Flow-dynamics assessment of mitral-valve surgery by intraoperative vector flow mapping

Abstract

OBJECTIVES: We assessed vortex patterns and energy loss in left ventricular flow in patients who underwent mitral valve repair or replacement with bioprosthetic valves.

METHODS: Vector flow mapping was performed before and after the procedure in 15 and 17 patients who underwent repair and replacement, respectively. The preprocedure mitral-septal angle was measured in all patients. Relationships between vortex patterns or energy loss change (ELC) and annuloplasty ring or bioprosthetic valve sizes or the effect of mitral leaflet resection in the repair group were statistically analysed.

RESULTS: Normal vortex patterns were observed in 13 and 1 patients who underwent repair and replacement, respectively. Abnormal vortex patterns were observed in 2 and 16 patients who underwent repair and replacement, respectively. ELC was significantly higher in the replacement group (196.6 ± 180.8) than in the repair group (71.9 ± 43.9). In the repair group, preoperative mitral-septal angles in patients with normal vortex patterns (79.2° ± 3.4°) were significantly larger than those in patients with abnormal vortex patterns (67.5° ± 3.5°). No significant differences were observed in the effects of annuloplasty ring and bioprosthetic valve sizes on vortex patterns and ELC, and in the effect of mitral valve resection (80.4 ± 56.3) and respect (without leaflet resection) (53.8 ± 28.4) on ELC in the repair group.

CONCLUSIONS: Mitral valve replacement alters the intraventricular vortex pattern and increases flow energy loss. A small mitral-septal angle is a risk factor for abnormal vortex patterns after mitral valve repair surgery.

Left ventricle, Mitral valve, Mitral valve annuloplasty

INTRODUCTION

Eddy currents in the sinus of Valsalva were depicted in the 1500s despite the lack of imaging modalities [1]. Kilner et al. [2] illustrated the chirally asymmetric paths of flow and changes in flow direction through the heart chambers, redirecting the flow momentum towards the next cavity, using multislice cardiac magnetic resonance imaging. Thereafter, various cardiovascular flow-visualization technologies have been developed [3–5], including vector flow mapping, which was only recently developed and has shown reasonable accuracy in vitro and in vivo [6, 7].

Transmitral inflow generates vortex rings; these rings show a stronger anterior component and weaker posterior component in 2D planes, such as the left ventricular long-axis view of echocardiography. The posterior vortex soon diminishes; the anterior vortex continues until the isovolumic contraction phase. Blood flow streams into the left ventricle and rotates to the anteroseptal wall during diastole (Fig. 1) [8, 9]. After mitral valve replacement, left ventricular blood flow changes and energy loss increases [9–11]. To assess the relationship between flow dynamics and mitral valve surgeries, we investigated the primary hypothesis that the postoperative left ventricular vortex pattern differs between patients who have undergone repair and those who have undergone replacement of the mitral valve. Second, we investigated the effect of these mitral valve surgeries on left ventricular energy loss change (ELC) and the effect of the preoperative angle between the mitral annulus and left ventricular anteroseptal wall (Fig. 2A) on vortex patterns. We compared postoperative and preoperative conditions using intraoperative vector flow mapping.

Figure 1

Vortex change after mitral valve surgery. Normal vortex patterns in a normal left ventricle, normal vortex patterns after mitral valve repair and abnormal vortex patterns after mitral valve replacement are shown in early diastole (left), late diastole (middle) and systole (right). Blood flow becomes dissipative because of the collision of transmitral inflow with the anteroseptal wall and ejection flow with the bioprosthetic valve after mitral valve replacement.

Figure 2

Echocardiographic images of mitral-septal angle and postoperative vortices. (A) Mitral-septal angle. The angle between the anteroposterior axis of the mitral annulus and the long axis of the midportion of the anteroseptal wall before a procedure is shown. (B) Vector flow mapping image of normal vortex patterns after repair surgery. (C) Colour Doppler image of normal vortex patterns after repair surgery. The red arrow points to the complete opening of the PML. (D) Vector flow mapping image of abnormal vortex patterns after repair surgery. (E) Colour Doppler image of abnormal vortex patterns after repair surgery. The red arrow points to the restricted opening of the PML. (F) Vector flow mapping image of normal vortex patterns after replacement surgery. The only case that showed normal vortex patterns after replacement surgery. The mitral-septal angle was 90° before the procedure. (G) Vector flow mapping image of abnormal vortex patterns after replacement surgery showing the transmitral inflow colliding with the anteroseptal wall.

