Determinants of Left Ventricular Diastolic Function and Exertional Symptoms in Adults with Coarctation of Aorta

Abstract

Background:

Coarctation of aorta (COA) results in chronic left ventricular (LV) pressure overload, and subsequently leads to LV diastolic dysfunction and heart failure over time. The goal of COA intervention is to prevent these complications. The timing of COA interventions is based on the presence of these COA severity indices: Doppler mean COA gradient, systolic blood pressure, upper-to-lower-extremity SBP gradient, aortic isthmus ratio, presence of collaterals, and exercise-induced hypertension. Although these indices are physiologically intuitive, the relationship between these indices and LV diastolic dysfunction and exertional symptoms has not been studied. The purpose of this study was to evaluate the association between the indices of COA severity and LV diastolic function and symptoms.

Methods:

In this cross-sectional study, multivariate linear and logistic regression analyses were used to assess the correlation between indices of COA severity, LV diastolic function (average e’ and E/e’), and exertional symptoms (NYHA II-IV and peak VO₂).

Results:

Of all the COA indices analyzed in 546 adult COA patients, aortic isthmus ratio had the strongest correlation with e’ (β [95% confidence interval]: 3.11 [2.02–4.31], p=0.014) per 1 cm/s; E/e’ (−13.4 [−22.3 - −4.81], p=0.009) per 1 unit; peak VO₂ (4.05 [1.97–6.59] per 1% change, p=0.019), and NYHA II-IV symptoms (odds ratio 2.16, 1.65–3.18, p=0.006).

Conclusions:

Of all the COA severity indices stipulated in the guidelines, aortic isthmus ratio had the strongest correlation with LV diastolic function and exertional symptoms. Since LV diastolic dysfunction, typically precede heart failure symptoms, we anticipate that the results of this study will improve and simplify patient selection for COA intervention, and potentially improve long term outcomes.

Among congenital heart disorders, coarctation of aorta (COA) is one of the leading causes of early cardiovascular mortality and heart failure-related morbidity.^{1, 2} COA results in left ventricular (LV) pressure overload because of mechanical obstruction at the aortic isthmus, vascular and endothelial dysfunction, and LV outflow tract obstruction in the setting of associated left-sided lesions.^3–6 Chronic LV pressure overload subsequently leads to cardiac myocyte hypertrophy, mismatch between myocardial oxygen demand and supply, and apoptosis with replacement fibrosis.⁷ These pathologic changes result in alteration in the viscoelastic properties and effective operative compliance of the LV which manifests clinically as diastolic dysfunction.⁷

LV diastolic dysfunction is common in patients with COA because of chronic pressure overload.⁸ Because of the known adverse effects of chronic LV pressure overload, the American College of Cardiology and the European Society of Cardiology have outlined indications for intervention in adults with COA.^{9, 10} These indications are based on the presence of the following COA severity indices: Doppler mean COA gradient, systolic blood pressure (SBP), upper-to-lower-extremity systolic blood pressure (ULE-SBP) gradient, exercise induced hypertension (EIH) and aortic isthmus ratio which is the ratio of the aortic isthmus diameter to the descending aorta diameter at the level of the diaphragm. ^{9, 10} However there are no data about the relative contributions of these COA indices to LV pressure overload, LV diastolic dysfunction or exertional symptoms. Such data will guide the clinician in determining which of the COA indices to prioritize when deciding on the timing of intervention. Since LV diastolic dysfunction reflects the deleterious effect of chronic pressure overload, the purpose of this study was to determine the association between indices of COA severity and LV diastolic function indices.

METHODS

Study Population

The data that support the findings of this study are available from the corresponding author upon reasonable request. We reviewed the MACHD (Mayo Adult Congenital Heart Disease) Registry and identified adult patients (age ≥18 years) with a diagnosis of COA that had at least one transthoracic echocardiogram and one cardiac cross-sectional imaging. The MACHD Registry contains data of all adults with congenital heart disease that received care at the Mayo Clinic Enterprise, from January 1, 1985 through December 31, 2018. The Mayo Clinic Institutional Review Board approved this study and waived informed consent for patients that provided research authorization. From this cohort, we excluded the following patients: (1) Patients with significant aortic valve disease defined as prosthetic aortic valve, native aortic valve peak velocity >2 m/sec or >mild aortic regurgitation. (2) Patients with significant mitral valve disease defined as prosthetic mitral valve, history of mitral valve repair, native mitral valve mean gradient >3 mmHg or >mild mitral regurgitation. (3) Patients with atrial fibrillation at the time of echocardiogram.

