Feasibility of Biventricular Repair in Right Dominant Unbalanced Atrioventricular Septal Defect: A New Echocardiographic Metric to Refine Surgical Decision-Making

Use of the mAVVI as a Reliable Relation Between Left and Total Diameters of the AVV

The AVV has an oval, rectangular-like, shape, with a lateral left to right diameter larger than the anteroposterior diameter. In the SLAO view and in direct surgical view, its anteroposterior diameter is practically preserved all along the valve, while the lateral diameters of the left and right components show relative variations (Figure 2).

OPEN IN VIEWER

Figure 2.

The anteroposterior diameter of the atrioventricular valve (AVV) is almost preserved all along the valve and therefore h1, h2, h3, and h4 have nearly the same length. For both subcostal left anterior oblique (SLAO) and apical four-chamber (4C) view, the reference points are the same as the ones depicted in Figure 4. The point L is the left vertex, R is the right vertex, and M is the point in which h3, the projection of the interventricular septum, perpendicularly crosses the lateral left to right diameter of the AVV.

These variations are related to the ventricular sizes and the position of the interventricular septum, which in turn determine the degree of unbalance. Therefore, anteroposterior diameters can be canceled out from the equations used to calculate the left and total AVV areas. By doing this, the left component area/total AVV area ratio depend almost only on the respective lateral left to right diameters. This means that the mAVVI is nearly equal to the left component diameter/total AVV diameter ratio measured in the same SLAO view for a given patient.

Besides, the ratio of left component diameter/total AVV diameter measured in the SLAO view is equal to the same ratio obtained from the apical 4-chamber (4C) view. The reason for that is that both views are almost perpendicular to each other, and the reference points are the same. These points are the two end points of the lateral left to right maximum diameter line, and the point in this line that divides the left from the right components of the AVV. The exact location of this division point is obtained in both echocardiographic views. In the SLAO view, it is determined by the line extended from both remnants of the interventricular septum crossing the line of the maximum lateral left to right diameter of the AVV, and in the apical 4C view, by the line extended from the crest of the interventricular septum crossing the AVV plane perpendicularly and dividing the valve in its left and right components (Figure 2).

As shown, the ratio of left and total areas, that is, the mAVVI, is nearly equal to the left component diameter/total AVV diameter ratio in the SLAO view. This ratio is also the same for the diameters obtained in the apical 4C view. We have tested this logical correspondence in 13 consecutive patients with AVSD. Areas and diameters in the SLAO view were obtained. Average of three measures was calculated. Diameters in the apical 4C view were also acquired. Ratios for each index were determined: left AVV area/total AVV area, left component diameter/total AVV diameter in the SLAO view, and left component diameter/total AVV diameter in the apical 4C view. As our interest was focused on the apical 4C view, we compared each value of mAVVI with its corresponding value of left component diameter/total AVV diameter ratio measured in the apical 4C view. We observed less than 3% differences, reflecting a tight correlation between both indexes (Figure 3).

OPEN IN VIEWER

Figure 3.

Each patient was evaluated for differences in the modified atrioventricular valve index (mAVVI) compared to left atrioventricular valve (AVV) diameter/total AVV diameter in the apical four-chamber (4C) view. The ratio of these measures for each patient is identified with an asterisk. Note that there is scarce dispersion from the central value of 1, meaning that both measures are almost equal for each patient.

This means we can “translate” the information from one echocardiographic view to the other, gathering all measures in one single plane, which allows us to use these data in the geometrical calculations described in the next section.

Geometrical Model of the Interplay Between mAVVI, RV/LV Inflow Angle and VSD Size

In the apical 4C view, the AVV, the VSD, and the right and left ventricular inflows constitute the edges of a triangle with the following sides: the AVV side (AVVs), the left primary annulus side (LPAs), and the right inflow side (RIs). The AVVs consists of a line extended between both AVV annulus hingepoints, the LPAs in a line from the left hingepoint of the AVV annulus to the crest of the interventricular septum and the RIs between the right hingepoint of the AVV annulus and the crest of the interventricular septum (Figure 4). The main triangle has also three angles. The most important one is called “α” and is located over the crest of the interventricular septum. It represents the right ventricle (RV)/left ventricle (LV) inflow angle described by Cohen et al. ³ Four points are also defined in this triangle: A is the vertex corresponding to the α angle, L is the left vertex, R is the right vertex, and M is the point at which a line passing through vertex A crosses the AVVs perpendicularly. This segment is VSD side (VSDs) and divides the AVVs in two portions, the so-called left component side and the right component side. Vertex A represents the crest of the interventricular septum, vertex L the left hingepoint of the AVV annulus, vertex R the right hingepoint of the AVV annulus, and M the projection of the ventricular septum that divides the AVV in left and right components.

