Rationale and design of Long-term Outcomes and Vascular Evaluation after Successful Coarctation of the Aorta Treatment study.

BACKGROUND: Coarctation of the aorta (CoA) can be treated using surgery, balloon angioplasty, or stent implantation. Although short-term results are excellent with all three treatment modalities, long-term cardiovascular (CV) morbidity and mortality remain high, likely due to persistently abnormal vascular function. The effects of treatment modality on long-term vascular function remain uncharacterized. The goal of this study is to assess vascular function in this patient population for comparison among the treatment modalities.
METHODS: We will prospectively assess vascular Afunction in large and small arteries fusing multiple noninvasive modalities and compare the results among the three groups of CoA patients previously treated using surgery, balloon angioplasty, or stent implantation after frequency matching for confounding variables. A comprehensive vascular function assessment protocol has been created to be used in 7 centers. Our primary outcome is arterial stiffness measured by arterial tonometry. Inclusion and exclusion criteria have been carefully established after consideration of several potential confounders. Sample size has been calculated for the primary outcome variable.
CONCLUSION: Treatment modalities for CoA may have distinct impact on large and small arterial vascular function. The results of this study will help identify the treatment modality that is associated with the most optimal level of vascular function, which, in the long term, may reduce CV risk.

INTRODUCTION

Current treatment techniques are equally effective at eliminating the stenosis in CoA patients.[1] However, a good anatomical result does not preclude late systemic hypertension in office visits (12%–65%),[123456] at peak exercise (10%–47%),[34578] or during ambulatory blood pressure (BP) monitoring (30%–59%).[78910] Furthermore, treated patients have reduced life expectancy [Figure 1],[2] mostly due to cardiovascular (CV) complications[211121314] and stroke.[15]

Figure 1

Survival after treatment of coarctation of the aorta. Survival curves of 819 surgical patients for over 60 years (reprinted with permission from Elsevier, license number 4131890880395)

Successfully treated CoA patients have stiffer large arteries[161718192021] and compromised vascular reactivity in small arteries,[8102223242526] their arterial pressure waveform is altered,[910232728] have imbalances in vascular function biomarkers,[2425272930] and increased left ventricular (LV) mass.[89192021263132] Vascular dysfunction is associated with older age at treatment,[21922293133] but early treatment does not guarantee normal vascular function.[1622]

Different treatment modalities may have varying effects on the stiffness of the repaired arterial segment:[34] Surgical repair results in a focal scar in the anastomosis; stenting creates a short, rigid segment; and balloon dilation (BD) produces a controlled tear of the intima and part of the media. Although it is possible that these differences translate into differences in vascular dysfunction, this has not been systematically compared. The largest, albeit observational and nonrandomized, comparison between the three modalities showed a lower BP in patients treated with BD versus those treated with stenting or surgery.[1] A small retrospective study showed less frequent exercise-induced hypertension in BD patients compared with other treatment types.[33] Conclusions drawn from these prior studies are hampered by methodological limitations and limited focus. In the general population, arterial stiffness is associated with major CV events.[35] Thus, choosing the CoA treatment option that optimizes vascular function is crucial for long-term outcomes in CoA.

Aim and hypothesis

The Long-term Outcomes and Vascular Evaluation after Successful Coarctation of the Aorta Treatment study aims to determine whether surgery, BD, and stenting are associated with differences in arterial stiffness in optimally treated patients. Our hypothesis is that patients who underwent successful BD will have better vascular function than patients who underwent successful surgical repair or stenting since this modality may least likely damage the biomechanical properties of the aortic wall.

METHODS

Study overview

This study is a cross-sectional prospective observational study of patients with CoA previously treated using one of three treatment modalities. Patients will be recruited at seven large pediatric cardiac centers from Europe and the United States of America [Appendix 1]. The study procedures will occur in a 1- or 2-day visit [Figure 2].

Figure 2

Long-term Outcomes and Vascular Evaluation after Successful Coarctation of the Aorta Treatment study workflow. ABPM: Ambulatory blood pressure monitoring, AIx: Augmentation index, BP: Blood pressure, CMR: Cardiac magnetic resonance imaging, PWA: Pulse wave analysis, PWV: Pulse wave velocity, RHI: Reactive hyperemia index

Recruitment

Selection criteria are depicted in Table 1. The study protocol was approved by Institutional Review Boards. Recruitment occurred between June 2013 and December 2017. The study data are collected and managed using REDCap software, hosted at Children's Hospital Boston.[36]

