Abstract
Background:
Chronic elevation of left heart filling pressure causes pulmonary vascular remodeling, pulmonary hypertension, and right heart dysfunction. Although diastolic dysfunction is relatively common in patients with coarctation of aorta (COA), there are limited data about the prevalence and prognostic implications of pulmonary hypertension and right heart dysfunction in this population. The purpose of the study was to assess right heart function and hemodynamics in patients with COA, and to determine the relationship between right heart indices and cardiovascular events defined as heart failure hospitalization, heart transplant or cardiovascular death.
Methods:
Right heart structure, function and hemodynamics were assessed with these indices: right atrial (RA) volume, RA pressure, RA reservoir strain, right ventricular (RV) global longitudinal strain, RV end-diastolic area, RV systolic pressure, and tricuspid regurgitation severity. Right heart hemodynamic score (RHHS), range 0–5, was generated based on the correlation between the right heart indices and cardiovascular events, using half of the cohort (derivation cohort, n=411), and then tested on the validation cohort (n=410). The goodness of fit and discrimination power was compared using c-statistics and risk score.
Results:
The median follow-up in the derivation cohort was 8.2 (4.0–11.1) years, and 59 (14%) patients had cardiovascular events during this period. RHSS was independently associated with cardiovascular events (hazard ratio 1.64, 95% CI: 1.38–2.17) for every unit increase in RHHS) after adjustment for clinical and echocardiographic indices (c-statistic 0.718, 95% CI: 0.682–0.746). The RHHS was also independently associated with cardiovascular events in the validation cohort (c-statistic 0.711, 95% CI: 0.679–0.741). The c-statistic difference (0.007, 95% CI: 0.014–0.022) and risk score (0.86, 95% CI: 0.54–1.17) suggest a good model fit.
Conclusions:
The current study underscores the prognostic importance of right heart dysfunction in patients with COA and suggests that right heart indices should be used for risks stratification in this population.
Keywords: Coarctation of aorta, Right heart remodeling, Right heart dysfunction, Pulmonary hypertension
INTRODUCTION
Patients with left ventricular (LV) systolic or diastolic dysfunction and left-sided valvular heart disease often develop increased left heart filling pressure (left atrial [LA] hypertension), which in turn leads to pulmonary vascular remodeling, pulmonary hypertension, and right heart dysfunction.1 The onset of right heart dysfunction is a marker of advanced disease, and patients who develop right heart dysfunction have increased risk of cardiovascular death, heart failure hospitalization, and hepatorenal dysfunction.1, 2 As a result, right heart indices are used for prognostication in patients with heart failure due to left-sided valvular or ventricular dysfunction.3–6
Coarctation of aorta (COA) is one of the leading causes of LV dysfunction and heart failure in adults with congenital heart disease.7, 8 It is characterized by chronic LV pressure overload, which in turn leads to LV remodeling and increased left heart filling pressure.7, 8 Although LA hypertension is relatively common in patients with COA,7–9 and chronic LA hypertension is known to cause right heart remodeling and dysfunction in older patients with acquired heart disease,1 there are limited data about the prevalence and prognostic implications of pulmonary hypertension and right heart dysfunction in adults with COA.9, 10 The purpose of this study was to assess right heart structure, function and hemodynamics in patients with COA, and to determine the relationship between right heart indices and clinical outcomes. We hypothesized that echocardiographic indices of right heart structure, function, and hemodynamics were independently associated with clinical outcomes.
METHODS
Study Population
The authors declare that all supporting data are available within the article. This is a retrospective study of adults (age ≥18 years) with COA that underwent transthoracic echocardiogram at Mayo Clinic from January 1, 2000 and December 31, 2018. The patients were identified through the Mayo Adult Congenital Heart Disease (MACHD) registry. The Mayo Clinic Institutional Review Board approved the study, and informed consent was waived for patients that provided research authorization. As COA and left-sided valvular lesions can coexist, and these concomitant left-sided valvular lesions significantly influence clinical presentation and mortality,9 we divided our cohort into 3 mutually exclusive groups based on the presence of LV outflow and inflow disease as assessed by Doppler echocardiography. (1) Isolated COA; (2) COA with LV outflow disease defined as having any of the following conditions: aortic valve prosthesis, sub-valvular, valvular or supra-valvular aortic stenosis (mean gradient >20 mmHg) or ≥moderate aortic regurgitation, in the absence of mitral valve disease; (3) COA with LV inflow disease (Shone complex) defined as having any of the following conditions: mitral valve prosthesis, sub-valvular, valvular or supra-valvular mitral stenosis (mean gradient >3 mmHg) or ≥moderate mitral regurgitation.
We divided the patients into 2 groups (derivation cohort and validation cohort) using a random assignment based the on last digit of their research identification number (even vs odd numbers).
Echocardiographic Assessment of Right Heart Indices
We used the first echocardiogram performed within the study period as the baseline echocardiogram, and all the indices used in the study were derived from the baseline echocardiogram. Comprehensive transthoracic echocardiogram was performed according to contemporary guidelines,11–13 with offline image analyses and measurements performed by an experienced research sonographer (J.W). Right atrial (RA) structure, function, and hemodynamics were assessed using the following indices: RA volume index (RA enlargement defined as RA volume index >27 ml/m2 in females or >32 ml/m2 in males); RA reservoir strain (RA dysfunction defined as RA reservoir strain <31%); estimate RA pressure (RA hypertension defined as RA pressure ≥8 mmHg). Right ventricular (RV) structure, function, and hemodynamics were assessed using the following indices: RV end-diastolic area (RV enlargement defined as RV end-diastolic area >25 cm2); RV global longitudinal strain (RV systolic dysfunction defined as RV global longitudinal strain >−18%); (3) Tricuspid regurgitation (significant regurgitation defined ≥moderate tricuspid regurgitation based on qualitative Doppler assessment). Pulmonary artery (PA) hemodynamics was assessed using estimated RV systolic pressure (pulmonary hypertension defined as estimated RV systolic pressure >40 mmHg in the absence of RV outflow tract obstruction). These indices were selected because of known association with clinical outcomes in patients with heart failure and pulmonary hypertension, and the cut-off points used for defining normal vs abnormal indices were based on published data.3–6, 11, 12, 14, 15 These 7 right heart indices were used to generate a composite right heart hemodynamic score (RHHS) as described under the Statistical Analysis subsection.