MATERIAL AND METHODS

Patients

This prospective study was approved by the institutional review board of our institution; written informed consent was obtained from all participating patients. We included 32 consecutive patients who underwent mitral valve repair or a replacement procedure with a bioprosthetic valve for severe mitral regurgitation from February 2015 to June 2016. The exclusion criteria were aortic valve regurgitation, aortic valve stenosis, mitral valve stenosis, ischaemic heart disease and arrhythmia. Fifteen patients underwent mitral valve repair, and 17 patients underwent mitral valve replacement with a bioprosthetic valve (Table 1). Energy loss in 7 patients (3 mitral valve repair; 4 mitral valve replacement) could not be calculated because of the low quality echocardiographic images.

Table 1

Baseline characteristics of patients

Characteristic	Repair (n = 15)	Replacement (n = 17)	P-value
Age, years	65 ± 12	71 ± 10	0.919
Male	11 (73)	7 (41)	0.09
Body surface area, m2	1.59 ± 0.2	1.54 ± 0.2	0.26
LVEDD, cm	5.5 ± 0.7	5.5 ± 1.2	0.533
LVESD, cm	3.5 ± 1.0	3.9 ± 1.5	0.832
LVEDV, ml	135 ± 36	136 ± 58	0.517
LVESV, ml	52 ± 37	64 ± 54	0.749
LVEF, %	63 ± 15	58 ± 16	0.155

Values are means ± SD or n (%).

LVEDD: left ventricular end-diastolic diameter; LVEDV: left ventricular end-diastolic volume; LVEF: left ventricular ejection fraction; LVESD: left ventricular end-systolic diameter; LVESV: left ventricular end-systolic volume; SD: standard deviation.

Image acquisition and processing

To compare the effect of mitral valve repair with that of mitral valve replacement using a bioprosthetic valve, left ventricular vortex and energy loss were assessed intraoperatively using vector flow mapping software. Digitized 2D colour Doppler cine-loop images obtained in the mid-oesophageal left ventricular long axis (ME-LAX) view by transoesophageal echocardiography were stored with the vector flow mapping configuration before and after the procedure. To equalize the load conditions, postoperative image acquisition was performed when the postoperative left ventricular diastolic diameter and heart rate were the same as those measured preoperatively. The ultrasound frequency was 5 MHz, with a frame rate of 30–40. The stored cine-loop images were analysed with a computer that was connected to the diagnostic ultrasound system, Aloka ProSound F75 Premier (Hitachi, Tokyo, Japan). The Nyquist limit for 2D colour Doppler imaging was set high enough to mitigate the aliasing phenomenon. One cardiac cycle was selected for analysis by determining 2 consecutive QRS complexes of the electrocardiogram as the beginning and ending points. The ventricular cavity–endocardial border was manually traced on the beginning frame; 2D speckle tracking was applied to detect cardiac wall motion. If the aliasing phenomenon was observed in the cine-loop images, the aliased pixels were manually corrected. Energy loss values were averaged over 3 cardiac cycles.

Principles of vector flow mapping

A novel technology, vector flow mapping, was applied to evaluate intracardiac flow dynamics. In this technology, a 2D continuity equation applied to colour Doppler echocardiography of blood flow and speckle-tracking echocardiography of the myocardium boundary allowed mathematical calculation of the velocity vectors of intraventricular flow [7, 12]. Briefly, the small pixels of the echocardiographic image are analysed using a continuity equation on the basis of the colour Doppler and speckle-tracking data. The 2 velocity vectors of each pixel that are calculated for the left-side and right-side boundaries are integrated by summation of the vectors according to a weight function [12]. Flow-energy loss, which increases if the size or direction of the adjacent velocity vectors changes, can be calculated according to the following equation:

$$\mathit{E}\mathit{n}\mathit{e}\mathit{r}\mathit{g}\mathit{y}\mathrm{ }\mathit{L}\mathit{o}\mathit{s}\mathit{s}=\int \mathit{\mu }\left.\left\{2{\left(\frac{\partial \mathit{u}}{\partial \mathit{x}}\right)}^2+2{\left(\frac{\partial \mathit{v}}{\partial \mathit{y}}\right)}^2+{\left(\frac{\partial \mathit{u}}{\partial \mathit{y}}+\frac{\partial \mathit{v}}{\partial \mathit{x}}\right)}^2\}\right.\right\}\mathbf{d}\mathit{A},$$

where μ is blood viscosity, u and v are velocity components of the Cartesian coordinates (x, y) and A is the area of the grid unit [12]. The kernel of the integral approximates the dissipated energy caused by viscous flow, which is integrated with the area increments, dA. The above equation expresses the total of the squares of the differences between neighbouring velocity vectors and energy loss increases at points where the size and direction of velocity vectors change. Because flow-energy loss is the rate of energy dissipation due to blood viscosity, flow-energy loss increases when the geometry of flow becomes less laminar.

Vortex patterns, energy loss change and mitral-septal angle

We defined 2 intraventricular vortex patterns: a normal pattern, defined as a counterclockwise rotating vortex, and an abnormal pattern defined as a clockwise rotating vortex in the ME-LAX view of transoesophageal echocardiography. We also defined ELC as the average energy loss of 1 post-procedure cardiac cycle divided by the average energy loss of 1 preprocedure cardiac cycle. Vortex patterns were assessed, and ELC was calculated using the vector flow mapping software. We named the angle between the mitral annulus and the anteroseptal wall of the left ventricular midportion in the ME-LAX view on transoesophageal echocardiography as the mitral-septal angle (Fig. 2A). The mitral-septal angle was measured during early diastole before the procedure.

After the operation, relationships between post-procedure vortex pattern and procedure type, ELC and procedure type and vortex pattern and mitral-septal angle were statistically analysed. Relationships between vortex pattern and annuloplasty ring size of the repair procedure, ELC and annuloplasty ring size of the repair procedure, vortex pattern and bioprosthetic valve size of the replacement procedure, ELC and bioprosthetic valve size of the replacement procedure, and ELC and the effect of mitral leaflet resection in the repair group were statistically analysed.

Statistical analysis

Statistical analysis was performed with JMP software (version 12.0.1 for Macintosh from SAS). Continuous variables are represented as means ± standard deviation. Welch’s t-test and the Wilcoxon rank sum test were used when appropriate. Categorical variables were compared by Fisher’s 2-tailed exact test. Spearman’s rank correlation was performed to evaluate the correlation between 2 numerical variables. P-values < 0.05 were considered to indicate significant differences.

RESULTS

Patient characteristics

Clinical characteristics of the patients are shown in Table 1. There were no significant differences in any of the baseline characteristics between the repair group and the replacement group.

Vortex pattern and energy loss change

Operative and analysis data from the repair and replacement groups are presented in Table 2. The preprocedure vortex pattern was normal in all 32 patients. After the repair procedure, 13 patients showed normal vortex patterns (Fig. 2B) and 2 patients showed abnormal vortex patterns (Table 2 and Fig. 2D). After the replacement procedure, 1 patient showed normal vortex patterns (Fig. 2F) and 16 patients showed abnormal vortex patterns (Table 2 and Fig. 2G). There were significant differences between the repair group and the replacement group in post-procedure vortex patterns (P < 0.0001, Fisher’s 2-tailed exact test).

