Author + information
- Received December 14, 1995
- Revision received May 9, 1996
- Accepted May 14, 1996
- Published online October 1, 1996.
- MERYL S COHEN,
- MARSHALL L JACOBS,
- PAUL M WEINBERG and
- JACK RYCHIK* ()
- ↵*Address for correspondence: Dr. Jack Rychik, Noninvasive Cardiovascular Laboratory, Division of Cardiology, The Children's Hospital of Philadelphia, Philadelphia, Pennsylvania 19104.
Objectives. This study was designed to define morphometric echocardiographic variables of unbalanced common atrioventricular canal (CAVC) that could aid in appropriate referral for surgical repair.
Background. Unbalanced CAVC has a high surgical mortality rate. This may be secondary to inappropriate referral of some patients for two-ventricle repair (closure of septal defects) instead of single-ventricle repair (Norwood palliation and Fontan operation).
Methods. The echocardiograms of 103 patients with CAVC were retrospectively reviewed. In the subcostal left anterior oblique view, the area of the atrioventricular (AV) valve aportioned over each ventricle was measured, and an AV valve index (AVVI) was calculated as left/right valve area. The ventricular cavity ratio between the two ventricles was estimated as left ventricular length times width divided by right ventricular length times width. These variables were correlated with surgical referral and outcome.
Results. Patients previously categorized as having balanced CAVC all had AVVI >0.67 (n = 77). Of the patients with unbalanced CAVC (n = 26), 11 had ductal-dependent circulation and underwent Norwood palliation (AVVI 0.21 ± 0.13, mean ± SD), and 15 had two-ventricle repair (AVVI 0.51 ± 0.12, p < 0.0001). Of these 15 patients, 9 have survived, with no difference in mean AVVI between survivors and nonsurvivors (0.52 ± 0.11 versus 0.49 ± 0.13, p = 0.72). For all 103 patients, AVVI correlated with ventricular cavity ratio. However, of the unbalanced CAVC group who underwent two-ventricle repair, three nonsurvivors had a discrepancy between AVVI and ventricular cavity ratio (low AVVI but normal ventricular size). A large ventricular septal defect was present in all six nonsurvivors but in only four of nine survivors (p < 0.05).
Conclusions. Echocardiographic morphometry is useful in defining unbalance in CAVC. If AVVI is <0.67 in the presence of a large ventricular septal defect, a single-ventricle approach to repair should be considered.
In common atrioventricular canal (CAVC), the common atrioventricular (AV) valve may be positioned equally over both ventricles (balanced) or unequally over the right or left ventricle (unbalanced), with variable degrees of associated ventricular hypoplasia. Unbalance of the common AV valve occurs in ∼10% of all patients with CAVC . Although successful repair of balanced CAVC through closure of the atrial and ventricular septal defect components (two-ventricle repair) is well established [4–6], surgical outcome for unbalanced CAVC is poor [2, 3]. Reparative strategies for unbalanced CAVC have included closure of the septal defects or, in cases of marked unbalance, palliative procedures, such as pulmonary arterial banding or the Norwood operation [1, 7]. Poor outcome for this lesion may be due in part to the inability to correctly choose the appropriate operative procedure. Difficulty exists in reliably predicting which patients will tolerate two-ventricle repair and which should have single-ventricle palliation with subsequent Fontan operation.
In the past, cardiac catheterization and angiographic estimates of relative ventricular size and volume were used to determine the degree of unbalance in CAVC [8, 9]. These methods, however, are subject to error because severity of unbalance of the common AV valve may not necessarily correlate with the degree of ventricular hypoplasia present. In addition, right ventricular volume has been found to be larger than left ventricular volume in most patients with balanced CAVC [9–11]; hence, it may be difficult to distinguish balance from unbalance solely on the basis of ventricular size. Furthermore, angiographic visualization of the common AV valve, the defining structure of balance or unbalance, can be difficult.
