Left ventricular hypertrophy persists after successful treatment for coarctation of the aorta

Abstract

Objectives. To evaluate cardiac size and function in patients with coarctation of the aorta (CoA) before and after treatment. Design. Ventricular size and function were examined by 2- and 3-dimensional echocardiography, and concentrations of natriuretic peptides measured in 15 paediatric patients before repair, and one, 6, and 12 months thereafter. Controls comprised 15 children. Results. Before repair, mitral inflow velocities and left ventricular (LV) size and wall thickness were higher in patients. Thicknesses of interventricular septum and LV posterior wall decreased after repair but increased to initial level one year thereafter. The LV end-diastolic diameter remained larger than in controls despite successful repair. The size of right ventricle increased and levels of natriuretic peptides decreased during follow-up. Levels of natriuretic peptides correlated with the smallest diameter of CoA segment and diastolic indices of LV function. Conclusion. LV hypertrophy persists and LV size remains larger than in controls after successful repair even in normotensive patients with normal growth of CoA segment. This may be due to remodelling of ventricles and the aorta caused by CoA.

• Coarctation of the aorta
• three-dimensional echocardiography
• two-dimensional echocardiography
• natriuretic peptides

Coarctation of the aorta (CoA) causes pressure overload of the left ventricle (LV) and predisposes the patient to systemic hypertension and significant morbidity 1. The generally used indication for treatment of CoA is a systolic arm-leg gradient of 20 mm Hg or more. In the presence of large collaterals or severe LV dysfunction, a pressure gradient may be less than assumed from the degree of obstruction and therefore CoA needs to be repaired despite a gradient of less than 20 mm Hg. After repair of CoA, the adaptation of LV to previous pressure overload persists even in normotensive patients without significant blood pressure gradients 2, 3. Little is known, however, of the degree and timing of changes in LV size and function and of levels of natriuretic peptides in children after repair of CoA.

Natriuretic peptide system has an important role in regulation of fluid homeostasis, blood pressure, and vascular tone and growth. Ventricular wall stress stimulates synthesis and release of both atrial and brain natriuretic peptides, concentrations of which have been shown to increase in heart failure 4, 5. Serum levels of N-terminal pro-brain natriuretic peptide (NT-proBNP) correlate with LV size in LV volume overload caused by a patent ductus arteriosus 6.

LV systolic and diastolic performance is evaluated in clinical practice most commonly with two-dimensional (2D) echocardiography. The measures of systolic function generally used are ejection fraction and fractional shortening. Diastolic function is evaluated by Doppler echocardiography with assessment of mitral inflow signal 7–9. Three-dimensional (3D) echocardiography is a new, non-invasive method to assess the LV size and function, its major advantage being independence of LV geometry. The measurement of LV volumes in 3D echocardiography correlates closely with cardiac catheterization and magnetic resonance imaging 10.

The purpose of the present study was to evaluate cardiac function in children with CoA by utilizing 2D and 3D echocardiography, and measurement of serum levels of natriuretic peptides, and to prospectively evaluate changes in these parameters during one year after treatment with percutaneous and surgical techniques.

Methods

Study population

At the Hospital for Children and Adolescents, University of Helsinki, Finland, 34 paediatric patients were diagnosed with unoperated (native) CoA or recurrent CoA (reCoA) requiring intervention between June, 2003, and November, 2004. Of these 34, the 15 children followed up in this hospital were enrolled in this study. None of them had a duct dependent CoA. They were compared at baseline with a control group of 15 healthy volunteer children matched for age, gender, height, weight, and body surface area. Characteristics of the patient and control groups are in Table I. All the parents of the participants agreed to participate in this clinical trial approved by the hospital ethics committee and gave their written informed consent.

Table I.  Characteristics of patient with coarctation of the aorta (CoA) and control subjects. Values are expressed as median (range).

