and hypertrophy may potentially slow LA dysfunction progression.
Clinical implications and future directions
The robust prognostic performance of LA strain imaging observed in this study has several potential clinical applications. First, LA strain imaging can help identify patients at higher risk for mortality, and these high-risk patients may be benefit from more intensive medical and interventional therapies. Secondly, LA strain can be used for longitudinal monitoring, and a temporal deterioration in LA strain should be a ‘red flag’ for patients at risk of future adverse outcomes. Finally, LA strain can potentially be used for risk stratification in the subset of patients with indeterminate diastolic function, where the conventional LVDD indices may not apply.
While the current study supports the clinical application of LA strain imaging for risk stratification in COA patients, it is unknown whether proactive interventions based on LA strain will improve survival. Therefore, there is a need for further studies to determine whether intervention to reduce LV hypertrophy and pressure overload (i.e. surgical or transcatheter COA intervention or strict blood pressure control) will result in improvement of LA strain and a reduction in long-term mortality similar to studies in patients with acquired heart disease.19
Limitations
This is a retrospective study, and it is therefore prone to bias inherent in this study design. Although we adjusted for the potential effect of surgical and transcatheter intervention on mortality using robust statistical methodology, it is possible that the observed correlation between LA strain indices and mortality could have been influenced by other factors that were not adjusted for in our models.
Conclusion
LA dysfunction and LVDD are common in adults with COA, especially in older patients with LV hypertrophy and pressure overload. LA dysfunction was associated with incremental risk of mortality during follow-up and thus can be used for prognostication in all patients including those with indeterminate diastolic function. Rather it suggests that the conventional criteria for LVDD severity classifications may not be generalizable to every disease subgroup. To the best of our knowledge, this is the first study evaluating and comparing the prognostic role of LA dysfunction and LVDD in adults COA patients, and as a result, further studies are required to confirm the generalizability of our results and explore the means of integrating these indices in routine clinical practice.
Supplementary data
Supplementary data are available at European Heart Journal - Cardiovascular Imaging online.
Funding
A.C.E. was supported by National Heart, Lung, and Blood Institute (NHLBI) grant K23 HL141448-03. The MACHD Registry was supported by the Al-Bahar Research grant.
Conflict of interest: none declared.
Supplementary Material
References
- 1. Diller GP, Kempny A, Alonso-Gonzalez R, Swan L, Uebing A, Li W. et al. Survival prospects and circumstances of death in contemporary adult congenital heart disease patients under follow-up at a large tertiary centre. Circulation 2015;132:2118–25. [DOI] [PubMed] [Google Scholar]
- 2. Verheugt CL, Uiterwaal CS, van der Velde ET, Meijboom FJ, Pieper PG, van Dijk AP. et al. Mortality in adult congenital heart disease. Eur Heart J 2010;31:1220–9. [DOI] [PubMed] [Google Scholar]
- 3. Lee MGY, Babu-Narayan SV, Kempny A, Uebing A, Montanaro C, Shore DF. et al. Long-term mortality and cardiovascular burden for adult survivors of coarctation of the aorta. Heart 2019;105:1190–6. [DOI] [PubMed] [Google Scholar]
- 4. Choudhary P, Canniffe C, Jackson DJ, Tanous D, Walsh K, Celermajer DS.. Late outcomes in adults with coarctation of the aorta. Heart 2015;101:1190–5. [DOI] [PubMed] [Google Scholar]
- 5. Lombardi KC, Northrup V, McNamara RL, Sugeng L, Weismann CG.. Aortic stiffness and left ventricular diastolic function in children following early repair of aortic coarctation. Am J Cardiol 2013;112:1828–33. [DOI] [PubMed] [Google Scholar]
- 6. Florianczyk T, Werner B.. Assessment of left ventricular diastolic function in children after successful repair of aortic coarctation. Clin Res Cardiol 2011;100:493–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Egbe AC, Miranda WR, Connolly HM.. Increased prevalence of left ventricular diastolic dysfunction in adults with repaired coarctation of aorta. Int J Cardiol Heart Vasc 2020;28:100530. doi: 10.1016/j.ijcha.2020.100530. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Egbe AC, Qureshi MY, Connolly HM.. Determinants of left ventricular diastolic function and exertional symptoms in adults with coarctation of aorta. Circ Heart Fail 2020;13:e006651. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Krogmann ON, Rammos S, Jakob M, Corin WJ, Hess OM, Bourgeois M.. Left ventricular diastolic dysfunction late after coarctation repair in childhood: influence of left ventricular hypertrophy. J Am Coll Cardiol 1993;21:1454–60. [DOI] [PubMed] [Google Scholar]