Study Design

The primary objective was to assess the correlation between indices of COA severity and LV diastolic function. LV diastolic function was measured using two indices: (1) tissue Doppler early velocity (e’) which is a measure of LV myocardial relaxation; (2) ratio of mitral inflow early velocity and tissue Doppler early velocity (E/e’) which is a measure of LV filling pressure.^{11, 12} The secondary objective was to assess the correlation between indices of COA severity and exertional symptoms. The New York Heart Association (NYHA) functional classification was used for subjective assessment of symptoms, and we defined symptoms as being in NYHA functional class II-IV. We used peak oxygen consumption (VO₂) on cardiopulmonary exercise test for objective assessment of symptoms. Exploratory analysis was performed to assess the temporal change in LV diastolic function indices and exertional symptoms at 5-years post intervention in the subgroup of patients that underwent COA intervention during follow-up.

For the purpose of this study, we defined COA severity using the indications for intervention stipulated in the American and European guidelines for the management of adults with congenital heart:^{9, 10} (1) Doppler mean gradient >20 mmHg or >10 mmHg in the presence of LV systolic dysfunction or collateral vessels; (2) SBP >140 mmHg; (3) ULE-SBP gradient >20 mmHg or >10 mmHg in the presence of LV systolic dysfunction or collateral vessels; (4) EIH; (5) aortic isthmus ratio which is the ratio of the aortic isthmus diameter to the descending aorta diameter at the level of the diaphragm ≤0.5.^{9, 10} Aortic isthmus ratio is measure of anatomic severity of COA while the other indices assessed the physiologic severity of COA.

Predictor and Outcome Variables

The first echocardiogram with tissue Doppler assessment performed within the study period was considered as the baseline echocardiogram. We also reviewed clinic notes, cross-sectional imaging studies and cardiopulmonary exercise test performed within 12 months from the time of the baseline echocardiogram. LV e’ and E/e’ were derived from this baseline echocardiogram. NYHA functional class and peak VO₂ were ascertained from the clinic notes and cardiopulmonary exercise test data respectively. Only peak VO₂ obtained from symptom-limited and maximum effort (respiratory exchange ratio >1.1) using a treadmill ergometer were analyzed in this study. Impaired exercise capacity was defined as peak VO₂ <80% predicted.¹³

For the predictor variables (COA severity indices), SBP was measured in the right arm, and ULE-SBP gradient was calculated as: SBP from the right arm minus SBP from the leg. Chest computed tomographic angiograms or magnetic resonance angiograms were used to determine the presence of collateral vessels and thoracic aortic dimensions. Collateral vessels were considered to be present or absent based on review of the imaging report. Aortic isthmus ratio was determined as a ratio of the aortic isthmus (the smallest COA diameter) divided by the diameter of the descending aorta at the level of the diaphragm.^{9, 10} EIH was defined as SBP >190 mmHg in females or >210 mmHg in males during symptom-limited treadmill exercise test with a respiratory exchange ratio >1.1.^{14, 15}

Imaging

Two-dimensional (2D) and Doppler echocardiography was performed according to contemporary guidelines.^{11, 16} Image analysis and off-line measurements were performed by two experienced sonographers (JW and KT). We used the average of the septal and lateral e’ when both variables were available, and we used either septal or lateral e’ in patients that had only one variable recorded.^{11, 12} LVMI was calculated using end-diastolic linear measurements of the interventricular septum, LV inferolateral wall thickness, and LV internal diameter derived from 2D echocardiography measured at the tissue-blood interphase.¹⁷ Similarly LVEF was calculated using 2D echocardiographic linear measurements of end-diastolic and end-systolic dimension obtained from the parasternal long-axis view.^{16, 17}

Pulsed wave Doppler was obtained about 2 to 5 mm proximal to the aortic isthmus (site of COA), and continuous wave Doppler was obtained across the aortic isthmus.^{16, 18} Only Doppler signals with optimal alignment defined as angle of insonation <20 degrees were analyzed for this study. Doppler peak velocity and time velocity integral were used to calculate uncorrected peak gradient (maximum instantaneous gradient) and mean gradients respectively.