Group	N	Gender (m/f)	Age (years)	Weight (kg)	Height (cm)	BSA (m^2)
Patients	15	6/9	2.00 (0.01–14.65)	14.00 (3.36–56.60)	91.5 (49.5–164.0)	0.59 (0.20–1.57)
Controls	15	6/9	2.94 (0.07–15.05)	15.00 (3.46–62.50)	95.2 (53.0–173.0)	0.62 (0.22–1.75)

BSA = body surface area, f = female, m = male, N = number.

The patients were examined prior to percutaneous or surgical treatment of CoA, and one, six, and 12 months thereafter. Control children were examined once. All patients and controls underwent clinical cardiovascular examination and blood test sampling for measurement of natriuretic peptides at the times of echocardiographic examinations.

Cardiac catheterization

The children with CoA considered suitable for percutaneous dilatation of the CoA segment were taken to the cardiac catheterization laboratory. Cardiac catheterization was performed under general endotracheal anaesthesia with systemic heparinization. Patients underwent standard haemodynamic cardiac catheterization, angiography of the aortic arch, and angioplasty of the CoA segment when appropriate.

Surgery

The patients aged less than 6 months with native CoA and those judged by invasive or uninvasive methods to be unsuitable for percutaneous treatment were referred to surgical repair with resection of the CoA segment and end-to-end anastomosis.

Echocardiography

We used the Acuson Sequoia C256 echocardiography system (Siemens, Mountain View, CA, USA) for 2D imaging and for Doppler and M-mode measurements. The echocardiographic examination took place with the patient in the supine position or in left lateral semirecumbency. An electrocardiographic tracing was recorded simultaneously with the echocardiogram. Transducer frequency was 7 MHz or 5 MHz, either or both used for each patient, to provide optimal two dimensional imaging and Doppler echocardiographic recordings. Standard parasternal, apical and subcostal views were used. Data were saved on magneto-optic disks for later analysis. Both 2D and 3D dimensional echocardiographic examinations were carried out by a single observer (A.E.).

Two-dimensional echocardiography

In 2D echocardiography, we measured the smallest diameter of the CoA segment and evaluated the aortic valve morphology. LV diastolic function was studied from the mitral inflow signal. We measured early peak flow velocity (E wave), atrial peak flow velocity (A wave), and velocity time integrals of the early (Evti), atrial (Avti), and total (EAvti) mitral inflow, and calculated the ratio of the early to the atrial peak flow velocities (E/A) as well as the deceleration time of the early peak velocity of mitral inflow (E decel).

M-mode echocardiography was performed from the parasternal long axis view. End-diastolic and end-systolic dimensions of LV, end-diastolic thickness of the interventricular septum and that of LV posterior wall, and end-diastolic diameter of the right ventricle were measured and the corresponding z scores determined. Fractional shortening and ejection fraction of the LV were then calculated as indices of LV systolic function 11. The mean of measurements from three cardiac cycles for each participant was saved for analysis.

Three-dimensional echocardiography

Three-dimensional echocardiography was performed with TomTec computer software (TomTec Imaging Systems GmHb, Munich, Germany). A series of cross-sectional echocardiographic images resulted from the apical view by rotating freehand scanning. Image acquisition was triggered by electrocardiogram. In free-hand scanning, a sensing device determined and registered the position and orientation of the transducer during the acquisition process. A complete cardiac cycle of images was taken on a selected image plane. Images of nine planes were collected by rotating the transducer at apical position by hand in a semicircle of 180°. Scanning of the plane took place if the heart rate was within ±20 beats/min of average. Digitized images were saved in computer memory during acquisition. After this, the 3D dataset underwent a postprocessing procedure.

The 3D datasets were analyzed with a detached computer. Manual tracing of the endocardium was performed on the white side of the black-white boundary. Papillary muscles, if discontinuous with the myocardium, were included in the ventricular volume. End-diastolic volume was calculated from the frame at the beginning of the R wave on the electrocardiogram or from the last frame with the mitral valve still open. End-systolic volume was calculated from the frame with the smallest cavity size when the mitral valve was still closed. Time-volume curves obtained served to determine end-diastolic and end-systolic volumes, stroke volume, and ejection fraction. From the calculated first derivatives of these curves, we measured peak filling rate and time to peak filling rate as indices of diastolic function and peak ejection rate as an index of systolic function.