RA and RV function were assessed using speckle tracking strain imaging obtained using Vivid E9 and E95 (General Electric Co, Fairfield, Connecticut) with M5S and M5Sc-D transducers (1.5–4.6 MHz) at frame rate of 40 to 80 Hz. Three-beat cine-loop clips were obtained from RV-focused apical four-chamber views. The ventricular septum was included for the assessment of RV global longitudinal strain. These images were exported (DICOM) and then analyzed offline using TomTec (TomTec Imaging Systems, Unterschleissheim, Germany). Adequate tracking by the software was visually verified and retraced if necessary, until adequate tracking was achieved.
Outcomes
The study outcome was the composite endpoint of heart failure hospitalization, heart transplant or cardiovascular death. All study outcomes were ascertained by comprehensive review of the medical records. Heart failure hospitalization was defined as a hospital admission for volume overload requiring intravenous diuretics.16 Exploratory analysis was performed to assess the relationship between right heart dysfunction and hepatorenal dysfunction because of recent data showing a high prevalence of hepatorenal dysfunction in patients with right-sided congenital heart lesions such as Ebstein anomaly.17 Hepatorenal function was assessed using the model for end-stage liver disease excluding international normalized ratio (MELD-XI score).17
Statistical Analysis
Data were presented as mean ± standard deviation, median (interquartile range) or count (%) as appropriate. Between-group comparisons were performed using analysis of variance test for continuous variables with normal distribution, Kruskal Wallis test for continuous variables with skewed distribution, and chi square test for categorical variables. Pairwise comparisons were performed using unpaired t test for continuous variables with normal distribution, Wilcoxon rank sum test for continuous variables with skewed distribution, and chi square test for categorical variables
To create the RHHS, we assessed the relationship between all 7 right heart indices and cardiovascular events using a multivariable Cox regression, and cardiovascular events were modeled as time-to-event variable from the first clinic encounter (time zero) to the last follow-up. The right heart indices were treated as categorical variables (normal vs abnormal). A composite RHHS was then generated by assigning 0.5 points for every 1 unit increase in hazard ratio (HR) for each of the right heart indices that had statistically significant relationship with cardiovascular events in the multivariable model.
We identified clinical, demographic, and echocardiographic indices associated with cardiovascular events using univariable Cox regression. Every variable used in the univariable model was based on the first (earliest) set of indices obtained within the study period. For the patients that underwent COA reintervention during follow-up (after 2000), the effect of COA intervention was modeled (COA intervention vs no intervention) as time-dependent variables to account for the differences in time to intervention. We addressed the problem of missing data using the conditional imputation method.18 We then assessed the prognostic power of RHHS for estimating the risk of cardiovascular events in the derivation cohort using multivariable Cox regression while adjusting for the indices selected a priori from univariable analysis with p<0.25 required for entry and p<0.1 required for a variable to remain in the multivariable model. We performed sensitivity analysis using complete case analysis in the subgroup of patients that had all 7 right heart indices (n=314). We also performed an exploratory analysis assessing the relationship between RHHS and hepatorenal function using Pearson correlation.
The Cox model from the derivation cohort was used to fit the data from the validation cohort. The discrimination power of the models in both cohorts was assessed by comparing the c-statistics from the derivation and validation cohorts. To assess the goodness of fit, we computed a risk score for subjects in the validation cohort using the estimates from the derivation cohort. The risk score was then used to compute calibration slope for the validation cohort. The regression slope was estimated by using the risk score as a predictor in the proportional hazard regression for subjects in the validation cohort. The model fit was assessed by checking that the beta estimate associated with the risk score was not different from 1 (suggesting reasonable model fit).
A p<0.05 was considered statistically significant. All statistical analyses were performed with JMP and SAS software (versions 14.1 and 9.4 respectively; SAS Institute Inc, Cary NC).
RESULTS
Baseline Characteristics
Of 821 COA patients that met the study inclusion criteria, 176 (21%) presented with native COA while 645 (79%) had prior COA repair. A total of 563 (69%) had isolated COA, 204 (25%) had COA with LV outflow disease, and 54 (6%) had COA with LV inflow disease. As compared to patients with isolated COA, those with associated left-sided valve lesions had higher prevalence of atrial arrhythmias, diuretic use, and higher N-terminal pro b-type natriuretic peptide levels (Table 1). The prevalence of abnormal right heart structure, function, and hemodynamics were as follows: RA dysfunction (16%, 127/795), RA enlargement (28%, 219/784), RA hypertension (17%, 132/814), RV systolic dysfunction (14%, 108/902), RV enlargement (9%, 66/741), tricuspid regurgitation (5%, 38/821), and pulmonary hypertension (20%, 149/749). Table 2 shows a comparison of right heart indices between the 3 groups. All 7 pre-specified right heart indices were significantly worse in patients with associated LV inflow and outflow disease as compared to those with isolated COA. There was modest correlation between RV systolic pressure and RVGLS (r=0.63, p<0.001) and RA reservoir strain (r=−0.58, p=0.004).