Chest computed tomographic angiogram and magnetic resonance angiogram that were performed within 12 months prior to procedure were reviewed. Thoracic aortic dimensions were measured as previously described.^19–21

Statistical Analysis

Data were presented as mean ± standard deviation, median (interquartile range) or count (%). Multivariate linear and logistic regression analyses were used to assess the correlation between outcomes variables (e’, E/e’, NYHA functional class and peak VO₂) and the indices of COA severity (Doppler mean gradient, SBP, ULE-SBP gradient, aortic isthmus ratio, and presence of collateral vessels) using stepwise backwards selection, and a threshold of p<0.10 was required to remain in the model. NYHA function class was modeled as binary outcome variable (NYHA II-IV vs NYHA I). All patients had Doppler mean gradient, Doppler peak gradient and SBP data, NYHA functional class assessment and cross-sectional imaging data but only 525 (96%) had ULE-SBP gradient data. We used the simple conditional imputation method to substitute for missing data, this entailed replacing missing valve with the mean or median values for the sample as previously described.²² Exercise data was available in 341 (63%) patients, and as a result, we created separate linear and logistic regression models in this subgroup of patients.

Collinearity was assessed using variance inflation factor, and as a rule of thumb, significant collinearity was considered to be present if the variance inflation factor was >10. All regression models were adjusted for age, sex, body mass index, LV ejection fraction, history of hypertension, coronary artery disease, antihypertensive therapy, bicuspid aortic valve, and age at the time of initial COA repair because these variables were known to be associated with LV diastolic function. In the patients without prior COA repair, the current age used at the age of COA repair. The strength of correlation for the different variables were expressed as β coefficient and 95% confidence interval (CI), and the comparison of the strength of correlation between the different predictor variables were made on the basis of p valve. A p<0.05 was considered statistically significant. All statistical analyses were performed with JMP software (version 14.1.0; SAS Institute Inc, Cary NC).

RESULTS

Of 546 patients that met the inclusion criteria, 402 (74%) had prior COA repair (surgical repair n=348 and transcatheter n=54) and 144 (26%) had native COA respectively. Table 1 shows the baseline characteristics of the cohort. The average e’ and E/e’ were 10±4 cm/s and 11±3 respectively, Table 1. The average peak VO₂ was 26±9 ml/kg/min and 68±17 % predicted. Based on the cut-off points stipulated in the guidelines, 192 (35%), 144 (26%), 51 (9%), and 63 (12%) patients met the criteria for intervention based on Doppler mean gradient, SBP, ULE-SBP gradient and aortic isthmus ratio respectively. In the subgroups of patients with exercise data (n=341), 61 (18%) met the criteria for intervention based EIH.

Table 1:

Baseline Characteristics (N=546)

Age, years	33±9
Male	334 (61%)
Body mass index, kg/m2	26±4
Body surface area, m2	1.93±0.31
Hypertension diagnosis	326 (60%)
SBP, mmHg	132±19
ULE-SBP gradient, mmHg	14 (0–29)
Bicuspid aortic valve	309 (57%)
NYHA II-IV	214 (39%)
Medications
Beta blockers	142 (26%)
Calcium channel blockers	59 (11%)
ACEI/ARB	129 (24%)
Thiazide	47 (9%)
Hydralazine	3 (0.5%)
Any BP medication	341 (62%)
Echocardiography
LV end-diastolic dimension, mm	49±6
LV end-systolic dimension, mm	30±5
LV ejection fraction, %	62±7
LV septal wall thickness, mm	9±2
LV posterior wall thickness, mm	10±2
LVMI, g/m2	99±21
Relative wall thickness	0.40±0.04
Aortic valve peak velocity, m/s	1.45±0.31
COA peak velocity, m/s	2.71±0.52
*COA peak gradient, mmHg	29±11
COA mean gradient, mmHg	16±9
Diastolic Function Indices
Mitral E velocity, m/s	1.1±0.3
Mitral A velocity, m/s	0.7±0.2
Mitral deceleration time, ms	193±41
Septal e’ velocity, cm/s	10±5
Lateral e’ velocity, cm/s	11±4
Average e’ velocity, cm/s	10±4
Average E/e’	11±3
Left atrial volume index, mL/m2	38±5
Tricuspid regurgitation velocity, m/s	2.7±0.5
Cross-sectional imaging data
Mid ascending aorta, mm	30±7
Distal ascending aorta, mm	25±6
Proximal aortic arch, mm	19±7
Distal aortic arch, mm	19±8
Aortic isthmus, mm	14±5
Proximal descending aorta, mm	22±7
Descending aorta at diaphragm, mm	20±5
Aortic isthmus ratio	0.7±0.2
Collateral vessels	48 (9%)
Exercise testing (N=341)
Peak oxygen consumption, ml/kg/min	26±9
Peak oxygen consumption, % predicted	68±17
SBP at peak exercise, mmHg	173±32