Serum natriuretic peptides

Serum samples were frozen at −20°C, and serum concentrations of N-terminal proatriopeptide (ANPN) measured by immunofluorometric assay. The reagents were manufactured by Medix Biochemica (Espoo, Finland) and the instruments by Delfia Research Fluorometer (Wallac, Turku, Finland). Serum concentrations of the NT-proBNP were measured by the electrochemiluminometric method. The reagent kit was manufactured by Roche (Mannheim, Germany), and the samples were analyzed at Limbach Laboratory (Heidelberg, Germany).

Statistical analysis

Analyses were performed with the Statistical Package for Social Science version 12.01 for Windows (SPSS Inc., Chicago, Illinois, USA). For variables derived from echocardiograms and blood samples, median and range were calculated. Because distribution of parameters tested by Kolmogorov-Smirnov′s goodness of fit test was not normal, the Mann-Whitney test was used for statistical analysis between groups, and the Wilcoxon signed rank test for analysis within groups. The patient group was compared with controls. During the one-, 6-, and 12-month follow-up, the results were also compared with those of the same patients at baseline. Spearman's correlation coefficient was used for calculating correlations between the smallest diameter of the CoA, blood pressure gradients, diastolic and systolic indices of LV size and function, and serum concentrations of ANPN and NT-proBNP peptide. The level of significance chosen was at p less than 0.05.

Results

No statistically significant differences emerged in heart rate between patients and controls. Eight patients had a bicuspic aortic valve. One patient had severe LV dysfunction and heart failure, and one was diagnosed with Turner's syndrome. One patient with reCoA had been successfully operated on for ventricular septal defect several years earlier. One patient was on medication for hypertension, six complained about headache or decreased exercise tolerance, and nine were asymptomatic. As for the control group, they showed no abnormalities in clinical examination or echocardiography.

Eleven patients had native CoA and four had reCoA. Nine children were treated with surgery, and six with balloon dilatation. 13 children completed the 6- and the 12-month follow-up (Table II).

Table II.  Systolic and diastolic blood pressures, findings in two- and three-dimensional echocardiography (2D and 3D echo), and serum levels of natriuretic peptides in controls and in patients at baseline, and 1, 6, and 12 months after repair of the coarctation of the aorta. Values are given as median (range).