Table 1:
Baseline Characteristics (n=821)
| All (n=821) | Isolated COA (n=563) | COA+LVOD (n=204) | COA+LVID (n=54) | p | |
|---|---|---|---|---|---|
| Age, years | 32 (21–46) | 32 (21–46) | 32 (22–47) | 36 (24–65) | 0.4 |
| Male | 479 (58%) | 337 (60%) | 115 (56%) | 27 (50%) | 0.3 |
| Valvular heart disease | |||||
| Supra-valvular aortic stenosis | 9 (1%) | --- | 8 (4%) | 1 (2%) | --- |
| Valvular aortic stenosis | 140 (17%) | --- | 128 (62%) | 12 (22%)* | --- |
| Sub-valvular aortic stenosis | 84 (10%) | --- | 65 (32%) | 19 (35%) | --- |
| Bicuspid aortic valve | 511 (62%) | 321 (57%) | 155 (76%) | 35 (65%)* | <0.001 |
| Sub-valvular mitral stenosis | 17 (2%) | --- | --- | 17 (32%) | --- |
| Valvular mitral stenosis | 34 (4%) | --- | --- | 34 (63%) | --- |
| Supra-valvular mitral stenosis | 3 (0.4%) | --- | --- | 3 (6%) | --- |
| Comorbidities | |||||
| Hypertension | 441 (54%) | 321 (57%) | 95 (47%) | 25 (46%) | 0.02 |
| Coronary artery disease | 54 (7%) | 35 (6%) | 16 (8%) | 3 (6%) | 0.7 |
| Diabetes | 41 (5%) | 25 (4%) | 12 (6%) | 4 (7%) | 0.5 |
| Atrial fibrillation | 61 (7%) | 34 (6%) | 14 (7%) | 13 (24%)* | <0.001 |
| Atrial flutter/tachycardia | 16 (2%) | 9 (2%) | 2 (1%) | 5 (10%)* | 0.008 |
| Medications | |||||
| Beta blockers | 241 (29%) | 166 (29%) | 59 (30%) | 16 (30%) | 0.9 |
| Calcium channel blockers | 105 (13%) | 73 (13%) | 23 (11%) | 9 (17%) | 0.6 |
| RAAS antagonist | 229 (28%) | 132 (27%) | 61 (30%) | 16 (30%) | 0.7 |
| Diuretics | 91 (11%) | 51 (9%) | 28 (14%) | 12 (22%)* | 0.01 |
| Laboratory data | |||||
| GFR, ml/min/1.73m2 | 95±21 | 95±20 | 98±22 | 91±21 | 0.3 |
| NT-proBNP, pg/ml | 220 (68–719) | 192 (41–484) | 221 (89–1244) | 459 (151–1810) | 0.02 |
COA: Coarctation of aorta; LVOD: left ventricular outflow disease; LVID: left ventricular inflow disease; RAAS: Renin angiotensin aldosterone antagonist; GFR: Glomerular filtration rate; NT-proBNP N-terminal pro b-type Natriuretic Peptide
Values were expressed as mean ± standard deviation for continuous variables with normal distribution, median (interquartile range) for continuous variables with skewed distribution, and count (%) for categorical variables. The p values were derived from between-group comparisons using analysis of variance test for continuous variables with normal distribution, Kruskal Wallis test for continuous variables with skewed distribution, and chi square test for categorical variables. Pairwise comparisons between LVOD vs LVID were performed using unpaired t test for continuous variables with normal distribution, Wilcoxon rank sum test for continuous variables with skewed distribution, and chi square test for categorical variables.
Denotes p<0.05 for pairwise comparison between LVOD vs LVID
Table 2:
Echocardiographic Indices (n=821)
| All (n=821) | Isolated COA (n=563) | COA+LVOD (n=204) | COA+LVID (n=54) | p | |
|---|---|---|---|---|---|
| RA Indices | |||||
| RA reservoir strain, % [n=795] | 43±11 | 45±12 | 41±11 | 35±14* | <0.001 |
| RA volume, ml/m2 [n=784] | 25±7 | 22±8 | 25±11 | 27±12 | 0.02 |
| RA pressure, mmHg [n=814] | 6±3 | 5±2 | 6±3 | 8±4* | <0.001 |
| RV Indices | |||||
| RV GLS, % [n=802] | −24±4 | −25±5 | −23±5 | −22±6 | <0.001 |
| RV EDA, mm [n=741] | 22±6 | 21±6 | 22±6 | 23±5 | 0.04 |
| RV FAC, % | 45±7 | 45±8 | 45±8 | 42±9 | 0.2 |
| RV s’, cm/s | 13±2 | 13±2 | 11±3 | 9±4* | <0.001 |
| TAPSE, mm | 22±7 | 23±5 | 20±5 | 18±6 | <0.001 |
| ≥Moderate TR [n=821] | 38 (5%) | 18 (3%) | 12 (6%) | 8 (15%)* | 0.003 |
| PA Hemodynamics | |||||
| TR velocity, m/s | 2.6±0.4 | 2.5±0.4 | 2.7±0.5 | 2.9±0.6* | <0.001 |
| RVSP, mmHg [n=749] | 34±13 | 32±11 | 37±14 | 45±17* | <0.001 |
| Other echo indices | |||||
| LV GLS, % | −21±4 | −21±4 | −20±3 | −20±4 | 0.01 |
| LV ejection fraction, % | 62±7 | 62±7 | 62±9 | 60±9 | 0.6 |
| LV mass index, g/m2 | 103±27 | 98±25 | 107±24 | 103±26 | 0.001 |
| LA reservoir strain, % | 36±5 | 38±5 | 36±5 | 34±5 | 0.007 |
| LA volume index, mL/m2 | 38±9 | 38±8 | 39±7 | 40±4 | 0.4 |
| Mitral E velocity, m/s | 1.1±0.3 | 1.1±0.4 | 1.0±0.3 | 1.3±0.3 | 0.2 |
| Septal e velocity, cm/s | 10±5 | 11±4 | 10±5 | 8±3 | 0.006 |
| Lateral e velocity, cm/s | 11±4 | 11±5 | 11±3 | 9±3 | 0.01 |
| MV mean gradient, mmHg | 5 (3–7) | --- | 4 (3–5) | 7 (5–10)* | --- |
| ≥Moderate mitral regurgitation | 16 (2%) | --- | ---- | 16 (30%) | --- |
| AV mean gradient, mmHg | 10 (6–19) | 8 (6–10) | 18 (9–31) | 14 (9–24) | <0.001 |
| ≥Moderate aortic regurgitation | 56 (6%) | --- | 53 (26%) | 3 (6%)* | --- |
| COA mean gradient, mmHg | 13 (8–19) | 14 (9–20) | 12 (7–16) | 13 (8–21) | 0.1 |
COA: Coarctation of aorta; LVOD: left ventricular outflow disease; LVID: left ventricular inflow disease; RA: Right atrium; RV: Right ventricle; GLS: Global longitudinal strain; EDA: End-diastolic area; FAC: Fractional area change; s’: Tissue Doppler systolic velocity: TR: Tricuspid regurgitation; RVSP: Right ventricular systolic pressure: PA: Pulmonary artery; E: mitral valve pulsed wave early velocity; e’: tissue Doppler early velocity; LV: Left ventricle; MV: Mitral valve: AV: Aortic valve: TAPSE: Tricuspid annular plane systolic excursion
[n=] denotes number of patients with each of the pre-specified right heart indices
Values were expressed as mean ± standard deviation for continuous variables with normal distribution, median (interquartile range) for continuous variables with skewed distribution, and count (%) for categorical variables. The p values were derived from between-group comparisons using analysis of variance test for continuous variables with normal distribution, Kruskal Wallis test for continuous variables with skewed distribution, and chi square test for categorical variables. Pairwise comparisons between LVOD vs LVID were performed using unpaired t test for continuous variables with normal distribution, Wilcoxon rank sum test for continuous variables with skewed distribution, and chi square test for categorical variables.
