Quantitative Echocardiographic Analysis of the Aortic Arch Predicts Outcome of Balloon Angioplasty of Native Coarctation of the Aorta
Background The use of balloon angioplasty for treatment of native aortic coarctation is controversial. Cineangiographic data suggest that aortic arch hypoplasia and isthmic narrowing are associated with angioplasty failure. This study of echocardiographic measurements of preangioplasty aortic arch morphology was performed to identify potential anatomic predictors of outcome noninvasively.
Methods and Results The preangioplasty echocardiograms of 105 patients 3 days to 17 years old with native coarctation of the aorta were analyzed off-line. Angioplasty was considered successful if the residual coarctation gradient was <20 mm Hg and no intervention for recoarctation occurred. Univariate analysis identified young age at angioplasty, presence of a patent ductus arteriosus, and the diameters of the aortic isthmus, distal transverse arch, and aortic valve as predictors of early and late outcomes. Multivariate analysis showed that the preangioplasty aortic isthmus z value was the best independent predictor of outcome, eliminating the effect on outcome of age and associated cardiac defects. An isthmus z value ≤−2.16 predicted early failure with 91% sensitivity and 85% specificity. Kaplan-Meier analysis demonstrated that 90% of patients with an isthmus z value >−1.0 remained free of recoarctation at late follow-up, whereas 89% of patients with a preangioplasty isthmus z value ≤−2.0 developed recoarctation within 36 months.
Conclusions Echocardiographic measurements of the aortic arch predict both early and late outcomes of balloon angioplasty for native aortic coarctation, and the preangioplasty aortic isthmus z value was the best independent predictor. Quantitative aortic arch analysis may improve selection of angioplasty candidates who are likely to benefit from the procedure.
Percutaneous BA for native coarctation of the aorta remains a controversial treatment option despite its increased use over the past decade.1 2 3 4 5 Early success rates for BA have ranged from 70% to 100%, but only 64% to 77% of patients are reportedly free of recoarctation at intermediate follow-up.6 7 8 9 10 11 12 BA has several advantages over surgical treatment for native coarctation of the aorta.3 10 12 Recent reports suggest that freedom from recoarctation at intermediate follow-up occurs in 75% to 80% of neonates and young infants treated by surgery.4 13 14 15 16 Enthusiastic reports of effective BA in patients with isolated coarctation are tempered by concerns about safety (such as femoral vessel obstruction17 and development of aneurysm at the coarctation site) and long-term efficacy.4 9
Factors that may be associated with outcome of BA for native coarctation of the aorta include age,9 10 12 18 associated cardiac defects,9 and aortic arch anatomy.3 6 8 10 Specifically, tubular narrowing and hypoplasia of the aortic isthmus are associated with poor outcome in several reports.3 6 8 9 10 11 18 19 20 21 In all studies, the measures of aortic arch anatomy are derived from cineangiography exclusively, yet these data do not allow precatheterization selection of patients who are most likely to benefit from BA. The purpose of this study is therefore to identify quantitative echocardiographic predictors of BA outcome in patients with native coarctation of the aorta.
One hundred twenty patients identified by a computerized search of the interventional database underwent BA as an initial intervention for native coarctation of the aorta between November 1, 1983, and March 31, 1994, at Texas Children's Hospital. Patients with a complete and technically adequate preangioplasty echocardiogram were included in this study. One hundred five patients met the inclusion criteria. All inpatient and outpatient records were reviewed for physical examination, echocardiographic, catheterization, MRI, and surgical data to determine patient clinical characteristics, demographics, and treatment outcome. Clinical and demographic information included patient age, sex, weight, height, and BSA at echocardiography and catheterization and all associated cardiac defects as determined by clinical, echocardiographic, catheterization, MRI, and surgical assessment. Indications for BA were variable but generally included angiographic evidence of aortic isthmic coarctation or a pressure gradient >20 mm Hg from ascending to descending aorta or both at cardiac catheterization.
Patients were categorized under early angioplasty success if (1) peak systolic coarctation gradient measured at catheterization was reduced to <20 mm Hg postangioplasty and (2) no intervention for recoarctation occurred within 1 month of the procedure. Patients who achieved early angioplasty success were categorized under angioplasty success at most recent follow-up if (1) resting arm-to-leg systolic blood pressure gradient was <20 mm Hg at last follow-up3 4 6 12 and (2) no intervention for recoarctation occurred during follow-up.