Variable	Controls (n = 15)	Baseline (n = 15)	One month after repair (n = 15)	Six months after repair (n = 13)	One year after repair (n = 13)
Systolic BP (mm Hg)	97 (76–120)	123 (91–151) ‡	104 (73–118)	105 (83–117)	103 (79–116)
Diastolic BP (mm Hg)	56 (42–69)	65 (42–96)*	57 (31–75)	58 (47–64)	61 (51–69)
2D echo
RVEDD (SD)	−0.25 (−1.50–2.00)	−0.75 (−2.25–0.75)	−1.00 (−3.50–3.00)	0.75 (−2.00–2.00)§	0.50 (−2.25–1.75)§
LVEDD (SD)	−0.25 (−1.50–2.00)	0.75 (−1.25–4.25)*	1.50 (−1.25–3.50) †	1.00 (−1.00–2.50)	1.00 (−0.50–2.00)*
IVSED (SD)	−0.25 (−1.25–1.50)	1.25 (−0.75–5.00) †	0.50 (−1.25–3.00)	0.25 (−1.25–2.50) §	1.50 (−0.75–2.75) †
LVPWED (SD)	−0.50 (2.25–0.75)	0.00 (−2.50–1.50) †	−0.50 (−2.00–1.50)	−0.50 (−1.50–1.50)	0.75 (−0.75–1.75) ‡
E (m/s)	0.84 (0.64–1.06)	1.11 (0.71–1.50) †	1.10 (0.85–1.24) ‡	1.22 (1.09–1.44) ‡	1.17 (1.05–1.32) ‡
A (m/s)	0.63 (0.30–0.98)	0.83 (0.39–1.16)*	0.86 (0.47–1.66) †	0.97 (0.31–1.35) †	0.73 (0.36–1.16)
E/A	1.50 (0.81–2.98)	1.22 (0.66–3.39)	1.30 (0.74–2.40)	1.36 (1.06–4.39)	1.57 (0.98–3.64)
EAvti (m)	0.119 (0.067–0.209)	0.140 (0.052–0.252)	0.198 (0.078–0.229)	0.171 (0.115–0.325) ‡§	0.174 (0.114–0.278) †§
Evti (m)	0.077 (0.029–0.146)	0.095 (0.014–0.146)	0.122 (0.036–0.170)	0.103 (0.062–0.191) †	0.115 (0.062–0.183)*§
Avti (m)	0.033 (0.019–0.048)	0.049 (0.020–0.089) †	0.057 (0.037–0.123) ‡	0.068 (0.040–0.090) ‡ §	0.054 (0.040–0.070) ‡
E decel (ms)	90 (48–220)	106 (33–202)	115 (53–194)	105 (74–209)	116 (65–218)
FS (%)	37 (32–48)	42 (19–52)	44 (32–49)	43 (34–52)	43 (31–51)*
EF (%)	69 (60–82)	75 (41–83)	77 (62–81)	75 (62–84)	75 (59–84)*
3D echo
LVEDV/BSA (ml/m^2)	52 (36–77)	65 (24–145)	64 (33–107)	61 (31–131)	71 (36–121)
PFR (ml/s)	74 (25–142)	57 (11–720)	87 (7–239)	96 (22–256)	124 (61–378) †
EF (%)	49 (38–59)	48 (31–54)	47 (32–60)	52 (35–70)§	56 (40–68)*§
ANPN (nmol/l)	0.43 (0.15–1.20)	0.60 (0.16–8.24)	0.49 (0.14–1.33)	0.46 (0.09–0.76)	0.34 (0.09–0.82)
proBNP (ng/l)	81 (26–321)	128 (5–19132)	125 (4–7342)§	86 (9–358)§	60 (32–309)§

* p < 0.05, † p < 0.01, ‡ p < 0.001 as compared with controls.

§ p < 0.05, p < 0.01, ¶ p < 0.001 as compared with baseline.

Abbreviations: A = atrial peak flow velocity, ANPN = N-terminal proatriopeptide, BP = blood pressure, BSA = body surface area, E decel = deceleration time of mitral early diastolic flow, E/A = ratio mitral early peak flow velocity to atrial peak flow velocity, E = mitral early peak flow velocity, EF = ejection fraction, FS = fractional shortening, IVSED = z score of the end-diastolic thickness of interventricular septum, LVEDD = z score of the left ventricular end-diastolic diameter, LVEDV/BSA = the left ventricular end-diastolic volume adjusted to body surface area, LVPWED = z score of the end-diastolic thickness of left ventricular posterior wall, PFR = peak filling rate, proBNP = N-terminal pro-brain natriuretic peptide, RVEDD = z score of the right ventricular end-diastolic diameter, SD = standard deviation, vti = velocity time integral.

Systolic and diastolic blood pressures, measurements obtained in 2D and 3D echocardiography, and serum concentrations of natriuretic peptides in the patient group and in controls are listed in Table II. Systolic and diastolic blood pressures were higher in the patient group.

Eight children underwent standard haemodynamic cardiac catheterization and angiography of the aortic arch: five of them had native CoA, and three had reCoA. Percutaneous balloon angioplasty was performed in six patients, one of them with stent implantation. In these patients, the invasive blood pressure gradient measured a median 30 (range 22 to 52) mm Hg before and 11 (range 0 to 16) mm Hg (p = 0.027) after the procedure. In angiography, the smallest diameter of the CoA segment measured a median 4.0 (range 2.5 to 7.0) mm before and 7.3 (range 4.4 to 12.5) mm (p = 0.028) after dilatation.