Because ultrasound provides excellent resolution of AV valve tissue as well as adjacent septal structures, it should be superior to angiography in delineating AV valve position. We therefore used echocardiography to develop an index that quantifies the degree of common AV valve unbalance in relation to the ventricular septum. The objective of this study was to morphometrically define by two-dimensional echocardiography the variables of AV valve position that could aid in the appropriate diagnostic classification of unbalanced CAVC, thus resulting in correct recommendations for operation and improved outcome.
Patients. The echocardiograms and medical records of patients who underwent surgical repair of CAVC at the Children's Hospital of Philadelphia from January 1990 to June 1994 were retrospectively reviewed. Patients were excluded if they had heterotaxy syndrome, dextrocardia or conotruncal anomalies (i.e., tetralogy of Fallot), alone or in combination. Of 195 patients identified, 103 met these criteria and had preoperative echocardiograms of good quality for review.
Patients were classified into two groups. Of the 103 patients, 77 (75%) were judged at the time of initial presentation by angiography or operative inspection, or both, concurrent with the impression of the preoperative two-dimensional echocardiogram (conventional means) to have a balanced CAVC (group I). The remaining 26 patients (25%) were judged to have some degree of common AV valve unbalance by angiography, echocardiography or surgical inspection (group II, 23 unbalanced to the right, 3 unbalanced to the left). Echocardiographic morphometric analysis of the AV valve (see later) was not performed a priori and thus was not considered in the diagnostic categorization of these patients before operation.
During review of the medical records, the following was noted for each patient: 1) the presence of Down syndrome; 2) the degree of AV valve insufficiency before operation (patients were categorized as having hemodynamically significant AV valve insufficiency if moderate or severe insufficiency was seen by Doppler color ); and 3) the size of the ventricular septal defect (patients were categorized as having a large unrestrictive defect or a small restrictive or no defect by two-dimensional and Doppler color flow echocardiography).
Operation. All patients in group I (balanced) underwent two-ventricle repair consisting of patch or suture closure of the atrial and ventricular septal defect components (n = 77, age at operation 17 ± 28 months [mean ± SD]). Patients in group II with a diagnosis of unbalanced CAVC were noted to have varying degrees of ventricular hypoplasia and either 1) presented in early infancy with marked left ventricular hypoplasia and a ductal dependent systemic circulation and hence underwent Norwood stage I palliation (group IIA, n = 11, age at operation 9 ± 11 days); or 2) presented in later infancy, without ductal dependent systemic circulation, were perceived to have two adequately sized ventricles and underwent two-ventricle repair (group IIB, n = 15, age at operation 4.8 ± 3.6 months). All patients unbalanced to the left (n = 3) underwent two-ventricle repair.
Echocardiography: morphometric analysis of the common AV valve and estimation of ventricular size. The echocardiograms of all 103 patients, acquired before surgical intervention were analyzed in blinded manner with regard to surgical referral. Patients between 3 weeks and 3 months of age were sedated with 60 to 75 mg/kg body weight of chloral hydrate before the study. Studies were initially recorded on VHS 0.5-in. format videotape using either a Hewlett-Packard Sonos 1000 or Acuson XP128 ultrasound system coupled with either a 5.0- or 3.5-MHz transducer.
The utility of the subcostal left anterior oblique plane in visualizing the ventricular septum on end, as well as the relation of the AV valve to the septum has been well described [13, 14]. The degree of balance or unbalance of the AV valve was therefore established by performing the following, utilizing an off-line analysis system (Digisonics) interfaced with a personal computer: 1) At end-diastole, the common AV valve was visualized in the left anterior oblique plane (Fig. 1). 2) The crest of the muscular septum and the tip of the infundibular septum were identified, and a line was drawn from one to the other, bisecting the common AV valve, thereby designating a portion of the valve to each ventricle. 3) With the common AV valve divided between the two ventricles, the respective designated valve areas over the left and right ventricle were traced, and a ratio of the two areas (left valve area divided by right valve area) was generated (AV valve index [AVVI]). In three cases where the common valve was unbalanced to the left, this ratio was inverted (right valve area/left valve area). An AVVI of 1.0 therefore represents equal balance of the common AV valve over both ventricles.