Images were obtained from the subxiphoid, apical, parasternal, and suprasternal notch views and recorded on 1.27-cm super VHS videotapes. Sedation with chloral hydrate (75 to 100 mg/kg; maximal dose, 1 g) was used when necessary in young, uncooperative patients. The echocardiograms were reviewed by an observer who was blinded to the results of cardiac catheterization, and selected still frames were identified for subsequent measurements. Each measurement was obtained in triplicate, and the average value was used for data analysis. Hard copies were obtained with a video page printer (Sony Videographic Printer UP-910), and measurements were performed with a digitizing tablet (Summagraphic II) attached to a personal computer (Compaq 386 SX) with commercially available software (Digisonics EchoPro 3.3).
The following measurements were obtained from video frames showing maximal dimension in systole: Aortic valve annulus and aortic root diameters were measured from the parasternal long-axis imaging plane. The aortic valve annulus was measured between the hinge points of the right coronary and the noncoronary aortic valve leaflets.22 The aortic root was measured as the widest diameter of the sinuses of Valsalva between the aortic valve annulus and the sinotubular junction. The ascending aorta, aortic arch, aortic isthmus, and proximal descending aorta were imaged from high parasternal and/or suprasternal notch windows,23 and measurements were obtained as shown in Fig 1⇓. The ascending aorta diameter was measured just proximal to the innominate artery. Distal transverse aortic arch diameter was measured at the base of the left common carotid artery. Aortic isthmus diameter was measured at the base of the left subclavian artery. The coarctation site diameter was measured at the point of maximal aortic isthmic narrowing. Distal transverse aortic arch length was measured from the midpoint of distal transverse arch diameter to the midpoint of aortic isthmus diameter. The aortic isthmus length was measured from the midpoint of aortic isthmus diameter to the midpoint of coarctation site diameter. When the coarctation site involved the base of the left subclavian artery, the diameters of the isthmus and the coarctation were considered to be the same. Descending aorta diameter was measured from the subxiphoid sagittal imaging plane at the level of the diaphragm. LV end-diastolic dimension and LV posterior wall thickness were measured from two-dimensionally directed M-mode tracings of the LV short axis in diastole.
Linear dimensions were indexed to the square root of BSA24 to allow comparison of measurements between patients of various sizes. Corresponding z values were calculated for the diameters of the aortic valve annulus, distal transverse arch, aortic isthmus, and LV dimensions. z values were computed as follows: z=(measured value − mean value of normal control) ÷ SD of normal control.
The normative echocardiographic data are derived from aortic arch measurements of 140 children 0 to 18 years old with no heart disease (courtesy of Steven D. Colan, MD, Children's Hospital, Boston, Mass). Maximal instantaneous gradient across the coarctation site was calculated from the peak spectral Doppler velocity with the modified Bernoulli equation.25
To determine interobserver variability, measurements were performed in a group of randomly selected patients by two investigators who were blinded to each other's measurements and to the results of cardiac catheterization. Simple linear regression analysis was used to calculate the correlation of measurements by the two observers. The absolute difference between observers' measurements was divided by the mean value of measurements and expressed as a percentage. Interobserver variability was expressed as the mean value (±SD) of these percentages.
The BA technique used at our institution for native coarctation of the aorta has been described in detail.7 10 Balloon size was selected such that the inflated balloon diameter was 0 to 2 mm greater than the aortic isthmus at the base of the left subclavian artery.10 In addition to hemodynamic and anatomic findings, preangioplasty and postangioplasty peak systolic coarctation gradient, maximum balloon diameter, maximum inflation pressure, number of inflations, operator, and calendar year of procedure were recorded.
When appropriate, data are expressed as mean±SD. Unpaired Student's t test was used to compare continuous variables between groups. Categorical variables were compared by contingency tables by use of χ2 analysis or Fisher's exact test. Variables found to be significantly associated with outcome by univariate analysis were evaluated simultaneously by multiple logistic regression analysis to identify independent predictors of outcome at early and most recent follow-up. Simple linear regression by the least-squares method was performed to identify significant correlation between continuous variables. Prediction models for early and follow-up outcomes were constructed by logistic regression analysis to test the interaction between specific echocardiographic variables and patient outcome. Standard definitions for calculation of specificity, sensitivity, accuracy, and predictive values26 were used to evaluate the ability of regression models and specific echocardiographic variables to predict outcome. Kaplan-Meier life-table analysis was used to compare outcome over the entire follow-up period for patients categorized by specific echocardiographic variables. A value of P<.05 was considered significant.