Denotes p<0.05 for pairwise comparison between LVOD vs LVID
Derivation Cohort (n=411)
Of the 821 patients, 411 and 410 were randomly assigned to the derivation and validation cohorts respectively. Compared to the derivation cohort, patients in the validation cohort had a higher prevalence of hypertension and use of renin angiotensin aldosterone system antagonists but were less likely to have LV inflow disease or use of diuretics (Supplementary Table S1).
The median follow-up for the 411 patients in the derivation cohort was 8.2 (4.0–11.1) years, and during this period, 38 (9%) patients were hospitalized for heart failure, 39 (10%) patients died from cardiovascular causes (end-stage heart failure n=24, arrhythmic/sudden cardiac death n=7, stroke-related deaths n=4, and postoperative death n=3), and 4 (1%) patients underwent heart transplant for end-stage left heart failure. The combined outcome of cardiovascular events occurred in 59 (14%) patients.
Right Heart Hemodynamic Score
All 7 right heart indices were associated with the combined outcome of cardiovascular events on univariable analysis when modeled as continuous variables and as categorical variables (normal vs abnormal values), Table 3. On multivariable analysis, only RA strain, RA volume index, RV global longitudinal strain and RV systolic pressure remained independently associated with the primary outcome, and a RHHS (range 0 to 5) was generated using these 4 indices (Table 4). We then stratified the patients into 3 risk groups based on their RHHS: low RHHS group (RHHS 0–1, [n=305, 74%]), intermediate RHHS group (RHHS 2–3, [n=67, 16%]) and high RHHS group (RHHS 4–5, [n=39, 10%]). As compared to the low RHHS group, the intermediate and high RHHS groups were older, and more likely to have associated LV inflow and outflow disease (Supplementary Table S2). The age at the time of COA repair was not significantly different between the low, intermediate, and high RHHS subgroups (3 [1–4] vs 3[1–4] vs 3 [1–4] years, p=0.6), Supplementary Table S2. There was no correlation between age at the time of COA repair and RVGLS (r=0.26, p=0.1), RA reservoir strain (r=0.21, p=0.2), or RHHS (r=0.17, p=0.4).
Table 3:
Univariable Cox Models Assessing Relationship between Right Heart Indices and Incident Cardiovascular Events
| Model using continuous variables | HR (95% CI) | p |
|---|---|---|
| RA indices | ||
| RA pressure, per 1 mmHg increase | 1.28 (1.11–1.33) | <0.001 |
| RA reservoir strain, per 1% decrease | 1.09 (1.03–1.15) | <0.001 |
| RA volume index, per 1 ml/m2 increase | 1.06 (1.01–1.11) | 0.008 |
| RV indices | ||
| RV global longitudinal strain, per 1% decrease | 1.07 (1.02–1.98) | 0.004 |
| RV end-diastolic area, per 1 cm2 increase | 1.05 (1.01–1.10) | 0.03 |
| Tricuspid regurgitation | --- | --- |
| PA indices | ||
| RVSP, per 1 mmHg increase | 1.06 (1.02–1.09) | 0.005 |
| Model using categorical variables | HR (95% CI) | p |
| RA indices | ||
| RA pressure ≥8 mmHg | 3.19 (1.43–6.91) | <0.001 |
| RA reservoir strain <31% | 2.66 (1.93–5.14) | <0.001 |
| RA volume index >27 ml/m2 (F) or >32 ml/m2 (M) | 2.89 (1.17–4.52) | 0.007 |
| RV indices | ||
| RV global longitudinal strain >−18% | 2.23 (1.18–4.76) | <0.001 |
| RV end-diastolic area >25 cm2 | 2.91 (1.08–7.24) | 0.02 |
| ≥Moderate tricuspid regurgitation | 2.17(1.25–5.17) | 0.008 |
| PA indices | ||
| RVSP >40 mmHg | 4.26 (2.92–9.14) | <0.001 |
RA: Right atrium; RV: Right ventricle; RVSP: Right ventricular systolic pressure; PA: Pulmonary artery; HR: Hazard ratio; CI: Confidence interval
The right heart indices were modeled as continuous variables (top) and as binary variables (normal vs abnormal values; bottom). The cut-off points for categorization into binary groups were based on published data described in the Methods section.