Table 1

Inclusion and exclusion criteria

Study procedures

A list with the main clinical and study tests variables are depicted in Tables 2 and 3. The comprehensive list of study variables is in Appendix 7

Table 2

List of main clinical variables

Table 3

List of main study test variables

Arterial stiffness

CoA treatments alter the biomechanics of the isthmus and may increase arterial stiffness. The velocity of the pulse wave velocity (PWV) travel in the arterial tree increases with arterial stiffness. Carotid–femoral PWV (cfPWV) is extensively validated in large studies a marker of aortic stiffness, and an independent predictor of CV events.[37] We will measure cfPWV with applanation tonometry, using either the NIHem (CV Engineering, Inc., Norwood, MA USA) or the SphygmoCor (AtCor Medical, West Ryde, NSW, Australia) devices.[37] This technique assumes a homogenous stiffness across the aorta and may potentially not accurately estimate the true carotid-to-femoral artery length. Cardiovascular magnetic resonance (CMR) measurements of PWV, on the other way, enables the detection of more subtle changes in segmental aortic PWV, above versus below the CoA site, and uses real aortic travel paths.[38] We will also use CMR to measure aortic area change during the cardiac cycle, paired with BP measurements, to quantify local arterial strain, compliance, distensibility, and the β-stiffness index [Appendix 2a and b].

Endothelial function

In CoA, the loss of central aortic pulsatility, which buffers systole, generates chronic shear stress downstream in smaller arteries, creating endothelial dysfunction, which is associated with CV events.[39] We will measure endothelial function with the reactive hyperemia index using endothelial pulse amplitude tonometry (endo-PAT), a novel noninvasive and reproducible technique that measures changes in pulsatile arterial volume with a fingertip probe.[40] Analysis of the pulse waveform allows for automated calculation of endothelial function in one arm, while the contralateral serves as control, making this is a patient standardized method [Appendix 3].

Pulse waveform analysis

In CoA, the stiff aorta and repaired isthmus may be important reflecting sites that impact the pulse waveform. Its analysis is an important clinical tool for monitoring of vascular function and predicting CV events.[37] We will measure three variables that express pulse waveform: central aortic pressure (CAP), pulse pressure (PP), and augmentation index (AIx; ratio of the amplitude of the reflected wave in the ascending aorta and the PP).[37] CAP, PP, and AIx can be measured noninvasively using applanation tonometry (and Endo-PAT for AIx), calibrated by the peripheral diastolic and mean arterial pressure.[37] There is a lack of consensus regarding the optimal method to estimate the CAP with tonometry. The NIHem system assumes that carotid artery pulse waveform accurately reflects the central aortic waveform and the pulsed wave analysis is automatically calculated from the carotid waveform. The SphygmoCor device uses a generalized transform function to generate a central aortic PP curve from the radial or carotid pressure tracings, which has not been validated in children. Considering our largely pediatric group and need to maintain consistency between data acquired on each device, we use the nonprocessed, signal-averaged SphygmoCor carotid tracing as the central aortic tracing which will be then digitized to calculate the CAP, following previously published approach [Appendixes 2a and 3].[41]

Blood pressure phenotype

BP phenotype is abnormal despite successful treatment of CoA. Office hypertension is a known risk factor for CV disease and the BP response during the ET is predictive of future development of resting hypertension in the general population.[42] Ambulatory blood pressure monitoring (ABPM) is superior to the office measurement in its ability to distinguish patients at the highest risk for target-organ damage.[43] We will assess BP phenotype with the manual auscultation technique to measure the right arm office BP; supine four extremity oscillometric BP measurement to assess for residual coarctation; ABPM to measure the circadian BP profile; and ET to assess the BP response to exercise and exercise-induced arm to leg BP gradient. Based on the office BP and ABPM results, we will classify our patients according to Table 4 [Appendix 4].

Table 4

Classification of BP Phenotype by ABPM

Biomarkers

We will measure asymmetric dimetilarginine (ADMA; NO's inhibitor),[44] and nitrite and nitrate (NOx, stable by-product of NO), biomarkers of endothelial function. Arterial stiffness is associated with increased systemic inflammation markers, which we will quantify with high-sensitivity C-reactive protein (hs-CRP) and local inflammatory cytokines of vascular wall function vascular adhesion molecule 1 (VCAM-1) and interleukin-1 beta (IL-1β).[2545] We will finally assess the molecular mechanisms of aortic wall response to vascular dysfunction, with matrix metalloproteases (MMP-2 and MMP-9),[46] and transforming growth factor beta-1 (TGF-β1, a smooth cell growth-modulating factor involved in the arterial wall response to hypertension).[30] NOx will be determined by chemiluminescence (Sievers NOAnalyzer 280i) and all remaining measurements will be performed with enzyme-linked immunosorbent assay kits: ADMA (Sunred Biological Technology, Shanghai, China); hs-CRP (BoosterBio, Pleasanton, USA); VCAM-1; IL-1β; matrix metalloproteases (MMP)-9; MMP-2; and TGFβ-1 (RayBiotech, Inc. Norcross, USA) [Appendix 5].