Early outcome of BA and outcome at most recent follow-up are illustrated in Fig 2⇓. Demographic and clinical data of the patients are summarized in Table 1⇓. Of the 105 patients included in this study, 83 (79%) achieved early angioplasty success and 22 (21%) were early angioplasty failures. Of the 22 early failure patients, 14 (64%) had postangioplasty peak systolic gradients ≥20 mm Hg and 8 (36%) required surgery or repeat angioplasty within 1 month of the initial procedure. Of the 83 patients who achieved early success, 75 (90%) had adequate follow-up data to determine outcome at most recent follow-up. Fifty-four patients (72%) had continued success at follow-up, but 21 (28%) were considered angioplasty failure, with a resting arm-to-leg systolic blood pressure gradient ≥20 mm Hg. Fifteen of 21 patients required an intervention for recoarctation, either surgery (n=8), repeat angioplasty (n=5), or both (n=2), at 14±12 months after the initial procedure. Patients classified as early angioplasty failures were significantly younger than successes at cardiac catheterization (29±49 versus 54±48 months, P=.04). Patients who developed recoarctation during follow-up were also younger at angioplasty (26±41 versus 62±46 months, P=.02). Length of follow-up was not significantly different between patients in the follow-up groups (Table 1⇓).
Associated Cardiac Defects
Only 8 of 62 patients (13%) with isolated coarctation failed angioplasty early compared with 6 of 10 (60%) with an associated PDA alone (P<.05). Three of 6 patients (50%) with an associated ventricular septal defect and subaortic stenosis also failed angioplasty early (P<.05). At follow-up, however, associated cardiac defects had no apparent effect on subsequent failure rates.
Table 2⇓ summarizes the salient echocardiographic morphometric measurements. Compared with early angioplasty failure patients, early angioplasty success patients had larger indexed aortic annulus (1.48±0.20 versus 1.35±0.23 cm/BSA0.5, P=.03), distal transverse aortic arch (1.15±0.20 versus 0.95±0.17 cm/BSA0.5, P<.001), and aortic isthmus (1.04±0.17 versus 0.79±0.10 cm/BSA0.5, P<.001) diameters. The corresponding z values for these arch measurements reflected similar disparity between groups. Indexed aortic isthmus length was significantly shorter in early angioplasty failure patients (0.90±0.53 versus 1.21±0.41 cm/BSA0.5, P=.02). Indexed coarctation site diameters were not significantly different between early successes and failures, nor were indexed diameters of aortic root, ascending aorta, left subclavian artery, descending aorta, or distal transverse arch length. The preangioplasty LV end-diastolic dimension and estimated Doppler gradients were similar between early outcome groups, but those with angioplasty success had increased LV posterior wall thickness (z values, 1.33±2.12 versus 0.21±1.91, P=.03).
Like early angioplasty success patients, late success patients had larger preangioplasty indexed aortic valve annulus, distal transverse arch, and aortic isthmus diameters and corresponding z values compared with patients who developed recoarctation (Table 2⇑). Late success patients also had larger preangioplasty coarctation site diameters than patients who developed recoarctation during follow-up (0.56±0.13 versus 0.49±0.14 cm/BSA0.5, P<.05), but the groups at recent follow-up were not significantly different in terms of preangioplasty LV end-diastolic dimension, posterior wall thickness, or Doppler gradient.
Table 3⇓ summarizes comparisons of the catheterization data from early and late follow-up groups. The preangioplasty echocardiographically measured aortic isthmus diameter correlated well with the angiographically measured isthmus diameter (r=.96, P<.0001). The echocardiographic measurements differed from the angiographic measurements by 6.2±7.4%, with an average actual difference of 0.49±0.59 mm. Preangioplasty coarctation gradients were similar for all outcome groups. Postangioplasty gradients were significantly higher in early failures compared with early successes (19.3±15.9 versus 6.9±5.3 mm Hg, P<.001). Of the 22 patients who failed early, 8 required intervention for recoarctation within 1 month despite having initial “success'' in the catheterization laboratory. Between groups at follow-up, there was a small, yet statistically significant, difference in postangioplasty residual coarctation gradient (5.6±5.5 versus 9.2±3.8 mm Hg, P=.001). There were no differences in angioplasty technique between outcome groups early or at most recent follow-up for the relationship of balloon size to aortic isthmus diameter, and no differences were observed between groups for operator or maximum inflation pressure. Early failure patients had more inflations per procedure than early success patients (5.5±2.4 versus 4.0±1.8 inflations, P=.003), presumably reflecting less satisfactory results in the catheterization laboratory in these patients. Calendar year of BA had no significant effect on outcome early (P>.25) or during follow-up (P>.5). Small aneurysm formation adjacent to the BA site was noted during follow-up in 8 patients. None of the aneurysms have progressed or required intervention.
