Three-Dimensional Color Doppler Transesophageal Echocardiography for Mitral Paravalvular Leak Quantification and Evaluation of Percutaneous Closure Success

Three-Dimensional Color Doppler Transesophageal Echocardiography for Mitral Paravalvular Leak Quantification and Evaluation of Percutaneous Closure Success Eduardo Franco, MD, Carlos Almerıa, MD, PhD, Jose Alberto de Agustın, MD, PhD, Viviana Arreo del Val, MD,  Jose Juan G omez de Diego, MD, PhD, Miguel Angel Garcıa Fernandez, MD, PhD, Carlos Macaya, MD, PhD, Leopoldo Perez de Isla, MD, PhD, and Eulogio Garcia, MD, PhD, Madrid, Spain

Background: Three-dimensional (3D) color Doppler transesophageal echocardiography (TEE) enables accurate planimetry of the effective regurgitant orifice (ERO) of a mitral paravalvular leak (PVL). The aim of this study was to evaluate the usefulness of this method to quantify paravalvular regurgitation and to assess percutaneous PVL closure success, compared with 3D planimetry of PVLs without using color-flow images (3D anatomic regurgitant orifice [ARO]). Methods: Forty-six patients (59 mitral PVLs) who underwent 3D TEE to evaluate the indication of PVL closure procedure were retrospectively included. Receiver operating characteristic curves were compared to identify degree III and IV paravalvular regurgitation of 3D color ERO and 3D ARO measures. Forty patients underwent percutaneous PVL closure procedures; analysis was conducted to determine whether the undersizing of the closure devices according to 3D color ERO and 3D ARO measures was associated with PVL closure failure. Results: Three-dimensional ERO measures showed better areas under the curve than 3D ARO measures and correlated better with the degree of paravalvular regurgitation. Three-dimensional color ERO major diameter $ 0.65 cm showed a positive predictive value of 87.1% and a negative predictive value of 94% to diagnose degree III and IV paravalvular regurgitation. For the 40 patients who underwent PVL closure procedures, the immediate technical success rate was 76.9%, and 1-year estimated survival was 69.5%. Closure device undersizing according to 3D color ERO length, but not other PVL measurements, was significantly associated with PVL closure failure (P = .007). Conclusion: Three-dimensional ERO was superior to 3D ARO at identifying the presence of degree III and IV paravalvular regurgitation. The undersizing of closure devices according to 3D color ERO length was associated with failed closure procedures. Confirmatory prospective studies are encouraged. (J Am Soc Echocardiogr 2014;27:1153-63.) Keywords: Paravalvular leak, Three-dimensional transesophageal echocardiography, Paravalvular mitral regurgitation, Effective regurgitant orifice, Percutaneous paravalvular leak closure procedure

Paravalvular leaks (PVLs) are a common complication after mitral valve replacement surgery.1 Three-dimensional (3D) transesophageal echocardiography (TEE) has emerged as the preferred imaging modality to evaluate the morphology and extent of these leaks, as it better demonstrates, compared with two-dimensional (2D)

From the Cardiovascular Institute, Hospital Clınico San Carlos, Madrid, Spain (E.F., C.A., J.A.A., J.J.G.D., M.A.G.F., C.M., L.P.I., E.G.); and the Paediatric Cardiology Department, Hospital Universitario La Paz, Madrid, Spain (V.A.V.). Reprint requests: Eduardo Franco, MD, Cardiovascular Institute, Hospital Clınico San Carlos, Calle Profesor Martin Lagos s/n, 28040 Madrid, Spain (E-mail: [email protected]). 0894-7317/$36.00 Copyright 2014 by the American Society of Echocardiography. http://dx.doi.org/10.1016/j.echo.2014.08.019

echocardiography, the shapes and sizes of the leaks.2 Threedimensional TEE is also the recommended technique to guide transcatheter PVL closure procedures.3-6 However, severity assessment of paravalvular regurgitation caused by PVL using 3D TEE is technically difficult. Common measurements of mitral regurgitation severity, such as jet width (vena contracta) and jet area on color-flow imaging, are useful in this context,7,8 but other parameters, such the proximal isovelocity surface area radius, have not been validated. PVL sizing can be performed by 3D transesophageal echocardiographic planimetry without using color-flow images, by measuring the areas of echo dropout that represent the presence of a leak.6 This method actually measures the anatomic regurgitant orifice (ARO) size rather than the effective regurgitant orifice (ERO). Moreover, it may overestimate or underestimate the sizes of PVLs, as the extension of these anechoic areas depends on parameters 1153

1154 Franco et al

such as the gain and compression values used for the acquisition of ARO = Anatomic regurgitant images, and resolution may be orifice limited.2 Three-dimensional transesoAUC = Area under the curve phageal echocardiographic CI = Confidence interval planimetry using color-flow images may improve PVL sizing, as ERO = Effective regurgitant it enables accurate planimetry of orifice the ERO of the leaks. Our aim PVL = Paravalvular leak was to evaluate the usefulness of ROC = Receiver operating this method to quantify paravalvcharacteristic ular regurgitation and its capacity to assess percutaneous PVL 3D = Three-dimensional closure success. We also analyzed 2D = Two-dimensional the differences between PVL dimensions calculated with the 3D color ERO and 3D ARO approaches, to detect if either of the two methods was associated with systematic over- or underestimation. Abbreviations

