Echocardiographic spectrum of the hypoplastic left heart syndrome: A clinicopathologic correlation in 19 newborns

Ultrasonic studies were performed in 19 neonates with the hypoplastic left heart syndrome whose diagnosis was confirmed at angiography or autopsy, or both. The patients were classified in two echocardiographic groups: Group I, 10 infants whose ventricular septum could be recorded, and Group II, 9 infants whose septum could not be recorded. The findings in these groups were compared with those in 60 neonates without congenital heart disease also studied with ultrasound. Two additional neonates who presented with signs of shock were also studied.The diagnostic echocardiographic features of hypoplastic left heart syndrome were: (1) a left ventricular end-diastolic dimension of less than 9 mm; (2) an aortic root diameter of less than 6 mm; (3) a ratio of left ventricular end-diastolic to right ventricular end-diastolic dimension of less than 0.6; and (4) a mitral valve echo that is absent or greatly distorted and of small amplitude. These echocardiographic criteria differed significantly from findings in the normal group (P < 0.01). Echocardiography proved valuable in neonates with shock. It is a safe, reliable technique that can be used to delineate the intracardiac anatomy in sick neonates with the hypoplastic left heart syndrome.     

Cross sectional echocardiographic identification of hypoplastic left heart syndrome and differentiation from other causes of right ventricular overload

To identify the echocardiographic features that can be used to distinguish between hypoplastic left heart syndrome and other causes of right ventricular overload in the sick neonate cross sectional echocardiographic studies of 10 neonates with hypoplastic left heart syndrome were analysed and compared with those in 15 neonates with other causes of right ventricular overload and 15 normal controls. Left ventricular and right ventricular cavity dimensions and the shape and size of the mitral valve annulus and aortic root were recorded and presented both as absolute values (mm) and corrected for body surface area (mm/m2). Logistic regression was used to produce a classification rule to estimate the probability of a neonate having hypoplastic left heart syndrome. The diameter of the mitral valve annulus was the single most discriminative variable. There was, however, considerable overlap of all the calculated features between neonates with hypoplastic left heart syndrome and those with other causes of right ventricular overload. The diagnosis of hypoplastic left heart syndrome should not be based on any one single echocardiographic feature but the general appearance of abnormal left heart valves, small cavity dimensions, and the size of the mitral valve annulus.     

Cross sectional echocardiographic identification of hypoplastic left heart syndrome and differentiation from other causes of right ventricular overload.

To identify the echocardiographic features that can be used to distinguish between hypoplastic left heart syndrome and other causes of right ventricular overload in the sick neonate cross sectional echocardiographic studies of 10 neonates with hypoplastic left heart syndrome were analysed and compared with those in 15 neonates with other causes of right ventricular overload and 15 normal controls. Left ventricular and right ventricular cavity dimensions and the shape and size of the mitral valve annulus and aortic root were recorded and presented both as absolute values (mm) and corrected for body surface area (mm/m2). Logistic regression was used to produce a classification rule to estimate the probability of a neonate having hypoplastic left heart syndrome. The diameter of the mitral valve annulus was the single most discriminative variable. There was, however, considerable overlap of all the calculated features between neonates with hypoplastic left heart syndrome and those with other causes of right ventricular overload. The diagnosis of hypoplastic left heart syndrome should not be based on any one single echocardiographic feature but the general appearance of abnormal left heart valves, small cavity dimensions, and the size of the mitral valve annulus.     

Left ventricular dysfunction in the fetus: relation to aortic valve anomalies and endocardial fibroelastosis.