To estimate the cavity size of the right and left ventricles relative to each other, measurements were made in the apical four-chamber view of ventricular length (from the AV valve annulus to apex) and width (from the crest of the ventricular septum to the free wall) for each ventricle. A ratio of the ventricular cavity dimensions was then calculated as left ventricular length times width divided by right ventricular length times width.
All measurements were made in at least three cardiac cycles at end-diastole by single-frame analysis, and the average was used for calculations. The frame before the onset of closure of the AV valve was designated as end-diastole.
Data analysis and statistics. Results are reported as mean value ± SD. Linear regression analysis was used to test for a correlation between the degree of unbalance of the CAVC (AVVI) and the ventricular cavity ratio. The Mann-Whitney rank sum test was used to assess differences between the groups for the echocardiographic indexes measured. Chi-square analysis (or the Fisher exact test, if n < 10) was used to determine differences in the presence of Down syndrome, degree of preoperative AV valve insufficiency, size of the ventricular septal defect component and survival between the groups. A p value <0.05 was considered significant.
Presence of Down syndrome, unrestrictive ventricular septal defect and preoperative AV valve insufficiency. The prevalence of Down syndrome and an unrestrictive ventricular septal defect was not specifically different between patients with balanced (group I) and unbalanced (group II) CAVC (Table 1). However, a significantly higher percentage of patients with unbalanced CAVC had hemodynamically important AV valve insufficiency than those with balanced CAVC (p = 0.03) (Table 1). Patients with Down syndrome were more likely to have an unrestrictive ventricular septal defect (complete CAVC) than those without Down syndrome (42 [86%] of 49 versus 21 [64%] of 33, p < 0.0001), but the degree of AV valve insufficiency was not significantly different between these two groups (p = 0.344).
Morphometric analysis: balanced (group I) versus unbalanced (group II) CAVC. The AVVI for all 103 patients correlated well with the ventricular cavity ratio (r = 0.75, p < 0.001) (Fig. 2). Mean AVVI and ventricular cavity ratios were not statistically different in relation to the size or presence of a ventricular septal defect (Table 2).
Patients categorized by conventional means as having balanced CAVC had morphometric variables that were significantly higher than patients categorized as having unbalanced CAVC. The AVVI for the balanced group ranged from 0.67 to 1.19, and ventricular cavity ratios ranged from 0.46 to 1.19. Patients determined to have unbalanced defects presented with a wide spectrum ranging from mild common AV valve unbalance to severe unbalance with marked ventricular hypoplasia. The AVVI for the unbalanced group ranged from 0.07 to 0.65; ventricular cavity ratios ranged from 0.01 to 0.87. No overlap was present in AVVI between the balanced and unbalanced groups (no balanced group patient had an AVVI lower, and no unbalanced group patient had an AVVI higher, than 0.66). However, significant overlap was noted for ventricular cavity ratio, with three patients in the unbalanced group having ventricular cavity ratios greater than the mean value for patients in the balanced group (Fig. 2). One patient had marked discrepancy between AVVI and ventricular cavity ratio manifested by significant common AV valve unbalance (AVVI 0.27) but relatively equal right and left ventricular cavity dimensions (ventricular cavity ratio 0.77) (Fig. 2).
Unbalanced CAVC.Group II: single- versus two-ventricle repair. For the 11 patients with unbalanced CAVC who underwent Norwood palliation, both mean AVVI and mean ventricular cavity ratio were significantly smaller than for the 15 who underwent two-ventricle repair (Table 2). However, overlap was present because patients with an AVVI between 0.27 and 0.41 and a ventricular cavity ratio between 0.26 and 0.34 underwent either a Norwood operation or two-ventricle repair (Fig. 2). No patient with an AVVI <0.3 or a ventricular cavity ratio <0.26 had a two-ventricle repair; all presented with ductal-dependent circulation in infancy.