Multiple logistic regression analysis was used to test simultaneously the effect on outcome of variables identified by univariate analysis: age, associated cardiac lesions, and echocardiographic measurements. For early and most recent follow-up outcome groups, the preangioplasty aortic isthmus z value was most significantly associated with angioplasty outcome (P<.0001 early; P<.0001 at follow-up) (Fig 3⇓). When corrected for the aortic isthmus z value, age at angioplasty and associated cardiac defects had no significant association with angioplasty outcome. Although a significant linear correlation between age and preangioplasty aortic isthmus z value (r=.43, P<.0001) indicates an association between these variables, patients without significant isthmic hypoplasia did well regardless of age at angioplasty. Nine of 11 patients (82%) who underwent angioplasty at ≤12 months of age and had a preangioplasty aortic isthmus z value >−2.0 had continued success at most recent follow-up. Distal transverse arch z value (P<.001) and indexed aortic isthmus length (P<.01) were also associated independently with early outcome. For follow-up groups, preangioplasty aortic annulus z value (P=.02) and distal transverse arch z value (P<.01) also predicted outcome independently.
Outcome in Neonates
Twenty-one neonates (mean age±SD, 17±10 days) underwent BA for native coarctation of the aorta. PDA was present in 17 of 21 patients (81%), and the average aortic isthmus z value was −2.6±0.6. In 18 of 21 neonates (86%), the isthmus z value was <−2.0; in 2, it was <−1.9; and in 1, it was −1.1. BA failed in 20 of 21 patients (95%).
Prediction of Outcome
By logistic regression analysis (Fig 4⇓), the probability of early angioplasty success [P(Success)] as a function of aortic isthmus z value (z) is described by the equation P(Success)=[1+e(−6.22−2.39z)]−1. An aortic isthmus z value ≤−2.16 [corresponding to a P(Success) ≤0.74] predicted immediate angioplasty failure with 91% sensitivity and 85% specificity (positive predictive value, 61%; negative predictive value, 97%). At most recent follow-up, the relationship of aortic isthmus z value and probability of continued success is expressed by the equation P(Success)=[1+e(−3.21−1.60z)]−1.
A preangioplasty aortic isthmus z value ≤−1.24 [corresponding to a P(Success) ≤0.77] predicted failure at follow-up with 81% sensitivity and 70% specificity (positive predictive value, 52%; negative predictive value, 90%).
Sensitivity, Specificity, and Predictive Values
Several echocardiographic variables and combinations of variables identified patients who failed angioplasty either early or during follow-up (Table 4⇓). The presence of a PDA in association with coarctation predicted early or late angioplasty failure with 93% specificity and 81% positive predictive value, but the sensitivity of this finding was only 40%. The most specific echocardiographic predictor of angioplasty failure was the combination of aortic isthmus z value ≤−2.0 and aortic valve annulus z value ≤−0.6 (specificity, 98%; positive predictive value, 96%). The presence of an aortic isthmus z value ≤−1.8 or an aortic valve annulus z value ≤−0.6 or both yielded the most sensitive predictive combination for angioplasty failure (sensitivity, 88%; negative predictive value, 89%).
The effect of isthmus hypoplasia on outcome over the entire duration of follow-up was examined by survival analysis. The Kaplan-Meier curves in Fig 5⇓ demonstrate the proportion of individuals who remain free of recoarctation over time in three groups stratified by aortic isthmus z value (group 1, aortic isthmus z value >−1.0; group 2, −1.0≥z value>−1.99; group 3, z value ≤−2.0). Throughout the follow-up period, a significantly higher proportion of group 1 patients remained free of recoarctation (group 1 versus group 2, P<.01; group 1 versus group 3, P<.001). Ninety percent of patients in group 1 remained free of recoarctation up to 9.6 years (mean±SD, 3.3±2.9 years) after angioplasty. Conversely, 74% of patients in group 3, all with aortic isthmus hypoplasia, failed angioplasty within 12 months, and only 11% were free of recoarctation at 36 months after the procedure.