Table 4:
Multivariable Cox Model Assessing Relationship between Right Heart Indices and Incident Cardiovascular Events
| Multivariable | HR (95% CI) | p | Points |
|---|---|---|---|
| RA reservoir strain <31% | 1.67 (1.12–3.89) | 0.004 | 1 |
| RA volume index >27 ml/m2 (F) or >32 ml/m2 (M) | 1.98 (1.02–4.15) | 0.02 | 1 |
| RV global longitudinal strain >−18% | 1.73 (1.08–3.03) | 0.008 | 1 |
| RVSP >40 mmHg | 3.26 (1.54–7.34) | <0.001 | 2 |
| RA pressure ≥8 mmHg | NS | NS | |
| RV end-diastolic area >25 cm2 | NS | NS | |
| ≥Moderate tricuspid regurgitation | NS | NS | |
| LV GLS, per unit increment | 1.14 (1.06–1.29) | 0.001 | |
| LV mass index per 5 g/m2 increment | 1.06 (1.01–1.15) | 0.006 | |
| Right Heart Hemodynamic Score (RHHS) | 5 |
RA: Right atrium; RV: Right ventricle; RVSP: Right ventricular systolic pressure; PA: Pulmonary artery; HR: Hazard ratio; CI: Confidence interval
A RHHS was created by assigning 0.5 points for every 1 unit increase in HR for each of 4 right heart indices in the model. The total score was then computed for each patient by summing up the points for each of the abnormal indices present in each patient. The composite score ranged from 0 to 5.
NS denote variables that were not statistically significant.
Of the 411 patients in the derivation cohort, 385 (94%) patients had no LV inflow disease (i.e. patients with isolated COA or COA with associated LV outflow disease). Of these 385 patients without LV inflow disease, the RVGLS was −24±5%, RV systolic pressure was 33±9 mmHg, RA reservoir stain was 43±11%, and the median RHHS was 1 (0–2). Aortic valve mean gradient (a measure of residual LV outflow obstruction) had a correlation with RV systolic pressure (r=0.32, p=0.03), but not with the other right heart indices (RVGLS, r=0.18, p=0.3; RA reservoir strain, r=0.11, p=0.5; and RHHS, r=0.27, p=0.09). There was no correlation between COA mean gradient (a measure of residual coarctation) and any of the right heart indices (RV systolic pressure, r=0.24, p=0.1; RVGLS, r=0.14, p=0.6; RA reservoir strain, r=0.08, p=0.7; and RHHS, r=0.25, p=0.1).
RHHS, LV global longitudinal strain, atrial fibrillation, and hypertension were independently associated with cardiovascular events after adjustment for clinical and echocardiographic indices (c-statistic 0.718, 95% confidence interval 0.682–0.746), (Table 5 and 6).
Table 5:
Univariable Cox Model for Predictors of Incident Cardiovascular events
| HR (95% CI) | p | |
|---|---|---|
| COA mean gradient, per 1 mmHg | 1.26 (0748–2.54) | 0.5 |
| Aortic mean gradient, per 1 mmHg | 1.08 (0.91–1.12) | 0.3 |
| LV GLS, per unit increment | 0.91 (0.85–0.97) | 0.008 |
| LV mass index, per 5 g/m2 | 1.11 (1.02–1.25) | 0.009 |
| Atrial fibrillation | 2.02 (1.28–4.04) | <0.001 |
| Hypertension | 1.53 (1.18–2.33) | 0.01 |
| Coronary artery disease | 1.96 (1.24–4.16) | 0.004 |
| GFR, per 5-unit increment | 0.93 (0.84–1.01) | 0.06 |
| COA intervention* | 1.42 (0.81–1.83) | 0.5 |
| Age at time of COA repair, years | 1.03 (0.89–1.21) | 0.4 |
| LVOT intervention† | 0.92 (0.80–1.06) | 0.3 |
| Mitral valve intervention‡ | 1.31 (0.92–1.74) | 0.4 |
COA: Coarctation of aorta; LV: Left ventricle; GLS: Global longitudinal strain; GFR: Glomerular filtration rate; HR: Hazard rate; CI: Confidence interval, LVOT: Left ventricular outflow tract
COA intervention* denotes surgical COA repair or transcatheter intervention; LVOT intervention† denotes surgical resection of subaortic stenosis or aortic valve replacement; Mitral valve intervention‡ denotes mitral valve repair or replacement
The variables with p<0.25 were then used to create the multivariable Cox model in Table 6
Table 6:
Multivariable Cox Model Assessing Relationship between Right Heart Hemodynamic Score and Incident Cardiovascular Event in the Derivation Cohort
| Model using continuous variables | HR (95% CI) | p |
|---|---|---|
| Right heart hemodynamic score, per unit | 1.64 (1.38–2.17) | <0.001 |
| LV GLS, per unit increment | 1.14 (1.03–1.39) | 0.004 |
| Atrial fibrillation | 1.51 (1.19–2.88) | 0.006 |
| Hypertension | 2.02 (1.14–3.08) | 0.007 |
| Model using categorical variables | HR (95% CI) | p |
| Right heart hemodynamic score | ||
| Low risk group (0–1) | Reference | --- |
| Intermediate risk group (2–3) | 2.21 (1.19–6.53) | 0.008 |
| High risk group (4–5) | 7.33 (2.72–18.41) | <0.001 |
| LV GLS, per unit increment | 1.08 (1.01–1.17) | 0.03 |
| Atrial fibrillation | 1.26 (1.04–31.98) | 0.01 |
| Hypertension | 2.16 (1.07–3.74) | 0.01 |
LV: Left ventricle; GLS: Global longitudinal strain; GFR: Glomerular filtration rate; HR: Hazard rate; CI: Confidence interval
Right heart hemodynamic score was modeled as a continuous variable (top), and as categorical variable (bottom)
The clinical variables included in these multivariable models were selected based on stepwise backwards selection of variables with p<0.25 and p<0.1 required for entry and exit from the multivariable models. These p values were obtained from a univariable Cox model shown in Table 5.