OBJECTIVE--To examine the relation between a characteristic form of left ventricular dysfunction in the fetus and abnormalities of the aortic valve and endocardial fibroelastosis of the left ventricle. DESIGN--A retrospective study to examine the correlation between echocardiographic findings in the fetus and postnatal or necropsy findings. SETTING--Tertiary referral centre for fetal echocardiography. PATIENTS--Thirty fetuses showing a characteristic echocardiographic picture of left ventricular dysfunction. MAIN OUTCOME MEASURES--The relation between the prenatal echocardiographic features and the postnatal and necropsy findings. RESULTS--At presentation the size of the left ventricular cavity was normal or enlarged in all cases. The measurements of the orifice of the aortic root and mitral valve were either normal or small for the gestational age. The echocardiographic diagnosis made at presentation was critical aortic stenosis in all cases. At necropsy or postnatal examination the aortic valve was dysplastic and stenotic in 15 cases and the left ventricle had become hypoplastic in one of these. Aortic atresia was present in seven patients, three of whom had a hypoplastic left ventricle. In six patients the aortic valve was bicuspid although not obstructive. One of these patients had hypoplasia of the aortic arch and one had a hypoplastic left ventricle but in the remaining four patients endocardial fibroelastosis of the left ventricle was the only abnormality found. No follow up information was available in two. Of 26 patients for whom there was postmortem information, 24 had evidence of some degree of endocardial fibroelastosis of the left ventricle. Sequential observations showed that five cases developed into the hypoplastic left heart syndrome. CONCLUSIONS--This type of left ventricular dysfunction in the fetus is the result of an overlap of diseases, including primary left ventricular endocardial fibroelastosis, critical aortic stenosis, and the hypoplastic left heart syndrome.     

Left ventricular dysfunction in the fetus: relation to aortic valve anomalies and endocardial fibroelastosis.

OBJECTIVE--To examine the relation between a characteristic form of left ventricular dysfunction in the fetus and abnormalities of the aortic valve and endocardial fibroelastosis of the left ventricle. DESIGN--A retrospective study to examine the correlation between echocardiographic findings in the fetus and postnatal or necropsy findings. SETTING--Tertiary referral centre for fetal echocardiography. PATIENTS--Thirty fetuses showing a characteristic echocardiographic picture of left ventricular dysfunction. MAIN OUTCOME MEASURES--The relation between the prenatal echocardiographic features and the postnatal and necropsy findings. RESULTS--At presentation the size of the left ventricular cavity was normal or enlarged in all cases. The measurements of the orifice of the aortic root and mitral valve were either normal or small for the gestational age. The echocardiographic diagnosis made at presentation was critical aortic stenosis in all cases. At necropsy or postnatal examination the aortic valve was dysplastic and stenotic in 15 cases and the left ventricle had become hypoplastic in one of these. Aortic atresia was present in seven patients, three of whom had a hypoplastic left ventricle. In six patients the aortic valve was bicuspid although not obstructive. One of these patients had hypoplasia of the aortic arch and one had a hypoplastic left ventricle but in the remaining four patients endocardial fibroelastosis of the left ventricle was the only abnormality found. No follow up information was available in two. Of 26 patients for whom there was postmortem information, 24 had evidence of some degree of endocardial fibroelastosis of the left ventricle. Sequential observations showed that five cases developed into the hypoplastic left heart syndrome. CONCLUSIONS--This type of left ventricular dysfunction in the fetus is the result of an overlap of diseases, including primary left ventricular endocardial fibroelastosis, critical aortic stenosis, and the hypoplastic left heart syndrome.     

Hypoplastic left heart syndrome: is echocardiography accurate enough to guide surgical palliation?

Two-dimensional echocardiography can diagnose hypoplastic left heart syndrome. However, with the advent of the possibility of palliative open heart surgery, complete anatomic diagnosis is necessary. The anatomic findings of 15 neonates with hypoplastic left heart syndrome (age 1 to 10 days, mean 4.1) who had two-dimensional Doppler echocardiographic studies were compared with the results obtained by angiography (6 cases), surgery (11 cases) and autopsy (8 cases). Complete two-dimensional echocardiographic examination of the aortic arch, pulmonary and systemic venous return, atrial septum, ductus arteriosus and proximal coronary arteries was possible in all 15 neonates and correctly diagnosed hypoplastic left heart syndrome in each. Anatomic two-dimensional echocardiographic assessment was accurate in 13 (86%) of the 15 neonates and there were no false positive results. Undiagnosed associated abnormalities were hypoplasia of a left pulmonary artery in one patient and left superior vena cava in another. Accurate quantitation of the size of the tricuspid valve anulus, ascending aorta, pulmonary anulus and right and left pulmonary arteries was possible. Doppler examination was performed in seven patients and confirmed retrograde aortic arch flow and right to left systolic shunting in the patent ductus arteriosus. In selected neonates, surgical palliation can be attempted without angiography.     