OPEN IN VIEWER

Figure 4.

(A) Apical four-chamber (4C) view of a patient with uAVSD showing the triangle; (B) the vertices, angles, and sides of the triangle are depicted. A indicates vertex of α angle; AVVs, AVV side; L, left vertex; LCs, left component side; LPAs, left primary annulus side; M, point at which a line passing through vertex A crosses the AVVs perpendicularly; R, right vertex; RCs, right component side; RIs, right inflow side; VSDs, ventricular septal defect side; α, RV/LV inflow angle; uAVSD, unbalanced atrioventricular septal defect.

From the trigonometric point of view, some mathematical relations can be established to describe the geometry of the AVV, the VSD, and the right and left ventricular inflows in the apical 4C view. This view is depicted in Figure 4, in which VSDs is represented as a perpendicular line extended from vertex A to AVVs, dividing the main triangle into two secondary right-angled triangles. Therefore, applying the Pythagoras theorem, we have:

Equation 1 describes the geometry of the left secondary right-angled triangle, whereas equation 2 defines the right secondary right-angled triangle. A third equation can be obtained from the main nonright-angled triangle with vertices R, L, and A by means of the cosine rule:

These three equations describe the geometry of the three triangles that can be identified in this area of the heart. As explained before, the mAVVI measured in the SLAO view and the left component diameter/total AVV diameter ratio in the apical 4C view are almost equal. Therefore, using two elements of these triangles, the mAVVI can be calculated as follows:

These equations will be used below to describe the interplay between mAVVI, RV/LV inflow angle, and VSD size.

Echocardiographic Protocol

Echocardiographic evaluation of patients with uAVSD should include the following four measures: left and total AVV area in SLAO view and VSD and AVV diameter in apical 4C view. First, in order to characterize the AVSD as balanced or unbalanced, the mAVVI is calculated. In SLAO view, end diastole, one frame before AVV closure, the AVV is evaluated in a frontal plane. A line is drawn between the crest of the muscular septum and the tip of the infundibular septum, bisecting the AVV and thereby assigning a portion to each ventricle. Left and total AVV areas are measured. The mAVVI is defined as the left valve area divided by the total AVV area. Average of three different measurements should be obtained. Second, in the apical 4C view, end diastole, one frame before AVV closure, the VSD and AVV annulus diameter are assessed. To measure the AVV annulus, a line is drawn between the two hingepoints of the AVV. The VSD diameter is acquired by tracing a line from the crest of the ventricular septum to the previously drawn line of the AVV annulus, achieving a 90° angle between the two lines. Again, the average of three different measurements should be obtained. The size of the VSD is divided by the AVV annulus diameter in the apical 4C view to obtain the new index described in the next section.

A Novel Metric: The Indexed VSD

We introduce a new index developed with the purpose of estimating the precise size of the VSD regardless of patient’s weight. The variable is called “indexed VSD” (inVSD) and is defined as the size of the VSD divided by the left to right AVV annulus diameter measured in the apical 4C view. Using the triangle depicted in Figure 4, it is calculated as follows:

The five aforementioned equations share in common components of the three described triangles, and they can be algebraically combined and expressed as a single final equation involving only mAVVI, inVSD, and α. Since the present study is focused on the behavior of the inVSD and its correlation with the RV/LV inflow angle and the mAVVI, the equation is presented in a form that allows finding the inVSD for given values of mAVVI and RV/LV inflow angle:

We entered different combinations of arbitrary values of mAVVI and RV/LV inflow angle into equation 6 to calculate the inVSD, as depicted in Table 1. For the mAVVI, we chose intervals of 0.03 in the range between 0 and 0.2 and intervals of 0.02 in the range between 0.2 and 0.4. For the RV/LV inflow angle, the values ranged from 80° to 170° at intervals of 10°.

OPEN IN VIEWER

Table 1.