Sensitivity Analyses
In the analyses described above, conditional imputation was used to account for missing data since all the 7 pre-specified right heart indices were not present in all patients. In order to assess the robustness of imputation, we performed sensitivity analyses using the complete case analysis method, and these analyses were based on the subset of patients that had data for all 7 right heart indices (n=314). The multivariable Cox regression analyses identified the same right heart indices with comparable prognostic power as the parent model (c-statistic 0.704, 95% confidence interval 0.679–0.738).
Exploratory Analyses
Exploratory analysis was performed to assess the relationship between right indices and hepatorenal function. Of the 411 patients, 395 (96%) had a comprehensive metabolic panel test at the time of baseline assessment, and the median MELD-XI score was 9.7 (9.4–10.5). There was a correlation between RHHS and baseline MELD-XI score (r=0.44, p=0.008).
Validation Cohort (n=410)
The median follow-up for the 410 patients in the validation cohort was 9.6 (5.1–11.8) years, and during this period, 35 (9%) patients were hospitalized for heart failure, 33 (8%) patients had cardiovascular death, and 3 (0.7%) patients underwent heart transplant. The combined outcome of cardiovascular events occurred in 52 (13%) patients. Using the same RHHS criteria as in the derivation cohort, the 410 patients in the validation cohort were classified into low RHHS group (RHHS 0–1, [n=303, 74%]), intermediate RHHS group (RHHS 2–3, [n=82, 20%]) and high RHHS group (RHHS 4–5, [n=25, 6%]).
The Cox model from the derivation cohort was used to fit the data from the validation cohort (Table 7). The resulting c-statistics for the models were similar between the two cohorts suggesting similar discriminatory power (derivation c-statistics 0.718, 95% CI 0.682–0.746, and validation c-statistics 0.711, 95% CI 0.679–0.741). A risk score was computed for subjects in the validation cohort using the estimates from the derivation cohort, and the risk score observed in the validation cohort was 0.86 (95% CI: 0.54–1.17) suggesting a good model fit.
Table 7:
Comparison of Multivariable Cox Model Assessing Relationship between Right Heart Hemodynamic Score and Incident Cardiovascular Event in Both Cohorts
| Derivation HR (95% CI) | Validation HR (95%CI) | |
|---|---|---|
| Right heart hemodynamic score, per unit | 1.64 (1.38–2.17) | 1.51 (1.32–2.04) |
| LV GLS, per unit increment | 1.14 (1.03–1.39) | 1.10 (1.02–1.23) |
| Atrial fibrillation | 1.51 (1.19–2.88) | 1.37 (1.15–2.34) |
| Hypertension | 2.02 (1.14–3.08) | 2.32 (1.24–2.32) |
| c-statistic (95% CI) | c-statistic (95% CI) | |
| 0.718 (0.682–0.746) | 0.711 (0.679–0.741) |
LV: Left ventricle; GLS: Global longitudinal strain; HR: Hazard rate; CI: Confidence interval
Of the 410 patients, 401 (98%) had a comprehensive metabolic panel test at the time of baseline assessment, and the median MELD-XI score was 9.6 (9.4–10.8). There was a correlation between RHHS and baseline MELD-XI score (r=0.39, p=0.02). There was no significant between-group difference in the robustness of correlation between RHHS and MELD-XI score in the derivation cohort vs the validation cohort (r=0.44 vs r=0.39, Meng test p=0.5)
DISCUSSION
Right Heart Dysfunction and Clinical Outcome
There are limited data about the prevalence and prognostic implications of right heart dysfunction in patients with COA.9, 10 The current study showed that RA dysfunction, RV dysfunction and pulmonary hypertension were present in 16%, 14%, 20% of our cohort respectively. We proposed the RHHS as a composite metric of right heart structure, function, and hemodynamics. For every unit increase in RHHS, we observed 64% increase in the risk of cardiovascular events, and similar estimates were observed both in the derivation and validation cohorts, suggesting that RHHS has robust prognostic power.
Two previous studies have described the prevalence and prognostic implications of pulmonary hypertension in patients with COA.9, 10 One of the studies was based on echocardiographic assessment of 159 adults with repaired COA, and in that study, Oliver et al10 reported that pulmonary hypertension (defined as RV systolic pressure >40 mmHg in the absence of RV outflow tract obstruction) was present in 19% of the patients, and that pulmonary hypertension was an independent risk factor for all-cause mortality and heart failure hospitalization. In a more recent study that was based on invasive hemodynamic assessment of 25 adults with Shone complex, Jain et al9 demonstrated that pulmonary hypertension (defined as mean PA pressure >20 mmHg) was present in 96% of the patients, and was associated with mortality. Collectively, these studies underscore the prognostic implication of pulmonary hypertension in the COA population. The significant difference in the prevalence of pulmonary hypertension from these 2 studies has to do with differences in the demographic characteristics of the patients in the studies. The Jain et al study was based on a selected cohort of patients with Shone complex referred for cardiac catheterization, hence the high prevalence of pulmonary hypertension. The demographic characteristics of the Oliver et al study was similar to that of the current study, hence the similar prevalence of pulmonary hypertension between both studies. In contrast to these prior studies, we did not study the prognostic implications of pulmonary hypertension as an isolated metric, but rather our results were based on a holistic assessment of right heart remodeling (changes in structure, function, and hemodynamics). Additionally, the large sample size in the current study allowed for multivariable adjustments for multiple confounders, hence improving the robustness of the results.