Table 2

Operative and analysis data of the repair and replacement group

Case number	Vortex pattern	ELC	Repair procedure	Annuloplasty ring size	Mitral-septal angle, °
1	Normal	19.4	P2 resection	30	75
2	Normal	34.4	P2 resection	28	80
3	Normal	46.8	A1,2 neochordae	26	80
4	Normal	55.5	P2,3 neochordae	32	85
5	Abnormal	160	P3 plication	32	70
6	Normal	43.7	Ring only	30	75
7	Normal	64.3	P3 resection	30	75
8	Normal	23.1	P2,3 neochordae	32	80
9	Abnormal	77.9	P2 resection	30	65
10	Normal	119.4	P1 resection	34	85
11	Normal	117.7	AC resection	30	80
12	Normal	100	A2,3 P3 neochordae	30	80
13	Normal	Imp	P3 resection	33	75
14	Normal	Imp	A2 cleft suture	32	80
15	Normal	Imp	Ring only	26	80
Case number	Vortex pattern	ELC	Bioprosthetic valve type/ subvalvular apparatus preservation	Bioprosthetic valve size	Mitral-septal angle, °
1	Normal	91.6	MOSAIC/PML preserved	33	90
2	Abnormal	146.1	MOSAIC/PML preserved	29	80
3	Abnormal	265.7	MOSAIC/PML preserved	31	80
4	Abnormal	331.9	MOSAIC/not preserved	29	85
5	Abnormal	114.3	MOSAIC/PML preserved	29	70
6	Abnormal	31.4	CEP Magna/not preserved	30	65
7	Abnormal	201.2	MOSAIC/PML preserved	29	75
8	Abnormal	64.4	MOSAIC/PML preserved	29	80
9	Abnormal	110.5	MOSAIC/PML preserved	27	70
10	Abnormal	80.8	EPIC/PML preserved	27	80
11	Abnormal	181.4	MOSAIC/PML preserved	29	80
12	Abnormal	208	MOSAIC/PML preserved	29	70
13	Abnormal	728.6	MOSAIC/PML preserved	31	80
14	Abnormal	Imp	EPIC/PML preserved	29	80
15	Abnormal	Imp	MOSAIC/not preserved	29	70
16	Abnormal	Imp	MOSAIC/not preserved	27	70
17	Abnormal	Imp	MOSAIC/not preserved	27	70

ELC: energy loss change; Imp: impossible to calculate energy loss because of low-quality echocardiographic image; PML: posterior mitral leaflet.

ELC of the replacement group was significantly higher than that of the repair group (replacement group, 196.6 ± 180.8; repair group, 71.9 ± 43.9; P = 0.009, Wilcoxon rank sum test) (Fig. 3A).

Figure 3

Distribution of ELC and mitral-septal angle. Red dots represent cases with normal vortex patterns in the repair group. Red circles represent cases with abnormal vortex patterns in the repair group. Blue dots represent cases with normal vortex patterns in the replacement group. Blue circles represent cases with abnormal vortex patterns in the replacement group. (A) Comparison of ELC between the repair and replacement groups. *ELC was significantly higher in the replacement group than in the repair group (P = 0.009). (B) Comparison of mitral-septal angle between normal-vortex-pattern cases and abnormal-vortex-pattern cases in the repair and replacement groups. *The mitral-septal angle of abnormal-vortex-pattern cases was significantly smaller than that of normal-vortex-pattern cases in the repair group (P = 0.019).

Relationship between mitral-septal angle and vortex patterns

The measured mitral-septal angles are presented in Table 2. In patients, who underwent mitral valve repair, the mitral-septal angle was significantly smaller in the group with abnormal vortex pattern (67.5° ± 3.5°) than in the group with normal vortex pattern (79.2° ± 3.4°) (P = 0.019, Wilcoxon rank sum test) (Fig. 3B). In patients, who underwent mitral valve replacement, the mitral-septal angle was not significantly different between the group with normal vortex pattern (90°) and the group with abnormal vortex pattern (75.3° ± 5.9°) (P = 0.083, Wilcoxon rank sum test) (Fig. 3B).

Effect of size of annuloplasty ring and bioprosthetic valve

There was no significant difference in the effect of annuloplasty ring size on vortex patterns between the group with normal vortex pattern (31.0 ± 1.4 mm) and the group with abnormal vortex pattern (30.2 ± 2.5 mm) (P = 0.723, Wilcoxon rank sum test). There was no significant correlation between annuloplasty ring size and ELC (ρ = −0.380, P = 0.223, Spearman rank correlation).