A total of nine patients underwent surgical repair with resection and end-to-end anastomosis. Length of hospital stay was a median one day (range 1 to 2) days for patients treated with percutaneous dilatation and six (range 5 to 20) days for those repaired surgically. No mortality was associated with percutaneous or surgical treatment.

Prior to repair, one child was on medication for hypertension. After repair, five patients were on medication: four for hypertension, and one for LV dysfunction. Antihypertensive medication was discontinued before the first follow-up visit in one patient and after the first visit in three. After one month's follow-up visit, the only patient on medication was the one with LV dysfunction which gradually improved, and she was weaned off medication by the end of the one-year follow-up.

The systolic arm-leg blood pressure gradient measured a median 30 (range 14 to 56) mm Hg at baseline. It had decreased to a median 0 (range 0 to 21) mm Hg (p = 0.001) by the time of the first follow-up visit. The gradient measured a median 5 (range 0 to 16) mm Hg (p = 0.001) 6 months, and 0 (range 0 to 12) mm Hg (p = 0.001) one year after repair.

Prior to repair, the z score of the LV end-diastolic diameter was higher in the patient group and remained so even a year after repair with no decrease during the follow-up (Table II). At baseline, the z scores of end-diastolic thicknesses of interventricular septum and LV posterior wall were higher than in controls. They decreased up to 6 months after the procedure but increased again to initial levels by the end of the one year follow-up. (Table II, Figure 1). No significant changes were seen during the one-year follow-up in the indices of LV systolic function as measured by M-mode echocardiography. In patients with CoA, the z score of the RV end-diastolic diameter increased during follow-up (Table II).

Figure 1.  Z scores of end-diastolic thicknesses of A. interventricular septum (IVSED) and B. LV posterior wall (LVPWED) in patients with coarctation of the aorta at baseline, and 1 month, 6 months, and 12 months after treatment, and in control children.* p < 0.05, † p < 0.01, ‡ p < 0.001 as compared with controls§ p < 0.05, p < 0.01, ¶ p < 0.001 as compared with baseline.

In Doppler echocardiography of the mitral inflow signal, the E and A waves and the Avti were all higher in the patient group at baseline. After successful repair, Evti and EAvti of mitral flow increased significantly as compared with baseline (Table II). These changes may be due to persistent abnormality of LV diastolic function as compared with controls.

In 3D echocardiography, LV ejection fraction increased during the one-year follow-up. At the end of the follow-up, LV peak filling rate was higher than in controls (Table II).

Serum levels of natriuretic peptides decreased significantly during the follow-up in patients with CoA (Table II). They correlated with the degree of aortic flow obstruction and with indices of LV compliance: Levels of ANPN and NT-proBNP had a negative correlation with the smallest diameter of CoA segment in 2D echocardiography (r = − 0.787, p = 0.001 and r = − 0.784 and p = 0.001, respectively), with E/A ratio in Doppler echocardiography (r = − 0.717, p = 0.003 and r = − 0.614, p = 0.015, respectively), and with peak filling rate in 3D echocardiography (r = − 0.887,p ≤ 0.001 and r = − 0.882, p = < 0.001, respectively).

Discussion

CoA is a lifelong disease process. The most common cardiac malformations associated with CoA include bicuspid aortic valve, mitral valve abnormalities, and ventricular septal defects. CoA is more common in patients with Turner's syndrome than in normal population. Therefore, our patient population represents a typical cohort of CoA patients. Patients who have undergone repair of CoA are at risk for reCoA, arterial hypertension, premature atherosclerotic disease, heart failure, cerebrovascular accidents, rupture of an aortic aneurysm, and sudden death even after successful treatment 12–14. Aortic remodelling has been shown to occur in patients after angioplasty for CoA 15. In our study, significant reduction in pressure gradient across CoA occurred after percutaneous and surgical repair.