ULE: Upper-to-lower extremity; SBP: Systolic blood pressure; ACEI: Angiotensin converting enzyme inhibitor; ARB: Angiotensin-II receptor blocker; LV: Left ventricle; LVMI: Left ventricular mass index; COA: Coarctation of aorta; E: Mitral early diastolic velocity; A: Mitral late diastolic velocity; e’: Mitral annular tissue Doppler early velocity; *COA peak gradient represents uncorrected maximum instantaneous gradient;

Data were presented as mean ± standard deviation, median (interquartile range) or number (%)

Of all the COA severity indices assessed, aortic isthmus ratio had the strongest correlation with e’ (β coefficients [95% CI] 3.11 [2.02–4.32], p=0.014) followed by Doppler mean gradient (0.19 [0.14–0.23], p=0.016), Table 2. Similarly, aortic isthmus ratio was the only COA severity metric that correlated with E/e’ (−13.4 [−22.3 - −4.81], p=0.009), Table 2. There were 44 (8%) patients that had only one e’ velocity (either septal or lateral) recorded. In a sensitivity analysis that excluded these 44 patients, we observed similar correlation between aortic isthmus ratio and e’ (3.19 [2.04–4.42], p=0.012), and between aortic isthmus ratio and E/e’ (13.7 [4.98–22.9], p=0.005).

Table 2:

Multivariate Predictors of LV Diastolic Function and Symptoms

	Full model		Final model
Predictors of e’ (R2=0.31, P<0.001)	β (95%CI)	p	β (95%CI)	p
Doppler mean gradient, per 5 mmHg	−0.21 (−0.26 - −0.15)	0.025	−0.19 (−0.23 - −0.14)	0.016
SBP, per 5 mmHg	−0.02 (−0.14 – 0.18)	0.432	---	---
ULE-SBP gradient, per 5 mmHg	0.01 (−0.24 – 0.31)	0.822	---	---
Aortic isthmus ratio	3.05 (1.14 – 5.22)	0.043	3.11 (2.02 – 4.31)	0.014
Presence of collateral vessels	0.01 (−1.87 – 2.02)	0.275	---	----
Predictors of E/e’ (R2=0.37 P<0.001)	β coefficient ±SE	p	β (95%CI)	p
COA severity indices
Doppler mean gradient, per 5 mmHg	0.01 (−0.11 – 0.13)	0.543	---	---
SBP, per 5 mmHg	0.06 (−0.02 – 0.11)	0.317	---	---
ULE-SBP gradient, per 5 mmHg	0.12 (−0.06 – 0.26)	0.417	---	---
Aortic isthmus ratio	−15.51(−21.2 - −10.02)	0.002	−13.4 (−22.3 - −4.81)	0.009
Presence of collateral vessels	0.87 (−0.56 – 0.294)	0.283	---	----
Predictors of peak VO2 (R2=0.32, P<0.001)	β coefficient ±SE	p	β (95%CI)	p
COA severity indices
Doppler mean gradient, per 5 mmHg	−0.16 (−0.27 – 0.06)	0.411	---	---
SBP, per 5 mmHg	−0.09 (−0.22 – 0.05)	0.451	---	---
ULE-SBP gradient, per 5 mmHg	−0.03 (−0.19 – 0.14)	0.677	---	---
Aortic isthmus ratio	4.22 (2.02 – 6.28)	0.007	4.05 (1.97–6.59)	0.019
Presence of collateral vessels	0.21 (−0.12 – 0.09)	0.676	---	----
Predictors of NYHA II-IV	Odds ratio (95% CI)	p	Odds ratio (95% CI)	p
COA severity indices
Doppler mean gradient, per 5 mmHg	1.35 (1.01–1.89)	0.012	1.27 (1.02–1.84)	0.019
SBP, per 5 mmHg	1.03 (0.83–2.14)	0.218	---	---
ULE-SBP gradient, per 5 mmHg	1.22 (0.74–2.84)	0.644	---	---
Aortic isthmus ratio	2.03 (1.44–3.19)	0.002	2.16 (1.65–3.18)	0.006
Presence of collateral vessels	1.33 (0.92–2.13)	0.208	---	----