RV dysfunction in the setting of left-sided lesions is not limited to adults with congenital heart disease, but is well described in patients with acquired heart disease.19–21 Patients with heart failure with preserved ejection fraction (HFpEF) that have RV systolic dysfunction at baseline, or developed RV systolic dysfunction during follow-up, have a higher risk of clinical deterioration and mortality during follow-up.20 Furthermore, RV systolic dysfunction was an independent risk factor for mortality in patients with severe aortic stenosis regardless of whether they received medical therapy or underwent aortic valve replacement, suggesting that perhaps interventions should be performed prior to onset of RV systolic dysfunction.22
We also observed a strong association between atrial fibrillation and outcomes (Table 6 and 7). We postulate that this may be related to the fact that atrial fibrillation is both a cause and an effect of advanced atrial remodeling, which in turn, is related to pulmonary hypertension and RV systolic dysfunction. Hence, atrial fibrillation was likely a hallmark of advanced cardiac remodeling as evidenced by the higher prevalence of atrial fibrillation in patients with high RHHS (Supplementary Table S2), and the higher risk of cardiovascular events in patients with atrial fibrillation (Table 6 and 7).
Another important finding from this study is that the deleterious effect of right heart dysfunction was not limited to the cardiovascular system but can affect multi organ-systems including the liver and kidney. While hepatorenal dysfunction is a well-recognized entity in adults with congenital heart disease, it is typically observed in patients with Fontan palliation and those with right-sided lesions that predispose them to chronic central venous hypertension, and in turn, hepatorenal dysfunction.17, 23, 24 The current study demonstrates for the first time that COA patients with advanced right heart remodeling (high RHHS) had a higher risk of hepatorenal dysfunction.
Clinical Implications and Future Directions
The current study highlights the importance of the right heart in the pathogenesis of heart failure, hepatorenal dysfunction, and mortality in patients with COA, and suggests that RHHS (a composite metric of right heart function and hemodynamics) can be used for risk stratification in this population. The results of the current study also provide the scientific premise for hypothesis-driven studies to determine whether medical, transcatheter and surgical interventions to unload the left heart will lead to improvement in right heart function and clinical outcomes. Addressing these questions will be critical to improving the long-term survival of COA patients, which remains suboptimal (median longevity ~50 years) despite having effective surgical and transcatheter therapies for COA repair.25, 26
Limitations
This is a retrospective study single center study, and it is therefore prone to selection and ascertainment bias, and this may limit generalizability of the results. However, we observed a consistent robust prognostic power of the proposed RHHS score both in the derivation and validation cohorts, suggesting that this score can potentially be applied to other populations. We did not have invasive hemodynamic indices, and hence we were unable to delineation of the mechanisms responsible for pulmonary hypertension and right heart dysfunction in this population.
Conclusions
In conclusion, more than one-fifth of patients with COA had abnormal right heart structure, function, and hemodynamics, and the severity of right heart remodeling was associated with clinical outcomes. The current study underscores the prognostic importance of right heart dysfunction in patients with COA and provides a foundation for further studies to determine whether the use of right heart indices for guiding therapies will improve clinical outcomes in this population.
Supplementary Material
Clinical Perspective.
In this study, we assessed right heart function and hemodynamics in patients with COA and evaluated the relationship between right heart indices and cardiovascular events. Our results showed that more than one-fifth of patients with COA had abnormal right heart structure, function, and hemodynamics, and the severity of right heart remodeling was associated with clinical outcomes. The study underscores the prognostic importance of right heart dysfunction in patients with COA and provides a foundation for further studies to determine whether the use of right heart indices for guiding therapies will improve clinical outcomes in this population.
Acknowledgement:
James Welper (J.W) for performing offline analysis of all echocardiographic data
Funding:
Dr. Egbe is supported by National Heart, Lung, and Blood Institute (NHLBI) grants (R01 HL158517 and K23 HL141448). The MACHD Registry is supported by the Al-Bahar Research
Abbreviations:
- RA
right atrium
- RV
right ventricle
- LA
left atrium
- LV
left ventricular
- PA
pulmonary artery
- COA
coarctation of aorta
- RHHS
right heart hemodynamic score
- MELD-XI
model for end-stage liver disease excluding international normalized ratio
- HR
hazard ratio
- E
pulsed wave Doppler early velocity
- e’
tissue Doppler early velocity
- HFpEF
heart failure with preserved ejection
Footnotes
REFERENCES
- 1.Haddad F, Doyle R, Murphy DJ and Hunt SA. Right ventricular function in cardiovascular disease, part II: pathophysiology, clinical importance, and management of right ventricular failure. Circulation. 2008;117:1717–31. [DOI] [PubMed] [Google Scholar]
- 2.Granada J, Uribe W, Chyou PH, Maassen K, Vierkant R, Smith PN, Hayes J, Eaker E and Vidaillet H. Incidence and predictors of atrial flutter in the general population. J Am Coll Cardiol. 2000;36:2242–6. [DOI] [PubMed] [Google Scholar]