There was no significant difference in the effect of bioprosthetic valve size on vortex patterns between the group with normal vortex pattern (33 mm) and the group with abnormal vortex pattern (28.8 ± 1.3 mm) (P = 0.075, Wilcoxon rank sum test). There was no significant correlation between bioprosthetic valve size and ELC (ρ = 0.222, P = 0.467, Spearman rank correlation).

Effect of resect or respect on energy loss change in the repair group

There was no significant difference in the effect of mitral valve resection (leaflet resection or plication) (80.4 ± 56.3) and respect (neochordae or ring only without leaflet resection) (53.8 ± 28.4) on ELC (P = 0.417, Wilcoxon rank sum test) in the repair group.

DISCUSSION

We demonstrated that in contrast to mitral valve repair, mitral valve replacement with a bioprosthetic valve altered the intraventricular vortex pattern and subsequently increased energy loss.

Vortices have an important role in normal cardiac function by keeping blood in motion inside the cardiac chambers, preserving momentum, avoiding excessive dissipation of energy, facilitating inflow into the ventricle and redirecting inflow towards the aortic valve [13–15]. The mechanism for forming vortices is as follows. When blood flows through a tubular structure, fluid layers at the centre of the jet move faster than those adjacent to the containing boundaries because of friction. If the boundaries disappear or abruptly expand, this friction generates a swirling tendency of the peripheral layers of fluid to spin away from the central jet. The transmitral inflow also generates a swirling flow at the tip of the leaflets and vortices form in the left ventricle. A longer anterior leaflet generates a stronger anterior vortex, and a shorter posterior leaflet generates a weaker posterior vortex. This leaflet’s asymmetry creates vortices of the best energetic performance [16]. If the mitral leaflets were the same size, energy dissipation would increase. Conversely, if the posterior leaflet was extremely short, energy dissipation would also increase because the flow would collide with the ventricular wall [16].

Abnormal flow patterns are observed in hearts with prosthetic valves in vivo [9, 17–19]. The Björk–Shiley prosthetic valve with an anterior orientation of the greater orifice reverses intraventricular flow and increases the mean prosthetic diastolic gradient [17]. In 38 of 40 cases of mitral valve replacement, the intraventricular flow is opposite to the normal pattern, and a significant increase in energy dissipation is found in a numerical model [18]. Patients with tissue prostheses show normal and abnormal blood-flow patterns, depending on the angle between the long axis of the prosthesis and septum [19]. Such an abnormal flow and increased energy loss in mitral valve replacement are partly responsible for the worse outcomes in the replacement group [20].

The superiority of mitral valve repair over replacement for short-term and long-term survival was established in the 1980s [21–28], but the mechanism is not fully understood. A possible factor contributing to its superiority is the preservation of the subvalvular apparatus in mitral valve repair, which maintains left ventricular geometry and allows for a reduction in left ventricular radius. The reduction in left ventricular radius leads to a reduction in wall stress according to the Laplace equation. Conversely, end-systolic wall stress is increased after mitral valve replacement without chordal preservation because of the loss of chordal support and the failure of left ventricular end-systolic volume to decrease [22–24]. Besides these structural defects, the intraventricular vortex and its energetic efficiency must affect the patient’s outcome [20, 29] because the intraventricular flow pattern in healthy subjects has a natural cardiac structure that is optimal in terms of minimization of energy dissipation [16].

Variety of repair techniques

Because patients with complicated mitral valve disease tend to be referred to our facility, the patients who had undergone mitral valve repair had been treated by various repair techniques. Although it would be simple to enrol only P2 repair patients, the repair techniques in this study were various because of the small number of enrolled cases. Studies are needed of patients undergoing the same repair technique.