LV: left ventricle; UL: upper-to-lower: SBP: systolic blood pressure; COA: coarctation of aorta; CI: confidence interval; VO₂: oxygen consumption; NYHA: York Heart Association

Aortic isthmus ratio also had a correlation with peak VO₂ (4.05 [1.97–6.59], p=0.019) and NYHA functional class (odds ratio 2.16, [1.65–3.18], p=0.006), Table 2. Incorporation of exercise data did not improve the robustness of the models, and EIH was not associated with e’, E/e’, peak VO₂ and NYHA functional class.

Aortic Isthmus Ratio

The aortic isthmus ratio for the cohort was 0.7±0.2 (median 0.7, interquartile range 0.6–0.9). Supplementary Figure 1 show the receiver operating characteristic curves used to identify the optimal cut-off points to predict elevated LV filling pressure and subjective symptoms. Aortic isthmus ratio ≤0.7 was the optimal cut-off point to detect high LV filling pressures defined as average E/e’ ≥14 with a sensitivity of 86% and specificity of 84% (area under the curve 0.704). Aortic isthmus ratio ≤0.07 could detect subjective symptoms (NYHA II-IV) with sensitivity of 88% and specificity of 81% (area under the curve 0.733), and impaired exercise capacity (peak VO2 <80%) with sensitivity of 92% and specificity of 47% (area under the curve 0.615).

Exploratory analysis

Of the 546 patients, 172 (32%) patients underwent COA intervention (stent n=44, surgery n=128), and 100 of the 172 patients had 5-year follow-up. The median aortic isthmus ratio at the time of intervention was 0.6±0.1. Compared to LV diastolic function indices prior to intervention, there was a net increase in LV e’ (1.5 [95%CI 1.3 – 1.7] cm/s, and decrease in LV E/e’ (−1.3 [95%CI −2.1 - −0.06]. Similarly, there was a net reduction in the number of patients with NYHA II-IV symptoms from 92 (92%) pre-intervention to 18 (18%) post-intervention. Temporal change in peak VO₂ was not assessed because of insufficient exercise data at 5 years post intervention

DISCUSSION

COA results in LV pressure overload, and LV hypertrophy occurs as an adaptive response to maintain wall stress.⁷ Chronic LV pressure overload leads to an increase in cardiac myocyte size and density, fibrosis and alteration of the viscoelastic properties and the effective operative compliance of the LV resulting in an increase in LV filling pressures and symptoms.⁷ The goal of COA intervention is to relieve LV pressure overload and prevent irreversible myocardial damage.^{9, 10} The American College of Cardiology and the European Society of Cardiology have proposed a number of COA severity indices that should prompt a referral for intervention.^{9, 10}These indices can broadly be classified as metrics of physiologic severity (Doppler mean gradient, SBP, ULE-SBP, or EIH) or anatomic severity (aortic isthmus ratio). Unfortunately, there are no data assessing the relationship between these indices of COA severity and LV diastolic function and symptoms. The current study showed that aortic isthmus ratio had the best correlation with diastolic function (average e’), LV filling pressure (E/e’), NYHA functional class and peak VO₂. Doppler mean gradient had some correlation with e’ and NYHA functional class but the other COA severity indices had no correlation with LV diastolic function indices or exertional symptoms.

LV diastolic dysfunction causing diastolic heart failure is responsible for half of all heart failure admissions in the United States.^{23, 24} The most common risk factors for diastolic dysfunction are LV pressure overload and ischemic heart disease, and these two factors are common in adults with COA.^{15, 25} In a case-control study, Lombardi et al⁸ demonstrated that LV diastolic dysfunction was more prevalent in COA patients compared to matched controls, and diastolic function indices correlated with aortic stiffness. COA patients had more aortic stiffness which resulted in higher arterial load on the LV.⁸ The etiology of LV pressure in the COA population is multifactorial, and results from a combination of mechanical obstruction at the aortic isthmus and vascular and endothelial dysfunction resulting in systemic hypertension.^{2, 4} Surgical and transcatheter therapies are both effective for treatment of COA.^{2, 21} However, these procedures are associated with some risk of complications, and some patients will require reinterventions, which further compounds the risk of morbidity and mortality. As a result, appropriate patient selection is important in order to optimize the benefits of these procedures.