Left ventricular mass

The altered BP phenotype that persists after CoA treatment represents an increase in afterload that leads to LV hypertrophy, strongly related to high BP and carrying a grave prognosis for cardiac events.[47] We will quantify LV mass by CMR, a well-established method for its calculation [Appendix 2b].

Cardiovascular health assessment

Patients with CoA experience increased CV disease compared to the general population. Literature in the general population has demonstrated that risk of cardiometabolic disease and accelerated atherosclerosis is mitigated by ideal CV health (ICVH),[48] defined as having optimal levels of health factors (BP, total cholesterol, and plasma glucose) and behaviors (smoking, body mass index, physical activity, and diet). We will implement a questionnaire to assess family history of CV disease and ICVH according to the guidelines of the American Heart Association [Appendix 6].[48]

Statistical considerations

Adjustment for confounders

We will adjust our treatment groups for three main documented confounders: (a) age at treatment; (b) current age; and (c) bicuspid aortic valve (associated with impaired aortic elasticity).[49] During recruitment, we will attempt to frequency match the three treatment groups. During analysis, the treatment groups will be compared for each of these three confounding variables and adjustments will be made using multivariable modeling with linear and logistic regression models.

Analytic plan

Our primary outcome variable will be cfPWV assessed by tonometry. Differences across groups will be explored using one-way analysis of variance. If differences in matching variables are detected among the groups, adjustment will be made using analysis of covariance. Post hoc analyses will be performed as necessary. Sample size estimates were obtained based on prior data that show that arch PWV measured by CMR is 3.3 ± 0.6 m/s in normal patients and 4.7 ± 1.1 m/sec after CoA surgery.[2050] Sample size estimates for comparison of PVW between three equal-sized treatment groups (assuming overall significance level = 0.05 and power = 0.8) are shown in Table 5. We plan on recruiting 24–30 patients in each group for a total sample size of 72–90.

Table 5

Sample size estimation

DISCUSSION

Methodological considerations

We chose a multicenter design to overcome recruitment challenges secondary to restrictive enrollment criteria (particularly the lower treatment age limit of 1 year, which excludes a majority of CoA patients that present in infancy, mostly managed by surgery) and need for matching treatment groups for confounders.

cfPWV is our primary outcome variable because it is validated as an accurate and reproducible measure of arterial stiffness with proven association to hard CV outcomes that can be reliably measured by applanation tonometry and CMR. We chose other parameters to complete a complementary and comprehensive assessment of vascular function in small and large arteries.

Importance of knowledge to be gained

This work will be the first systematic and comprehensive comparison of vascular function between three different treatment modalities in CoA patients. We postulate that the integrity of the arterial wall is best preserved with balloon dilatation, compared to stenting or surgery. We are aware that our population is highly selected, but believe that this is the only way to compare the three treatment types. The results of our selected population may be relevant when several modalities are applicable to one patient. Currently, the preservation of vascular function is not considered when choosing between treatment modalities. Ultimately, the results of our study may help clinicians choose treatment modalities based not only on relief of anatomic stenosis but also on their ability to preserve long-term vascular health.

Study limitations

Our results will reflect vascular function in a selected group of optimally treated CoA patients and may not be generalizable to all CoA patients. We will compare vascular function after treatment but not before the treatment. Variation in antihypertensive medication protocols between different institutions may affect vascular parameters.

CONCLUSION

There is ample evidence to suggest that CoA is a systemic arterial disease and not merely a focal stenosis of the aortic isthmus. However, the current management paradigm continues to focus on alleviating the anatomic stenosis. Our study aims to refine this treatment paradigm by adding the preservation of vascular function to the goals of successful treatment. The strengths of this study include its multicenter design and the use of multiple noninvasive modalities to perform a comprehensive and prospective assessment of vascular function and CV health.

Financial support and sponsorship

The study was supported by two grants, from the Millennium Foundation BCP (awarded on December 18, 2012) and the Fundação Luso-Americana para o Desenvolvimento (Proj A-1 287/2012).

Conflicts of interest

There are no conflicts of interest.