Indexed VSD as a function of mAVVI and RV/LV inflow angle.^a

mAVVI / α	80°	90°	100°	110°	120°	130°	140°	150°	160°	170°
0.02	0.25	0.14	0.07	0.04	0.03	0.02	0.01	0.01	0	0
0.05	0.32	0.21	0.14	0.10	0.07	0.05	0.03	0.02	0.01	0
0.08	0.37	0.27	0.19	0.14	0.10	0.08	0.05	0.04	0.02	0.01
0.11	0.41	0.31	0.23	0.18	0.13	0.10	0.07	0.05	0.03	0.01
0.14	0.44	0.34	0.26	0.20	0.16	0.12	0.09	0.06	0.04	0.02
0.17	0.47	0.37	0.29	0.23	0.18	0.14	0.10	0.07	0.05	0.02
0.20	0.49	0.40	0.32	0.25	0.20	0.16	0.12	0.08	0.05	0.02
0.22	0.51	0.41	0.33	0.27	0.21	0.17	0.13	0.09	0.06	0.03
0.24	0.52	0.42	0.34	0.28	0.22	0.18	0.13	0.10	0.06	0.03
0.26	0.53	0.43	0.35	0.29	0.23	0.18	0.14	0.10	0.06	0.03
0.28	0.54	0.45	0.36	0.30	0.24	0.19	0.15	0.11	0.07	0.03
0.30	0.55	0.45	0.37	0.31	0.25	0.20	0.15	0.11	0.07	0.03
0.32	0.56	0.46	0.38	0.31	0.25	0.20	0.16	0.11	0.07	0.03
0.34	0.57	0.47	0.39	0.32	0.26	0.21	0.16	0.12	0.08	0.04
0.36	0.57	0.48	0.40	0.33	0.27	0.21	0.17	0.12	0.08	0.04
0.38	0.58	0.48	0.40	0.33	0.27	0.22	0.17	0.12	0.08	0.04

Abbreviations: AVV, atrioventricular valve; inVSD, indexed ventricular septal defect; mAVVI, modified atrioventricular valve index; RV/LV, right ventricle/left ventricle; uAVSD, unbalanced atrioventricular septal defect; VSD, ventricular septal defect.

^aDifferent combinations of values of mAVVI and RV/LV inflow angle (α) are entered to obtain the respective inVSD in the uAVSD group. This variable represents the size of the VSD indexed to the AVV diameter (0.5 means that the VSD is exactly half the size of the AVV).

Subsequently, we built a graphic of inVSD versus mAVVI and depicted the curves for different angles at intervals of 10° for the uAVSD group (Figure 5).

OPEN IN VIEWER

Figure 5.

The indexed ventricular septal defect (inVSD) is plotted as a function of modified atrioventricular valve index (mAVVI) for different values of the right ventricle/left ventricle (RV/LV) inflow angle. Slashed gray lines show the results from 70° to 170° in steps of 10°. The curves for 81°, 108°, and 135° are depicted in full lines. Notice the reduction in the inclination of each curve as the mAVVI grows. For each curve, variation in the inVSD is much higher between mAVVI of 0 and 0.2 than between 0.2 and 0.39.

Setting the Limits of the inVSD on a Clinical Basis

As a first step toward assessing the clinical applicability of this new index, we decided to employ the data published in the multi-institutional study of 112 patients conducted by the Congenital Heart Surgeons’ Society in 2013. ³ In the Supplemental Table of this article, the authors reported the mean and standard deviation (SD) of many variables on a large sample of patients with presumed right dominant uAVSD compared to another sample of patients with presumed balanced AVSD, basing the definition on the mAVVI. Among all the variables that were studied, they found that the RV/LV inflow angle was a strong data-derived predictor of right dominant uAVSD. Mean value of the angle in uAVSD group was 108°. The values of mean plus and minus 1 SD were 135° and 81°, respectively. As this study included a considerably large number of patients and a robust statistical analysis, these values might be representative of the uAVSD population. We shall then consider that the normal range of the inVSD for these patients is delimited by the curves that correspond to the mean angle plus and minus 1 SD, that is, 135° and 81°, respectively. We show the curves for these three values of the RV/LV inflow angle in Figure 5.

mAVVI, inVSD, and Surgical Decision-Making

The size of the VSD is a key point as it reflects mortality within the gray zone of mAVVI between 0.2 and 0.39. Our geometrical model gathers mAVVI, RV/LV inflow angle, and VSD in a single triangle. This allows us to mathematically correlate the inVSD with the mAVVI, which in turn proved to accurately predict unbalance. The two variables are logical factors affecting inflow dynamics of the LV.