The current standard of care is to refer patients with COA for intervention if they are considered to have hemodynamically significant COA based on the COA severity indices recommended by the practice guidelines.^{9, 10} However, it remains unclear how many of the severity indices have to be present to prompt a referral. The current study showed that aortic isthmus ratio, which is a measure of anatomic severity, had the best correlation with LV myocardial relaxation and filling pressures. Since LV filling pressure predicts heart failure symptoms,²⁴ we postulate that aortic isthmus ratio is better predictor of disease severity and symptomatology. Consistent with our finding, other studies have shown that aortic isthmus size was associated with LV hypertrophy and was also predictive of cardiovascular adverse events.^{6, 26}

Clinical Implications

Based on the results of the current study demonstrating strong correlation between aortic isthmus ratio and LV diastolic dysfunction and exertional symptoms, as well as the results of previous studies showing the prognostic importance of aortic isthmus size on cardiovascular outcomes during pregnancy, we expect that our results will simplify (at least to some extent) the risk stratification in this population. Aortic isthmus ratio can easily be obtained from cross-sectional imaging, and is non-invasive. It is not subject to the limitations of using Doppler COA gradients in patients with long segment stenosis, suboptimal acoustic windows and low cardiac output states, as these conditions increase the margin of error with these techniques. Doppler COA gradients, SBP, and ULE-SBP gradients are still very important, and hence will provide complementary data to guide management. Based on the current data, aortic isthmus ratio ≤0.7 may be used as the threshold to make a diagnosis of hemodynamically significant COA, but further data are required to determine if using this threshold to guide timing of intervention will result in an improvement in clinical outcomes.

Limitations

Although we demonstrated a correlation between aortic isthmus ratio and LV diastolic dysfunction as well as exertional symptoms, the study design does not provide proof of causality. We controlled for potential confounders using appropriate statistical methods, but there may still be some residual confounders because this is a retrospective study. Further studies are required to determine if the aortic isthmus ratio at the time of COA intervention is predictive of event-free survival.

Conclusion

Of all the COA severity indices stipulated in the guidelines, aortic isthmus ratio before intervention or reintervention had the strongest correlation with LV diastolic function indices and exertional symptoms. The aortic isthmus ratio is a measure of the anatomic severity of COA, and hence a measure of LV pressure overload. Since LV diastolic dysfunction, typically precede heart failure symptoms, we anticipate that the results of this study will potentially simplify patient selection for COA intervention. Aortic isthmus ratio is independent of loading condition, and can be obtained by non-invasive cross-sectional imaging. Cross-sectional imaging is one of the routine imaging evaluations that are performed in this population, and provides information about the severity, length and location of COA, and also screen for thoracic aortic aneurysm and pseudo aneurysm that can occur in this population.

WHAT IS NEW?

• Of all the severity indices for coarctation of aorta (COA) stipulated in the guidelines, aortic isthmus ratio had the best correlation with left ventricular diastolic function as measured by indices of myocardial relaxation and filling pressures.

• Aortic isthmus ratio also had the best correlation with subjective exertional symptoms (New York Heart Association functional class) and objective exercise capacity (peak oxygen consumption)

WHAT ARE THE CLINICAL IMPLICATIONS?

• The current study suggests the use of aortic isthmus ratio for risk stratification in COA patients because it has a better correlation with left ventricular remodeling and exertional symptoms as compared to the other COA severity indices.

• Aortic isthmus ratio provides a measure of anatomic severity, and it is independent of loading conditions, and hence more reproducible.

• Assessment of aortic isthmus ratio requires cross-sectional imaging which provides other important clinical data such as the morphology of the aortic isthmus, thoracic aorta dimensions, and screening for pseudo-aneurysms.

Abbreviations:

COA

coarctation of aorta

LV

left ventricle

EIH

exercise-induced hypertension

SBP

systolic blood pressure

ULE-SBP

upper-to-lower-extremity SBP

NYHA

New York Heart Association

VO₂

oxygen consumption

CI

confidence interval