Hypoplastic left heart syndrome: experience with palliative surgery

Aortic atresia is a form of hypoplastic left heart syndrome always complicated by associated severe hypoplasia of the ascending aorta and various degrees of mitral valve and left ventricular hypoplasia. At present it is a universally fatal lesion in early infancy. This is a report of a new palliative procedure for hypoplastic left heart syndrome that has resul;ed in early ongoing survival of two infants with aortic atresia. On the basis of experience with a third patient, an operation for future physiologic correction is proposed.     

Coronary artery abnormalities and right ventricular histology in hypoplastic left heart syndrome.

OBJECTIVES. To determine whether right ventricular preservation is compromised in any anatomic subgroup of hypoplastic left heart syndrome, we assessed the coronary artery anatomy and myocardial histologic features of 151 postmortem specimens with hypoplastic left heart syndrome. BACKGROUND. Although previous studies have shown that coronary artery abnormalities are more prominent in the subgroup of patients who have a patent mitral valve and obstructed aortic valve, it has not been established that these abnormalities compromise right ventricular perfusion. METHODS. Eighty-nine specimens with a patent mitral valve and aortic atresia, 52 with mitral and aortic atresia and 10 with patent mitral and aortic valves were examined. Histologic sections of the right ventricle, left ventricle and coronary arteries were obtained from 64 study hearts and 5 control hearts. RESULTS. Gross coronary anomalies included coronary-cameral communications (n = 29), single left coronary artery (n = 2), single right coronary artery (n = 1) and tortuosity (n = 19). Coronary-cameral communications and tortuosity were significantly associated with the subgroup that had mitral hypoplasia and aortic atresia. Coronary artery wall thickness relative to lumen diameter was not different among the subgroups. No areas of coronary artery stenosis or interruption were discovered. Although endocardial fibroelastosis of the left ventricle was strongly associated with the mitral hypoplasia and aortic atresia group, the right ventricular histologic findings were similar in all subgroups and were not differentially affected by older age. CONCLUSIONS. Although the incidence of coronary abnormalities is greater in patients with mitral hypoplasia and aortic atresia, in this study there was no apparent difference in perfusion of the right ventricle among the anatomic subgroups of hearts with hypoplastic left heart syndrome.     

M mode echocardiography in hypertrophic cardiomyopathy: Diagnostic criteria and prediction of obstruction

To assess the reliability of the classic echocardiographic features in the heart with hypertrophic cardiomyopathy as criteria that differentiate it from normal heart and as predictors of outflow tract obstruction versus nonobstruction, 70 patients with clinical and angiographic evidence of hypertrophic cardiomyopathy were studied with M mode echocardiography. The diagnostic sensitivity and specificity of the classic features were assessed: ventricular septal thickness, 83 and 94 percent (sensitivity and specificity, respectively); ventricular septal amplitude of movement, 71 and 89 percent; ventricular septal thickness to left ventricular posterior wall ratio, 79 and 94 percent; left ventricular end-systolic dimension, 54 and 86 percent; septal-mitral valve distance at the onset of systole, 29 and 100 percent; systolic anterior motion of the mitral valve, 61 and 100 percent; and mid systolic closure of the aortic valve, 61 and 100 percent.No single M mode echocardiographic feature was consistently abnormal in hypertrophic cardiomyopathy. In nonobstructive hypertrophic cardiomyopathy, ventricular septal thickness greater than or equal to 13 mm (sensitivity 68 percent and specificity 94 percent) and ventricular septal thickness to posterior wall ratio greater than or equal to 1.5 (sensitivity 82 percent and specificity 94 percent) were the individual features with the greatest diagnostic value from the norm. Patients with obstruction at rest and labile obstruction (gradient only on provocation) had echocardiographically identical features. Ventricular septal thickness greater than or equal to 13 mm plus systolic anterior motion of the mitral valve or mid systolic closure of the aortic valve were the features that in combination best differentiated obstructive (resting and labile) from nonobstructive hypertrophic cardiomyopathy (sensitivity 82 percent and specificity 68 percent) and the heart with obstructive hypertrophic cardiomyopathy from the normal heart (sensitivity 82 percent and specificity 100 percent).     