Group IIB: survivors versus nonsurvivors of two-ventricle repair. Nine (60%) of the 15 patients with unbalanced CAVC who underwent two-ventricle repair (group IIB) have survived (Table 3). No significant difference was present in mean AVVI or mean ventricular cavity ratios between survivors and non-survivors (Table 2). Although Down syndrome as well as hemodynamically important (moderate to severe) preoperative AV valve insufficiency was more prevalent in nonsurvivors, these were not statistically significant risk factors. Five (83%) of six nonsurvivors and four (44%) of nine survivors had Down syndrome (p = 0.29); four nonsurvivors (67%) and only one survivor (11%) had moderate AV valve insufficiency (p = 0.08). However, the presence of a large, unrestrictive ventricular septal defect was a significant risk factor for poor surgical outcome in patients with unbalanced CAVC undergoing two-ventricle repair. All six nonsurvivors (100%) and only four (44%) of nine survivors had an unrestrictive ventricular septal defect (p < 0.05).
Of the six nonsurvivors, three had a marked discrepancy between AVVI and ventricular cavity ratio (low AVVI with a ventricular cavity ratio equal to the mean value for the balanced group) (Table 3, Patients 11, 12 and 13).
One survivor of two-ventricle repair has developed coarctation postoperatively requiring reoperation and has acquired left-sided pulmonary veno-occlusive disease (Table 3, Patient 4). Another patient underwent successful valve replacement for severe postoperative AV valve insufficiency (Patient 6). The other seven are doing well at follow-up.
Surgical outcome for patients with unbalanced CAVC has been extremely poor, with mortality rates in some reports as high as 57% to 100% [1–3]. Two-ventricle repair by closure of the atrial and ventricular septal defects may be successfully achieved in some cases; however, in those with marked unbalance of the AV valve, alternate strategies are needed. Improved survival for patients with hypoplastic left heart syndrome and other forms of aortic outflow obstruction by Norwood palliation followed by hemi-Fontan and Fontan operations has allowed for the successful application of this approach in patients with markedly unbalanced CAVC and associated left ventricular hypoplasia. However, in cases of moderate unbalance of the AV valve, deciding which surgical strategy to use may be difficult. Performance of a two-ventricle repair in patients with extensive unbalance of the common AV valve may result in significant mortality. Predictive criteria for determining the viability of function of the left-sided structures after two-ventricle repair of this lesion are lacking.
Echocardiographic morphometry. In the present study, we investigated an echocardiographically defined, quantifiable variable of unbalance of the common AV valve (AVVI) and analyzed the relation of this variable to relative ventricular cavity size, surgical referral and outcome. In all 77 patients with an AVVI >0.67, diagnostic categorization by morphometric analysis concurred with the previous diagnosis of a balanced CAVC by angiography or subjective echocardiographic assessment, whereas all 26 patients with AVVI <0.67 concurred with the previous diagnosis of an unbalanced CAVC. In addition, morphometric measurements were helpful in further classifying patients by surgical referral. There was a lower limit of the AVVI (0.27) at which all patients had ductal-dependent circulation and hence underwent Norwood operation (seven patients). Four patients with an AVVI between 0.27 and 0.40 had ductal-dependent circulation and also underwent the Norwood operation. All remaining patients underwent a two-ventricle repair.