- 3.Mohammed SF, Hussain I, AbouEzzeddine OF, Takahama H, Kwon SH, Forfia P, Roger VL and Redfield MM. Right ventricular function in heart failure with preserved ejection fraction: a community-based study. Circulation. 2014;130:2310–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Meyer P, Filippatos GS, Ahmed MI, Iskandrian AE, Bittner V, Perry GJ, White M, Aban IB, Mujib M, Dell’Italia LJ, et al. Effects of right ventricular ejection fraction on outcomes in chronic systolic heart failure. Circulation. 2010;121:252–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ye Y, Desai R, Vargas Abello LM, Rajeswaran J, Klein AL, Blackstone EH and Pettersson GB. Effects of right ventricular morphology and function on outcomes of patients with degenerative mitral valve disease. J Thorac Cardiovasc Surg. 2014;148:2012–2020 e8. [DOI] [PubMed] [Google Scholar]
- 6.Galli E, Guirette Y, Feneon D, Daudin M, Fournet M, Leguerrier A, Flecher E, Mabo P and Donal E. Prevalence and prognostic value of right ventricular dysfunction in severe aortic stenosis. Eur Heart J Cardiovasc Imaging. 2015;16:531–8. [DOI] [PubMed] [Google Scholar]
- 7.Egbe AC, Miranda WR and Connolly HM. Increased prevalence of left ventricular diastolic dysfunction in adults with repaired coarctation of aorta. Ijc Heart Vasc. 2020;28.100530 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Egbe AC, Qureshi MY and 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.Jain CC, Warnes CA, Egbe AC, Cetta F, DuBrock HM, Connolly HM and Miranda WR. Hemodynamics in Adults With the Shone Complex. Am J Cardiol. 2020;130:137–142. [DOI] [PubMed] [Google Scholar]
- 10.Oliver JM, Gallego P, Gonzalez AE, Sanchez-Recalde A, Bret M and Aroca A. Pulmonary hypertension in young adults with repaired coarctation of the aorta: an unrecognised factor associated with premature mortality and heart failure. Int J Cardiol. 2014;174:324–9. [DOI] [PubMed] [Google Scholar]
- 11.Rudski LG, Lai WW, Afilalo J, Hua L, Handschumacher MD, Chandrasekaran K, Solomon SD, Louie EK, Schiller NB. Guidelines for the echocardiographic assessment of the right heart in adults: a report from the American Society of Echocardiography endorsed by the European Association of Echocardiography, a registered branch of the European Society of Cardiology, and the Canadian Society of Echocardiography. J Am Soc Echocardiogr 2010;23:685–713; quiz 786–8. [DOI] [PubMed] [Google Scholar]
- 12.Lang RM, Badano LP, Mor-Avi V, Afilalo J, Armstrong A, Ernande L, Flachskampf FA, Foster E, Goldstein SA, Kuznetsova T, et al. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr. 2015;28:1–39 e14. [DOI] [PubMed] [Google Scholar]
- 13.Mitchell C, Rahko PS, Blauwet LA, Canaday B, Finstuen JA, Foster MC, Horton K, Ogunyankin KO, Palma RA, Velazquez EJ. Guidelines for Performing a Comprehensive Transthoracic Echocardiographic Examination in Adults: Recommendations from the American Society of Echocardiography. J Am Soc Echocardiogr. 2019;32:1–64. [DOI] [PubMed] [Google Scholar]
- 14.Padeletti M, Cameli M, Lisi M, Malandrino A, Zaca V and Mondillo S. Reference values of right atrial longitudinal strain imaging by two-dimensional speckle tracking. Echocardiogr. 2012;29:147–52. [DOI] [PubMed] [Google Scholar]
- 15.Morris DA, Krisper M, Nakatani S, Kohncke C, Otsuji Y, Belyavskiy E, Radha Krishnan AK, Kropf M, Osmanoglou E, Boldt LH, et al. Normal range and usefulness of right ventricular systolic strain to detect subtle right ventricular systolic abnormalities in patients with heart failure: a multicentre study. Eur Heart J Cardiovasc Imaging. 2017;18:212–223. [DOI] [PubMed] [Google Scholar]
- 16.Egbe AC, Miranda WR, Ammash NM, Ananthaneni S, Sandhyavenu H, Farouk Abdelsamid M, Yogeswaran V, Kapa S, Fatola A, Kothapalli S, et al. Atrial Fibrillation Therapy and Heart Failure Hospitalization in Adults With Tetralogy of Fallot. JACC Clin Electrophysiol. 2019;5:618–625. [DOI] [PubMed] [Google Scholar]
- 17.Egbe AC, Miranda WR, Dearani J, Kamath PS and Connolly HM. Prognostic Role of Hepatorenal Function Indexes in Patients With Ebstein Anomaly. J Am Coll Cardiol. 2020;76:2968–2976. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Ali AM, Dawson SJ, Blows FM, Provenzano E, Ellis IO, Baglietto L, Huntsman D, Caldas C and Pharoah PD. Comparison of methods for handling missing data on immunohistochemical markers in survival analysis of breast cancer. Br J Cancer. 2011;104:693–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Obokata M and Borlaug BA. Right ventricular dysfunction and pulmonary hypertension in heart failure with preserved ejection fraction. Eur J Heart Fail. 2016;18:1488–1490. [DOI] [PubMed] [Google Scholar]
- 20.Obokata M, Reddy YNV, Melenovsky V, Pislaru S and Borlaug BA. Deterioration in right ventricular structure and function over time in patients with heart failure and preserved ejection fraction. Eur Heart J. 2019;40:689–697. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Borlaug BA, Kane GC, Melenovsky V and Olson TP. Abnormal right ventricular-pulmonary artery coupling with exercise in heart failure with preserved ejection fraction. Eur Heart J. 2016;37:3293–3302. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Bohbot Y, Guignant P, Rusinaru D, Kubala M, Marechaux S and Tribouilloy C. Impact of Right Ventricular Systolic Dysfunction on Outcome in Aortic Stenosis. Circ Cardiovasc Imaging. 2020;13:e009802. [DOI] [PubMed] [Google Scholar]
- 23.Konno R, Tatebe S, Sugimura K, Satoh K, Aoki T, Miura M, Suzuki H, Yamamoto S, Sato H, Terui Y, et al. Prognostic value of the model for end-stage liver disease excluding INR score (MELD-XI) in patients with adult congenital heart disease. PloS one. 2019;14:e0225403. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Assenza GE, Graham DA, Landzberg MJ, Valente AM, Singh MN, Bashir A, Fernandes S, Mortele KJ, Ukomadu C, Volpe M et al. MELD-XI score and cardiac mortality or transplantation in patients after Fontan surgery. Heart. 2013;99:491–6. [DOI] [PubMed] [Google Scholar]
- 25.Lee MGY, Babu-Narayan SV, Kempny A, Uebing A, Montanaro C, Shore DF, d’Udekem Y and Gatzoulis MA. Long-term mortality and cardiovascular burden for adult survivors of coarctation of the aorta. Heart. 2019;105:1190–1196. [DOI] [PubMed] [Google Scholar]
- 26.Choudhary P, Canniffe C, Jackson DJ, Tanous D, Walsh K and Celermajer DS. Late outcomes in adults with coarctation of the aorta. Heart. 2015;101:1190–5. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.