Relation of the echocardiographic estimate of left ventricular size to mortality in infants with severe left ventricular outflow obstruction

The relation of left ventricular size, as estimated with echocardiography, to mortality was evaluated in three groups of infants with severe left ventricular outflow obstruction. Group I consisted of 17 patients with combined aortic and mitral stenosis or atresia associated with definite hypoplasia of the left ventricle. Group II consisted of eight patients with the primary diagnosis of severe aortic stenosis. Group III consisted of 12 patients with severe coarctation of the aorta. The left ventricular enddiastolic dimension measured with M mode echocardiography and the cross-sectional area of the left ventricular cavity as seen in the parasternal long axis view of the two dimensional echocardiogram were used as indexes of left ventricular volume.All patients with symptomatic outflow obstruction and a left ventricular end-diastolic dimension of less than 13 mm died in infancy. However, five patients with a hypoplastic left ventricle proved at angiography or at autopsy, or both, were found to have a ventricular end-diastolic dimension of 13 mm or greater. Two dimensional echocardiography showed that the left ventricle in these patients was foreshortened and spherical in shape. The cross-sectional area of the left ventricle of each patient in group I was less than 1.6 cm². This was below the range of cross-sectional areas found in a group of normal infants (1.8 to 3.5 cm² ± 2 standard deviations about the mean). Three patients in groups II and III had a slightly reduced left ventricular area (1.7 cm²) and none of these patients survived infancy.Measurement of the cross-sectional area of the left ventricle is a useful method of determining left ventricular size in infants suspected of having the hypoplastic left ventricle syndrome. Patients who have reduced left ventricular volume as assessed by this technique are at very great risk even if surgical relief of the outflow obstruction is attempted.     

Echocardiographic and clinical spectrum of mitral anular calcification

To establish the reliability of echocardiography in making the diagnosis of mitral anular calcification 10 consecutive patients with dense echoes at the mitral anular area were examined fluoroscopically, using an image intensifier. Nine of these had distinct mitral anular calcification. The echocardiographic reliability thus confirmed, 30 cases with similar findings were reviewed and the total group of 40 cases were used to define the echocardiographic spectrum of mitral anular calcification. This lesion extended variably from the anulus to involve the posterior left ventricular wall, mitral valve leaflets, aortic root, aortic valve and interventricular septum. The pleomorphic echocardiographic findings introduce difficulties in the diagnosis of pericardial effusion and mitral stenosis. Echocardiography is a reliable method of diagnosing mitral anular calcification, but care must be taken to avoid confusing this condition with others it resembles.     

Anatomic types of congenital mitral stenosis: Report of 49 autopsy cases with consideration of diagnosis and surgical implications

In a series of 49 autopsy cases of congenital mitral stenosis, four anatomic types were found: (1) typical congenital mitral stenosis with short chordae tendineae, obliteration of interchordal spaces and reduction of interpapillary distance (24 cases, 49 percent); (2) hypoplastic congenital mitral stenosis almost always associated with a hypoplastic left heart syndrome (20 cases, 41 percent); (3) supramitral ring (6 cases, 12 percent); and (4) parachute mitral valve (4 cases, 8 percent). The median ages at death were as follows: parachute mitral valve, 9 $\text{11}\text{12}\text{years}$; supramitral ring, 5 $\text{6}\text{12}\text{years}$; typical congenital mitral stenosis, 6 months, and hypoplastic congenital mitral stenosis, 5 days. Thus, parachute mitral valve had the best natural history and the hypoplastic type of mitral stenosis the worst; that of the other two types was intermediate. Associated malformations were present in 96 percent of cases, endocardial sclerosis or florid endocardial fibroelastosis occurring in 47 percent, hypoplasia of the aortic isthmus in 37 percent, aortic valve stenosis in 29 percent, aortic atresia in 29 percent and coarctation of the aorta in 27 percent of the cases. On the basis of an understanding of the pathologic anatomy, real time twodimensional echocardiography probably will make it possible to make the differential diagnosis among the four major anatomic types of congenital mitral stenosis. Mitral valvuloplasty of several new types appears to merit trial as the initial surgical procedure. At present, mitral valve replacement should be deferred as long as possible.     

Echocardiographic manifestations of valvular vegetations

Eight patients with autopsy or surgically proved valvular vegetations were examined using echocardiography. Five of these patients had lesions on the aortic valve and three had lesions on the mitral valve. The echocardiographic finding in these patients was a non-uniform thickening of valve leaflets which exhibited unrestricted motion. Often the abnormal echoes which produced the thickened valve had a shaggy appearance. In all eight patients the location of the echocardiographic abnormality correlated with the anatomic findings at surgery or autopsy. In one patient the diagnosis of bacterial endocarditis was first suspected following the echocardiographic examination and only subsequently was a heart murmur heard. These findings indicate that echocardiography may play a useful role in elucidating the pathological anatomy of the bacterial endocarditis with vegetation; however, the length of time from the onset of clinical illness to echocardiographic diagnosis remains unknown.     