The AVVI predicted relative ventricular cavity size in most but not all patients (Fig. 2). Some patients with a well balanced CAVC (AVVI 0.9 to 1.19) had a large right ventricular cavity relative to the left (ventricular cavity ratio 0.7 to 0.8). This finding has been previously reported and may be due to predominance of atrial septal defect physiology with right ventricular volume overload. Conversely, three patients were noted to have marked unbalance of the AV valve (AVVI 0.3 to 0.5) with a relatively normal ventricular cavity ratio (0.7 to 0.9). Ventricular size assessment alone would have misjudged the degree of common AV valve unbalance present in these patients. This finding suggests that AV valve inflow is not the sole determinant of ventricular cavity size.
Unbalanced CAVC: survival after two-ventricle repair. Fifteen patients with unbalanced CAVC (AVVI <0.67) underwent two-ventricle repair (Table 3). Of the nine survivors, the spectrum of unbalance was relatively wide (AVVI 0.30 to 0.65). Neither degree of unbalance of the common AV valve nor relative size of the ventricle influenced survival within this group. Of note, the patient with the smallest AVVI and ventricular cavity ratio who underwent two-ventricle repair survived the operation and is presently alive. He has subsequently undergone repair of coarctation of the aorta and has developed left pulmonary veno-occlusive disease (Patient 4, Table 3).
Hemodynamically significant preoperative AV valve insufficiency was more prevalent in nonsurvivors; however, statistical significance was not reached (p = 0.08). Of interest, all six nonsurvivors had a large unrestrictive ventricular septal defect (p < 0.05). Both significant AV valve insufficiency and a large ventricular septal defect result in a volume load on the left ventricle. It is plausible that this volume load may alter the dimensions of the left ventricle before repair and result in spurriously greater morphometric values than would exist were these factors not present. This phenomenon was evident in three patients (Patients 11, 12 and 13, Table 3) who had marked unbalance of the common AV valve and normal ventricular cavity size. All three had unrestrictive ventricular septal defects, and two of the three had significant AV valve insufficiency before operation. All three expired soon after two-ventricle repair.
Although dimensional analysis was the focus of our investigation, functional abnormalities influencing outcomes should be considered as well in patients with significant AV valve insufficiency or an unrestrictive ventricular septal defect, or both. Reduced afterload, such as exists with significant AV valve insufficiency, may make the left ventricle of an unbalanced CAVC with poor ventricular function appear to contract well. In addition, a large ventricular septal defect may act as a conduit for systemic circulatory support by allowing right to left shunting from the right ventricle directly into the aorta. One would expect these patients to have mild aortic desaturation. We reviewed the cardiac catheterization data for all patients who underwent two-ventricle repair (12 of 15 had preoperative cardiac catheterization data available for review) and found that 5 of 12 (2 survivors, 3 nonsurvivors) had aortic desaturation. It is difficult to determine the site of the right to left shunt in patients with a large atrial septal defect. With the potential for either pulmonary venous desaturation, atrial level shunting or ventricular level shunting present, in addition to streaming effects, interpretation of aortic desaturation data is difficult and was not helpful in distinguishing between patients in our study. Nevertheless, it appears quite likely that given a small but adequate atrioventricular inflow, a poorly functioning ventricle may supply good systemic perfusion with the aid of right ventricular support across a ventricular septal defect. Only after closure of the defect would left ventricular dysfunction and the potential inability to support systemic output become unmasked. Morphometric analysis must therefore be interpreted in conjunction with an assessment of valve insufficiency and the presence of a ventricular septal defect to appropriately refer patients for operation.
Limitations of the study. Volumetric estimates of ventricular cavity size were not performed. Volume measurements such as the Simpson's method have not been proved to be accurate for the right ventricle and are probably not accurate for the left ventricle in patients with CAVC because the shape of the left ventricle in this disease is different from that of the normal ventricle. Linear dimensions in the single-plane view were readily available in all cases and were used to estimate relative cavity dimension rather than volume between the right and left ventricle in each patient. These measurements are also readily reproducible.