Adult populations have shown increased morbidity and mortality due to arterial hypertension and atherosclerotic disease despite successful repair of CoA 13, 14. Prevalence of arterial hypertension has also been detected in children after CoA repair 14, 16. In our study, at baseline, systolic blood pressure was higher in patients than in control children, but after repair, no differences were detected in systolic or diastolic blood pressures. Risk for reCoA after surgery in the neonatal period has been up to 19%, and after the neonatal period up to 3% in the mean follow-up of 4 years 17, 18. Risk for reintervention after balloon dilatation has varied up to 80% in the neonatal and infant period and up to 13% in children aged over one year with up to 9 years’ follow-up 19, 20. None of our patients developed reCoA during their one-year follow-up.

Evaluation and follow-up of LV diastolic function in patients with CoA is important, since diastolic dysfunction may precede systolic dysfunction in patients with heart failure 21. Because diastolic indices of LV function are age and heart rate-dependent 22, we chose age- and gender matched controls. In our study, differences were detectable in the indices of diastolic LV function between patients and controls at baseline. That increased afterload causes impaired relaxation and decreased compliance of LV 23, 24 explains why the E and A waves of mitral inflow were higher in our patients than in controls at baseline. Similar findings have appeared in patients late after CoA repair 25, 26.

Although the thicknesses of the interventricular septum and the LV posterior wall were within normal limits in most patients, as a group they differed from controls at baseline. The thicknesses of the interventricular septum and LV posterior wall decreased up to 6 months after the procedure. However, one year after repair, they had increased again, although the patients were normotensive at rest. Animal studies suggest that after CoA repair, humoral factors may be involved in cardiac hypertrophic response 27. Systolic and diastolic hypertension in ambulatory blood pressure measurement, impaired flow-mediated vasodilatation, and increased thicknesses of intima and media of carotid artery have been reported to occur years after CoA repair in children and young adults 14, 28, 29. In our patients, the z score of the right ventricular end-diastolic diameter increased during follow-up due to remodelling of the ventricles after cessation of LV pressure overload. No differences were detected in LV systolic function at baseline between groups. In 3D echocardiography, ejection fraction increased after repair. LV contractility in 2D and 3D echocardiography was higher in patients with CoA after one-year follow-up. Similar findings with hyperdynamic and hypertrophied LV have been reported several years after operation for CoA, suggesting that some degree of LV remodelling takes place after CoA repair, but the LV geometry and hypertrophy do not normalize entirely 2.

In paediatric patients, plasma concentrations of NT-proBNP 30 and natriuretic peptide type B have been shown to correlate with clinical signs of heart failure but in children with LV pressure overload, the levels of natriuretic peptides have been almost equal to control levels 31, 32. Similarly, in our study, no difference was seen in serum levels of natriuretic peptides between patients and controls at baseline. However, levels of natriuretic peptides decreased significantly in patients after treatment. In addition, serum levels of natriuretic peptides correlated with the smallest diameter of CoA segment and diastolic indices of LV function. Based on our results, we recommend for a paediatric population that measurement of serum levels of natriuretic peptides be included in follow-up before and after treatment of CoA. They can be used as individual indicators of response to treatment.

Limitations

Due to a small number of patients in this study, it was not possible to analyse separately the patients treated with surgery and those treated with balloon dilatation. Longer follow-up may be needed to see if LV hypertrophy resolves eventually.

The age range in our study was wide because the study group comprised of both patients with native CoA and those with reCoa. Therefore, at baseline, patients were compared with age matched controls. During the follow-up they were compared with both control group and with themselves at baseline. Most of the parameters used for follow-up were indexed for body surface area.

Conclusion

CoA causes significant changes in diastolic and systolic LV function. Remodelling and growth of the CoA segment takes place months after repair. After repair, the right ventricular size and the velocity time integrals of mitral inflow increase. Serum levels of natriuretic peptides decrease during follow-up after repair of CoA. Close follow-up is needed despite cessation of LV pressure overload since LV hypertrophy and dilatation and abnormal diastolic function persist even in normotensive patients with growth of the CoA segment.

Acknowledgements

This study was supported by grants from the Foundation for Paediatric Research, Helsinki, Finland, and the Finnish Foundation for Cardiovascular Research, Helsinki, Finland, the special governmental subsidy for health sciences research, Helsinki, Finland.