Echographic-quantified severity of mitral anulus calcification: Prognostic correlation to related hemodynamic,valvular, rhythm,and conduction abnormalities

Clinical and echocardiographic findings in 123 patients with mitral anulus calcification (MAC) were analyzed. In all patients M-mode echocardiography demonstrated a dense band of echoes posterior to the mitral valve, moving parallel and anterior to the left ventricular endocardium. Thirty-three per cent of patients were classified as having minimal to mild MAC (< 5 mm) and 67% had moderate to severe MAC (≥ 5 mm). There was a significant correlation between the degree of MAC to left atrial enlargement, congestive heart failure, aortic valve sclerosis, mitral regurgitation, atrial fibrillation, and AV-fascicular conduction defects. ECG evidence of conduction disturbances was significantly associated with MAC ≥ 5 mm in width. The echocardiographic demonstration of MAC ≥ 5 mm was significantly associated with the clinical implications known to occur with MAC; this echographic finding has important prognostic value in the evaluation of patients with mitral anulus calcification.     

Hypoplastic left heart syndrome and 45X karyotype

A review of 63 patients with 45X karyotype (Turner's syndrome) admitted to a hospital from 1972 to 1985 showed that 20 (32%) had one or more major cardiac malformations (mostly coarctation and aortic stenosis). Four (20%) died in the neonatal. One infant had mitral stenosis and severe aortic stenosis and died at the age of 35 days. The three (15%) other patients who died had a typical hypoplastic left heart syndrome, with an atretic aortic valve in two and pinpoint aortic valve in one. Turner's syndrome was clinically suspected in three of the cases. One of these had mosaicism (46XX,45X) the others had a 45X pattern. During the same period (1972-85) 39 patients (14 girls and 25 boys) were admitted with diagnosis of hypoplastic left heart syndrome. Three (21.4%) girls had a 45X karyotype. The high incidence of hypoplastic left heart syndrome in Turner's syndrome and of Turner's syndrome in girls with hypoplastic left heart syndrome suggests that hypoplastic left heart syndrome can be another expression of the 45X karyotype.     

Hypoplastic left heart syndrome and 45X karyotype.

A review of 63 patients with 45X karyotype (Turner's syndrome) admitted to a hospital from 1972 to 1985 showed that 20 (32%) had one or more major cardiac malformations (mostly coarctation and aortic stenosis). Four (20%) died in the neonatal. One infant had mitral stenosis and severe aortic stenosis and died at the age of 35 days. The three (15%) other patients who died had a typical hypoplastic left heart syndrome, with an atretic aortic valve in two and pinpoint aortic valve in one. Turner's syndrome was clinically suspected in three of the cases. One of these had mosaicism (46XX,45X) the others had a 45X pattern. During the same period (1972-85) 39 patients (14 girls and 25 boys) were admitted with diagnosis of hypoplastic left heart syndrome. Three (21.4%) girls had a 45X karyotype. The high incidence of hypoplastic left heart syndrome in Turner's syndrome and of Turner's syndrome in girls with hypoplastic left heart syndrome suggests that hypoplastic left heart syndrome can be another expression of the 45X karyotype.     

Flail aortic valve leaflets: M-mode and two-dimensional echocardiographic manifestations

Four patients with documented flail aortic valve leaflets were studied using M-mode and two-dimensional echocardiography (2-D echo). Two had aortic valve endocarditis, one had endocarditis involving a congenital heart defect, and one had a myxomatous aortic valve. Mitral valve flutter and early mitral valve closure led to the diagnosis of severe aortic insufficiency in three patients. Diastolic aortic valve flutter, considered to be specific for a flail aortic leaflet, was present in three patients. In the fourth patient left ventricular outflow tract (LVOT) echoes were present, but did not distinguish between a flail aortic leaflet and an aortic vegetation. Two-D echo confirmed LVOT echoes in all patients. Discrimination between a flail leaflet and a vegetation(s) without leaflet disruption was accomplished by noting the hinge point of the LVOT diastolic echoes, which was the aortic wall in patients with a flail leaflet.The combination of these M-mode and 2-D echocardiographic findings permitted the diagnosis of a flail aortic leaflet to be made accurately and noninvasively. In two patients surgery was performed without prior cardiac catheterization.     