The number of patients with unbalanced CAVC was relatively small but was proportionately greater within our cohort (25%) than previously reported (10%) . This finding may be related to selection bias because our center is a referral site for single-ventricle palliation by Norwood operation. Of note, no patient without a patent ductus arteriosus and right to left flow supplying systemic circulation underwent Norwood operation. There were only three patients in the study with unbalance to the left; thus not many conclusions can be drawn about this entity. The physiology may very well be different in these patients because the small ventricular chamber is nonsystemic; however, the problem of relative AV valve stenosis after repair as a cause of morbidity or mortality, or both, would exist in right and left dominant ventricles.
Summary. We suggest that in patients with suspected unbalanced CAVC, the AVVI should be calculated using two-dimensional echocardiography in the subcostal left anterior oblique view. As a supplement to other means of evaluation, we believe that the algorithm shown in Fig. 3 should form the basis for appropriate stratification of patients for operation. Because this disease has a high mortality rate, the Fontan operation, which has a low mortality rate in most experienced institutions, may be a better option for some patients with unbalanced CAVC. A prospective application of morphometric analysis in unbalanced CAVC is warranted and should aid in improving surgical outcome for this complex lesion.
A.1 Abbreviations and Acronyms
AV = atrioventricular
AVVI = atrioventricular valve index
CAVC = common atrioventricular canal
- Received December 14, 1995.
- Revision received May 9, 1996.
- Accepted May 14, 1996.
- THE AMERICAN COLLEGE OF CARDIOLOGY
- ↵Corno A, Marino B, Catena G, Marcelletti C. Atrioventricular septal defects with severe left ventricular hypoplasia. J Thorac Cardiovasc Surg 1988;96:249–52.
- ↵Mehta S, Hirschfeld S, Riggs T, Liebman J. Echocardiographic estimation of ventricular hypoplasia in complete atrioventricular canal. Circulation 1979;59:888–93.
- Bharati S, Lev M. The spectrum of common atrioventricular orifice (canal). Am Heart J 1973;86:553–61.
- ↵Mair DD, McGoon DC. Surgical correction of atrioventricular canal during the first year of life. Am J Cardiol 1977;40:66–9.
- Chin AJ, Keane JF, Norwood WI, Castenada AR. Repair of complete common atrioventricular canal in infancy. J Thorac Cardiovasc Surg 1982;84:437–45.
- Studer M, Blackstone EH, Kirklin JW, et al. Determinants of early and late results of repair of atrioventricular septal (canal) defects. J Thorac Cardiovasc Surg 1982;84:523–42.
- ↵Norwood WI. Hypoplastic left heart syndrome. Ann Thorac Surg 1991;52:688–95.
- ↵Thies WR, Bargeron LM, Bini RM, Colvin EV, Soto B. Spectrum of hearts with one underdeveloped and one dominant ventricle. Pediatr Cardiol 1986;7:129–39.
- ↵Thanopaulos BD, Fisher EA, DuBrow IW, Hastreiter AR. Ventricular volumes in common atrioventricular canal. Circulation 1978;57:991–5.
- Espinosa-Caliani JS, Alvarez-Guisado L, Munoz-Castellanos L, et al. Atrioventricular septal defect: quantitative anatomy of the right ventricle. Pediatr Cardiol 1991;12:206–13.
- Jarmakani JM, George B, Wheller J. Ventricular volume characteristics in infants and children with endocardial cushion defects. Circulation 1978;58:153–7.
- ↵Wu YT, Chang AC, Chin AJ. Semiquantitative assessment of mitral regurgitation by Doppler color flow imaging in patients aged <20 years. Am J Cardiol 1993;71:727–32.
- ↵Chin AJ, Yeager SB, Sanders SP, et al. Accuracy of prospective two-dimensional echocardiographic evaluation of left ventricular outflow tract in complete transposition of the great arteries. Am J Cardiol 1985;55:759–64.
- Silverman N. Atrioventricular septal defects (atrioventricular canal defects). In: Pediatric Echocardiography. Baltimore, MD: Williams & Wilkins, 1993:143–66.