The importance of the Congenital Heart Surgeons’ Society study relies on the fact that it has addressed the RV/LV inflow angles of one of the largest samples of patients with uAVSD. We used published mean and SD of this variable to identify different regions within the gray zone, assuming that this cohort represents the population of patients with uAVSD. This is an indirect method of confining the limits of the inVSD, strictly based on the cohort studied by the Congenital Heart Surgeons’ Society. Clinical validation of the normal range of this novel index is still needed before correlating it with individual outcomes.

The regions are determined by the curves of the mean value of the angle plus and minus 1 SD (Figure 7). Patients with a small inVSD (smaller than 0.16 if the mAVVI is 0.2 or smaller than 0.22 if the mAVVI is 0.39) fall below the 135° curve in the mAVVI versus inVSD graphic. These patients have an LV that is already supporting the systemic circulation by itself, as the VSD is too small to unload the LV to the RV and the pulmonary circulation and therefore could be considered suitable for BVR, especially those with mild unbalance.

OPEN IN VIEWER

Figure 7.

Suggested approach in relation to the indexed ventricular septal defect (inVSD) and the modified atrioventricular valve index (mAVVI).

There are two regions located around the mean angle of 108° within the gray zone. The first one gathers patients with inVSD between the mean angle and +1 SD. These cases, with angle more obtuse than the mean for uAVSD, show an inVSD between 0.16 and 0.28 if the mAVVI is 0.2, or an inVSD between 0.22 and 0.37 if the mAVVI is 0.39. In such context, we would probably recommend to consider BVR, as these patients exhibit an inVSD smaller than the mean expected value.

The second region, between the mean angle and the −1 SD curves, includes patients displaying an inVSD larger than the mean for uAVSD, between 0.28 and 0.49 if the mAVVI is 0.2, or between 0.37 and 0.57 if the mAVVI is 0.39. This could perhaps be the most ambiguous area for surgical decision-making, because the VSD is large but not large enough to consider UVR. We think that alternative strategies should be taken into consideration in this subset of patients, rather than strictly choosing between UVR and BVR. Foker et al described a technique to induce catch-up growth of a hypoplastic LV by undergoing a staged repair design to increase flow across the AVV. It consisted of partial closure of the VSD and the atrial septal defect along with pulmonary artery banding. This would keep the RV at systemic pressure in order to augment the systemic output through the planned residual intracardiac shunting. After adequate LV growth, the VSD and the atrial septal defect were closed, and the pulmonary artery banding was removed. ⁵

The area above the 81° curve includes those patients with an inVSD larger than 0.49 if the mAVVI is 0.2, or larger than 0.57 if the mAVVI is 0.39. Such a large VSD might preclude BVR, as it is associated with high mortality. Therefore, UVR should be considered, particularly in those cases displaying mAVVI near 0.2 within the gray zone. In instances of mild unbalance, a large inVSD may not by itself mandate UVR, but it should at least confer uncertainty with respect to the choice of a BVR strategy. The decision whether to pursue a single ventricle track should be considered more seriously if the inVSD is larger than 0.9 for mAVVI of 0.2 or larger than 0.97 for mAVVI of 0.39. These values correspond to an RV/LV inflow angle of mean−2 SD, that is, 54°. Patients with mAVVI < 0.2 show severe unbalance and shall undergo UVR, independently of the VSD size.

In addition to the mAVVI and the inVSD, some other criteria must obviously be taken into consideration, especially for difficult cases. These include ductal dependence of the systemic circulation, AVV straddling, apex-forming LV, LV/RV volumes, and the left ventricular inflow index. ⁶ Besides, trisomy 21, AVV regurgitation and pulmonary hypertension play an important role, as they are related to higher morbidity and mortality after UVR. Figure 8 shows a suggested algorithm for the management of uAVSD based on the mAVVI and the inVSD.

OPEN IN VIEWER

Figure 8.

Theoretical algorithm for the management of unbalanced atrioventricular septal defect (uAVSD).

Study Limitations

The main limitation of this study is that the inVSD has not yet been tested against clinical outcomes. We used the Congenital Heart Surgeons’ Society patient data presented by Cohen et al to set plausible normal limits of the RV/LV inflow angle for the uAVSD population. This angle and the inVSD are mathematically correlated in an almost unequivocal way. However, normal limits of both indexes and their relevance in terms of surgical decision-making still need direct validation. Individual clinical outcomes need to ratify the utility of this new metric before it can be employed for clinical recommendations.