- 10. Singh A, Addetia K, Maffessanti F, Mor-Avi V, Lang RM.. LA strain for categorization of LV diastolic dysfunction. JACC Cardiovasc Imaging 2017;10:735–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Thomas L, Marwick TH, Popescu BA, Donal E, Badano LP.. Left atrial structure and function, and left ventricular diastolic dysfunction: JACC state-of-the-art review. J Am Coll Cardiol 2019;73:1961–77. [DOI] [PubMed] [Google Scholar]
- 12. Nagueh SF, Smiseth OA, Appleton CP, Byrd BF III, Dokainish H, Edvardsen T. et al. Recommendations for the evaluation of left ventricular diastolic function by echocardiography: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr 2016;29:277–314. [DOI] [PubMed] [Google Scholar]
- 13. Singh A, Medvedofsky D, Mediratta A, Balaney B, Kruse E, Ciszek B. et al. Peak left atrial strain as a single measure for the non-invasive assessment of left ventricular filling pressures. Int J Cardiovasc Imaging 2019;35:23–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Telles F, Nanayakkara S, Evans S, Patel HC, Mariani JA, Vizi D. et al. Impaired left atrial strain predicts abnormal exercise haemodynamics in heart failure with preserved ejection fraction. Eur J Heart Fail 2019;21:495–505. [DOI] [PubMed] [Google Scholar]
- 15. Reddy YNV, Obokata M, Egbe A, Yang JH, Pislaru S, Lin G. et al. Left atrial strain and compliance in the diagnostic evaluation of heart failure with preserved ejection fraction. Eur J Heart Fail 2019;21:891–900. [DOI] [PubMed] [Google Scholar]
- 16. Reddy YNV, Obokata M, Verbrugge FH, Lin G, Borlaug BA.. Atrial dysfunction in patients with heart failure with preserved ejection fraction and atrial fibrillation. J Am Coll Cardiol 2020;76:1051–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Sato K, Grant ADM, Negishi K, Cremer PC, Negishi T, Kumar A. et al. Reliability of updated left ventricular diastolic function recommendations in predicting elevated left ventricular filling pressure and prognosis. Am Heart J 2017;189:28–39. [DOI] [PubMed] [Google Scholar]
- 18. Morris DA, Belyavskiy E, Aravind-Kumar R, Kropf M, Frydas A, Braunauer K. et al. Potential usefulness and clinical relevance of adding left atrial strain to left atrial volume index in the detection of left ventricular diastolic dysfunction. JACC Cardiovasc Imag 2018;11:1405–15. [DOI] [PubMed] [Google Scholar]
- 19. Huynh QL, Kalam K, Iannaccone A, Negishi K, Thomas L, Marwick TH.. Functional and anatomic responses of the left atrium to change in estimated left ventricular filling pressure. J Am Soc Echocardiogr 2015;28:1428–1433.e1. [DOI] [PubMed] [Google Scholar]
- 20. Ye Z, Miranda WR, Yeung DF, Kane GC, Oh JK.. Left atrial strain in evaluation of heart failure with preserved ejection fraction. J Am Soc Echocardiogr 2020; 33:1490–9. [DOI] [PubMed] [Google Scholar]
- 21. Labombarda F, Bonopera M, Maragnes P, Milliez P, Manrique A, Beygui F.. Impaired left atrial function in adults and adolescents with corrected aortic coarctation. Pediatr Cardiol 2021; 42:199–209. [DOI] [PubMed] [Google Scholar]
- 22. Egbe A, Miranda W, Connolly H, Dearani J.. Haemodynamic determinants of improved aerobic capacity after tricuspid valve surgery in Ebstein anomaly. Heart 2021;107:1138–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Morris DA, Takeuchi M, Krisper M, Kohncke C, Bekfani T, Carstensen T. et al. Normal values and clinical relevance of left atrial myocardial function analysed by speckle-tracking echocardiography: multicentre study. Eur Heart J.Cardiovasc Imag 2015;16:364–72. [DOI] [PubMed] [Google Scholar]
- 24. Caballero L, Kou S, Dulgheru R, Gonjilashvili N, Athanassopoulos GD, Barone D. et al. Echocardiographic reference ranges for normal cardiac Doppler data: results from the NORRE Study. Eur Heart J.Cardiovasc Imag 2015;16:1031–41. [DOI] [PubMed] [Google Scholar]
- 25. Egbe AC, Reddy YNV, Obokata M, Borlaug BA.. Doppler-derived arterial load indices better reflect left ventricular afterload than systolic blood pressure in coarctation of aorta. Circulation Cardiovasc Imag 2020;13:e009672. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Egbe AC, Allison TG, Ammash NM.. Mild coarctation of aorta is an independent risk factor for exercise-induced hypertension. Hypertension 2019;74:1484–9. [DOI] [PubMed] [Google Scholar]
- 27. Melenovsky V, Hwang SJ, Redfield MM, Zakeri R, Lin G, Borlaug BA.. Left atrial remodeling and function in advanced heart failure with preserved or reduced ejection fraction. Circ Heart Fail 2015;8:295–303. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.