The mean interobserver variability for aortic arch measurements was 9±13% (0.9±1.3 mm). The correlation coefficient by simple linear regression analysis for measurements by the two observers was .96 (P<.0001).
This study found that outcome of BA for native coarctation of the aorta can be predicted from preangioplasty morphometric echocardiographic measures of the aortic arch. For both early and follow-up outcome groups, the most discriminating echocardiographic variable was the preangioplasty z value of aortic isthmus diameter. The effect of isthmus diameter z value on BA outcome was independent of age and associated lesions. This is consistent with previous reports that have shown that tubular narrowing of the aortic isthmus is associated with poor outcome early and at long-term follow-up.3 9 10 11 18 19 In previous studies, the diagnosis of aortic isthmic hypoplasia was made by angiographic assessment based on various criteria. Descriptions have ranged from qualitative assessment of aortic arch hypoplasia9 to critical threshold ratios of aortic isthmus diameter to either ascending aorta or descending aorta diameters.3 18 19 In this study, aortic isthmus diameters were indexed to body size and expressed as z values, allowing comparison between patients of different ages and body sizes. Of the entire study group, 37% of patients undergoing the procedure (38/105) had a preangioplasty aortic isthmus z value of ≤−2.0 by echocardiographic assessment. Of these patients, 84% (32/38) failed angioplasty either immediately or at follow-up, composing 74% (32/43) of all study patients who failed.
The independent effect of aortic valve annulus diameter on angioplasty outcome is of interest. Although, in general, aortic annulus diameters were within normal range for age and body size, small but statistically significant differences between groups were observed, with smaller annulus size in those who failed (Table 2⇑). Decreased antegrade aortic flow during fetal life has been proposed to be a contributing mechanism in the pathogenesis of aortic coarctation.27 28 The reason that a relatively minor decrease in aortic valve diameter significantly affects angioplasty outcome is not obvious. Speculatively, a small aortic valve annulus may limit postnatal growth potential of the aortic arch and isthmus, either through limitation of aortic blood flow or through an unrelated mechanism. Interestingly, Rhodes et al29 found that in neonates with critical aortic stenosis, small changes in mitral valve area and aortic root diameter were independent predictors of outcome. Both in the study by Rhodes et al and in the present study, small differences in dimensions of structures upstream or downstream of the obstructive lesion were significantly associated with outcome.
Young age at angioplasty has been proposed as a predictor of angioplasty failure.9 10 12 18 Rao et al18 identified age <12 months as an independent risk factor for angioplasty failure at follow-up in 7 of 20 patients. In the present study, patients who failed angioplasty were younger in both follow-up groups. However, when age was corrected for aortic isthmus z value by multivariate analysis, the effect of age on outcome was eliminated. Nine of 11 patients (82%) who underwent angioplasty at ≤12 months of age and had a preangioplasty aortic isthmus z value >−2.0 had continued success. The correlation between age and aortic isthmus z value found in this study is most likely related to earlier and more symptomatic presentation of patients with severe aortic arch hypoplasia in association with coarctation. Of the 39 patients undergoing angioplasty at age ≤12 months, 37 patients (95%) received medications for symptoms of congestive heart failure at the time of the procedure, whereas only 6% of patients >12 months of age at angioplasty (4/66) required similar treatment. Neonates with severe aortic coarctation necessitating early therapeutic intervention pose a special problem. Severe aortic isthmic hypoplasia and a PDA are particularly common in this group of patients (86% and 81%, respectively). Indeed, 20 of 21 neonates in this series required additional intervention after initial BA for recurrent coarctation.
The data presented here may be helpful for assigning treatment to patients with native coarctation of the aorta. If high-risk patients could be identified by echocardiography before a proposed intervention, catheterization might be avoided in patients unlikely to benefit from BA. Some investigators, citing high rates of restenosis after early surgery, have suggested that BA in infants with severe arch hypoplasia may be warranted as a palliative measure so that definitive surgical treatment can be performed at a later age, when long-term results are more favorable.6 8 Currently, use of this approach varies by institution according to local surgical results and preference. However, recent experience reported by Sandhu et al30 and Conte et al31 with the extended end-to-end anastomosis, designed specifically to deal with arch hypoplasia, has demonstrated that excellent surgical results are achievable even in neonates with severe aortic arch and isthmic hypoplasia.