Mitral-septal angle and vortex patterns

Here mitral valve repair tended to preserve vortex patterns and generate lower energy loss. Conversely, mitral valve replacement tended to alter vortex patterns and generate higher energy loss. This change in vortex patterns with mitral valve replacement is understandable. Because the bioprosthetic valve opens in a direction perpendicular to the mitral annular line, the transmitral inflow is towards the anteroseptal wall (Fig. 1). Conversely, the transmitral inflow after mitral valve repair drifts towards the apex due to steering by the long anterior mitral leaflet and complete opening of the posterior mitral leaflet (PML) (Figs 1 and 2C). However, there were 2 exceptional patients in the repair group whose post-procedure vortex patterns changed. They had undergone PML resection and plication. The rigidity of the PML caused by the resection restricted the opening of PML, and transmitral inflow seemed to stream towards the anteroseptal wall (Fig. 2E). The mitral-septal angle in the 2 exceptional repair cases did not exceed 70°. Patients who underwent mitral valve repair and whose preoperative mitral-septal angle is ≤ 70° are considered to be at risk of having an abnormal vortex.

In 1 exceptional case, vortex patterns were preserved in the mitral valve replacement group (Fig. 2F). In that case, the mitral-septal angle was 90°. Basically, in most cases, the preoperative mitral-septal angle was < 90°. To make the postoperative vortex normal, we hope that there was an angled (tilted) prosthetic valve with an angle that is 90° subtracted from the mitral-septal angle.

Energy loss change

Abnormal vortex patterns after the mitral valve replacement increase energy loss because of a more dissipative blood flow through the entire cardiac cycle (Figs 1 and 4). The transmitral inflow collides with the anteroseptal wall, and blood flow becomes dissipative during diastole [30]. The collision of blood flow with the prosthetic valve during the systolic phase generates unnecessary energy loss [11]. Although acceleration of the inflow and outflow generates energy loss mainly in the repair group, dissipation of the inefficient intraventricular vortex generates unnecessary energy loss in patients who undergo replacement.

Figure 4

Postoperative energy loss images. (Top) Post-procedure repair group. (Bottom) Post-procedure replacement group. Brightness indicates energy loss. The 6 phases of 1 cardiac cycle (early systole, mid-systole, late systole, early diastole, mid-diastole and late diastole) are shown from left to right. In the repair group, normal vortex patterns and efficient blood flow generate energy loss mainly because of flow acceleration. The bright yellow area (high-energy loss area) is concentrated on the outflow during systole and on the tips of the mitral leaflets during diastole in the repair group. Conversely, abnormal vortex patterns create dissipative blood flow and generate unnecessary energy loss in the replacement group. The bright area is dispersed around the prosthetic valve during systole and around the whole left ventricle during diastole in the replacement group.

Although mitral leaflet resection in the repair group possibly reduces effective orifice area and increases energy loss due to the increase of inflow velocity, there was no significant difference between the effects of leaflet resection and respect on ELC. In performing mitral leaflet resection, it is important to avoid rigidity of the posterior mitral leaflet.

Future prospects

In the case of mitral valve repair, intraoperative real-time vector flow mapping enables us to check the intraventricular vortex pattern and energy loss. If the vortex pattern is abnormal and energy loss is high, we can provide the operational direction whether a second pump run is needed or not.

Study limitations

Because intraventricular flow is a 3D entity, assessment of the flow in 3D would be desirable. However, 3D vector flow mapping is not currently available. Currently, the best imaging plane to analyse is that in the ME-LAX view, because it allows imaging of the maximum intraventricular area, including inflow and outflow.

Because of the relatively small number of subjects, our results should be considered preliminary findings that must be verified in a larger number of subjects.

CONCLUSIONS

Mitral valve replacement alters intraventricular vortex direction and increases flow-energy loss. Mitral valve repair tends not to alter vortex direction and not to increase flow-energy loss. In mitral valve repair cases, a careful procedure that does not impair the mobility of the posterior leaflet is required if the mitral-septal angle is ≤ 70°. Real-time vector flow mapping enables evaluation of vortex patterns and energy loss during surgery.

Conflict of interest: Keiichi Itatani is an endowed chair of Kyoto Prefectural University of Medicine, financially supported by Medtronic Japan and has a stock option of Cardio Flow Design. The other authors declare no conflict of interest.