Safe transthoracic left ventricular puncture performed with echocardiographic guidance

Retrograde aortic left ventricular pressure recording is difficult to obtain and often unreliable in the presence of a mechanical aortic valve prosthesis. Under echocardiographic guidance, we have safely performed transthoracic left ventricular puncture for pressure recording in seven patients with an aortic valve prosthesis. The transducer is placed at the left sternal border and then slid laterally until the right ventricular cavity and interventricular septal echoes disappear, but before the echoes from the left lung are recorded. The chest wall is marked and the transducer is angled until the mitral valve echoes are identified. The transducer angle and direction are recorded with external markers, and the left ventricular cavity depth is measured. An 18-gauge needle is then inserted the determined depth into the left ventricle, guided exactly by the echocardiographic external markers. The procedure was successful in each patient on the first attempt without a single complication. We conclude that this technique affords the operator greater control when left ventricular pressure must be recorded by transthoracic puncture and that it may reduce the risk of potentially serious complications.     

Assessment of left ventricular endocardial fibroelastosis in fetuses with aortic stenosis and evolving hypoplastic left heart syndrome

Systematic evaluation of left ventricular (LV) endocardial fibroelastosis (EFE) in the fetus has not been reported. The role of EFE in the pre- and postnatal evolution of hypoplastic left heart disease, and the implications of EFE for outcomes after prenatal intervention for fetal aortic stenosis with evolving hypoplastic left heart syndrome have also not been determined. A 4-point grading system (0-3) was devised for the assessment of fetal LV echogenicity, which was presumed to be due to EFE. Two reviewers independently graded EFE on the preintervention echocardiograms of fetuses treated with in utero aortic valvuloplasty for evolving hypoplastic left heart syndrome from 2000 to 2008. Intra- and interobserver reproducibility was determined for the EFE grade and characterization of related echocardiographic features. The relations among EFE severity, other left heart anatomic and physiologic variables, and postintervention outcomes were analyzed. The assessment and grading of EFE was possible for both observers in all 74 fetuses studied. By consensus, the EFE severity was grade 1 in 31 patients, grade 2 in 32, and grade 3 in 11. Fetuses with mild (grade 1) EFE had significantly greater maximum instantaneous aortic stenosis gradients (e.g., higher LV pressures) and less globular LV geometry than patients with grade 2 or 3 EFE on preintervention echocardiogram. The severity of EFE was not associated with the size of the aortic valve or LV. From preintervention to late gestation, the time-indexed change in LV end-diastolic volume was significantly greater in fetuses with grade 1 EFE than those with more severe EFE. Incorporation of EFE severity into our previously published threshold score improved the sensitivity and positive predictive value for the postnatal biventricular outcomes. In conclusion, echocardiographic grading of EFE is possible, with reasonable intra- and interobserver reliability in midgestation fetuses with evolving hypoplastic left heart syndrome. EFE severity corresponded to some indexes of left heart size, geometry, and function and with the probability of a biventricular outcome postnatally. Additional experience and external validation of the EFE grading scoring system are necessary.     

Selective aortic root angiography in the hypoplastic left heart syndrome.

Selective aortic root angiography was performed in 18 patients with the clinical diagnosis of hypoplastic left heart syndrome in order to allow more precise correlation with echocardiography. Aortic valve atresia and severe hypoplasia of the left heart chambers subsequently were substaniated at necropsy in each instance. Wide variation in the size and shape of the aortic root and ascending aorta was demonstrated angiographically in these patients. In some patients, despite aortic valve atresia, the dimensions of the aortic root and sinuses of Valsalva at angiography were almost normal. In addition, this technique allowed the demonstration of primitive vascular communications between the coronary artery and left ventricular cavity in 1 patient (coronary-cameral fistula). The realization that the aortic root may approach nearly normal size in some patients with the hypoplastic left heart syndrome is essential to the echocardiographic diagnosis of this disease and to the differential diagnosis from other cardiac and noncardiac causes of severe neonatal congestive heart failure.