The retrospective nature of this study imposes inherent limitations. In addition, it might be argued that our institutional technique, which recommends use of an angioplasty balloon diameter only 1 to 2 mm greater than the diameter of the aortic isthmus, limits potential outcome and influences the strong association between isthmus diameter and angioplasty outcome.7 10 Yet, Rao et al18 and Redington et al9 reported poor outcome in patients with isthmus hypoplasia when larger angioplasty balloons were used, and similar results have also been reported with more conservative approaches to balloon size selection.3
We conclude that quantitative echocardiographic measurements of the aorta may be used to predict outcome of BA in patients with native coarctation of the aorta. The combination of aortic isthmus size and valve annulus diameter is a specific and sensitive predictor of angioplasty outcome independent of patient age at angioplasty or associated cardiac defects.
Selected Abbreviations and Acronyms
|BSA||=||body surface area|
|MRI||=||magnetic resonance imaging|
|PDA||=||patent ductus arteriosus|
We thank James E. Lock, MD, for his comments and Emily Flynn-McIntosh for artwork.
Guest editor for this article was Welton M. Gersony, MD, College of Physicians and Surgeons, Columbia University, New York, NY.
- Received September 27, 1995.
- Revision received March 6, 1996.
- Accepted March 13, 1996.
- Copyright © 1996 by American Heart Association
Lock JE, Bass JL, Amplatz K, Fuhrman BP, Castaneda-Zuniga W. Balloon dilation angioplasty of aortic coarctations in infants and children. Circulation. 1983;68:109-116.
Shaddy RE, Boucek MM, Sturtevant JE, Ruttenberg HD, Jaffe RB, Tani LY, Judd VE, Veasy LG, McGough EC, Orsmond GS. Comparison of angioplasty and surgery for unoperated coarctation of the aorta. Circulation. 1993;87:793-799.
Redington AN, Booth P, Shore DF, Rigby ML. Primary balloon dilatation of coarctation of the aorta in neonates. Br Heart J. 1990;64:277-281.
Rao PS, Galal O, Wilson AD. Five-to-nine year follow-up results of balloon angioplasty of native aortic coarctation. J Am Coll Cardiol. 1995;25:124A. Abstract.
Knott-Craig CJ, Elkins RC, Ward KE, Overholt ED, Razook JD, McCue CA, Lane MM. Neonatal coarctation repair: influence of technique on late results. Circulation. 1993;88:198-204.
Ziemer G, Jonas RA, Perry SB, Freed MD, Castaneda AR. Surgery for coarctation of the aorta in the neonate. Circulation. 1986;74(suppl I):I-25-I-31.
Burrows PE, Benson LN, Babyn P, MacDonald C. Magnetic resonance imaging of the iliofemoral arteries after balloon dilation angioplasty of aortic arch obstructions in children. Circulation. 1994;90:915-920.
Lappen RS, Muster AJ, Duffy CE, Berdusis K, Paul MH. Predictability of coarctation of the aorta from the degree of transverse aortic arch hypoplasia: an echocardiographic-angiocardiographic correlation. In: Doyle EF, Engle MA, Gersony WM, Rashkind VJ, Talner NS, eds. Pediatric Cardiology: Proceedings From the Second World Congress of Pediatric Cardiology. New York, NY: Springer-Verlag Inc; 1986:172-175.
Hanse´us K, Bjo¨rkhem G, Lundstro¨m N-R. Dimensions of cardiac chambers and great vessels by cross-sectional echocardiography in infants and children. Pediatr Cardiol. 1988;9:9-17.
Hatle L, Angelsen B. In: Doppler Ultrasound in Cardiology: Physical Principles and Clinical Implications. Philadelphia, Pa: Lea & Febiger; 1985:24.
Labarthe DA, Kosinetz C, Jones TM. Epidemiology. In: Garson A, Bricker JT, McNamara DG, eds. The Science and Practice of Pediatric Cardiology. Philadelphia, Pa: Lea & Febiger; 1990:143.
Ho SY, Anderson RH. Coarctation, tubular hypoplasia, and the ductus arteriosus. Br Heart J. 1979;41:268-278.
Shinebourne EA, Elseed AM. Relation between fetal flow patterns, coarctation of the aorta, and pulmonary blood flow. Br Heart J. 1974;36:492-498.
Rhodes LA, Colan SD, Perry SB, Jonas RA, Sanders SP. Predictors of survival in neonates with critical aortic stenosis. Circulation. 1991;84:2325-2335.