METHODS Study Population We retrospectively included every consecutive patient with a diagnosis of mitral PVL who underwent 3D TEE at our center to consider the indication of a PVL closure procedure between January 2009 and July 2011. No method for sample size calculation was used. Instead, we studied every available subject. All information was recorded in a dedicated database, and all echocardiographic images were stored in Digital Imaging and Communications in Medicine 3.0 and row data in a digital system. Of the 53 patients initially included, seven lacked the 3D transesophageal echocardiographic color-flow images necessary for the quantification of 3D color ERO and were excluded from the analysis. Hence, a total of 46 patients, with a total of 59 mitral PVLs, were included in the analysis. Of these 46 patients, 40 (52 leaks) underwent percutaneous transcatheter PVL closure procedures. All procedures were performed with 3D transesophageal echocardiographic guidance, and the devices used in all cases were Amplatzer Vascular Plug III occluders (St Jude Medical, Plymouth, MN). At our institution, these procedures are performed in all cases using both retrograde aortic and antegrade transseptal approaches, with the creation of an arteriovenous rail through the leak to snare and exteriorize the wire, so that more support for sheath and device delivery is achieved. In 31 patients, the retrograde aortic approach was first performed to cross the leak with the wire. In the remaining nine patients, the antegrade transseptal approach was first selected to cross the leak; in one of these patients, transseptal puncture was not possible, so only the retrograde aortic approach without the use of an arteriovenous rail was used. The size selection of the device was carried out intraprocedurally using a sizing approach that measured the 3D dimensions of the leak without using color imaging and without image postprocessing. Color-flow images were also used, but because no postprocessing was carried out, the 3D color ERO measures (or the 3D ARO measures) were not calculated then. The operator then selected, among the available device sizes, the one or ones that best fit the size of the defect. Twenty-eight patients (34 leaks) underwent postoperative follow-up 3D TEE at our center and were included in the analysis of PVL closure success. The study complied with the Declaration of Helsinki.

Journal of the American Society of Echocardiography November 2014

Demographic and Clinical Variables Demographic and clinical data from patients were retrospectively collected from their medical records. In patients who underwent PVL closure, we identified three possible indications for the procedure: heart failure, symptomatic hemolytic anemia, or a combination of the two. Echocardiographic Technique and Measurements Conventional 2D transthoracic echocardiographic studies were performed before TEE in all patients, to measure atrial and ventricular dimensions and left ventricular ejection fraction. All transesophageal echocardiographic studies were conducted with a Philips iE33 ultrasound system and an X7-2t transesophageal transducer (Philips Medical Systems, Andover, MA). A first approach with 2D TEE was used to locate the best planes. Afterward, a 3D subvolume that included the mitral valve and the left atrial walls was recorded. Additional 3D zoom loops focused on the locations of the PVLs were also acquired. Next, 3D color-flow loops focused on paravalvular regurgitation were recorded. All 3D transesophageal echocardiographic loops included at least three cardiac cycles. Three-dimensional transesophageal echocardiographic loops were processed using QLAB (Philips Medical Systems). The number of PVLs, as well as their dimensions (length, width, and area), calculated by planimetry of the areas of echo absence (3D ARO), were recorded; to measure the dimensions of each PVL, a 2D plane was extracted from the 3D subvolume of interest using the QLAB multiplanar reconstruction tool (Figure 1). To calculate the 3D color ERO, the 3D transesophageal echocardiographic color-flow loops were processed using the multiplanar reconstruction tool of QLAB (similarly to the analysis of 3D subvolumes), to select the 2D plane that represented the ERO. Monochromatic color-flow mode was selected to facilitate the measurements. Once the best 2D plane was selected, planimetry of the colored area was performed to calculate the length, width, and area of the 3D color ERO (Figure 2). The procedure was feasible for 58 of the 59 PVLs. We also calculated the eccentricity index, defined as the ratio between the length and width of the PVL, using both 3D ARO and 3D color ERO measurements. The location of each PVL was defined with respect to the aortic valve, using a clocklike approach similar to that used in heart surgery,9,10 with the aortic valve represented as 12 o’clock, the left atrial appendage represented as 9 o’clock, the posterior mitral annulus represented as 6 o’clock, and the interatrial septum represented as 3 o’clock (Figure 3). Anterior location was defined for PVLs situated between 12 and 3 o’clock, septal location for PVLs between 3 and 6 o’clock, posterior location for PVLs between 6 and 9 o’clock, and lateral location for PVLs between 9 and 12 o’clock. The paravalvular regurgitation severity of each PVL was established as mild (degree I), mild to moderate (degree II), moderate (degree III), or severe (degree IV), according to the recommendations for native valve regurgitation.11 To set the degree of paravalvular regurgitation of each PVL, the mean degree of severity using several parameters was calculated (Table 1): the jet width, length, and area on color-flow images; the relationship between the regurgitant jet area and the left atrial area; and the pulmonary vein pulsed-wave Doppler pattern. The grading was performed by an expert echocardiographer, who also conducted the intraprocedural transesophageal echocardiographic studies used to guide the PVL closure procedures. To guarantee independent grading of paravalvular regurgitation, the

Journal of the American Society of Echocardiography Volume 27 Number 11

Franco et al 1155

Figure 1 Three-dimensional ARO quantification. Three-dimensional transesophageal echocardiographic loops are recorded for this purpose. A midsystolic frame is selected for quantification (A); the position of the leak that will be measured is highlighted (arrow). Using the multiplanar reconstruction tool in QLAB (Philips Medical Systems), the 2D plane that includes the anatomic orifice of the leak is extracted (B). Finally, the length, width, and area of this anatomic orifice (3D ARO) are measured using planimetry (C).

Figure 2 Three-dimensional color ERO quantification of the same leak as in Figure 1. Three-dimensional transesophageal echocardiographic color-flow loops are recorded for this purpose (A); this patient had two leaks (asterisks), and the one being measured is highlighted (arrow). Using QLAB (Philips Medical Systems) with a monochromatic color-flow mode, the loops are played, and the frame in which the origin of the regurgitant jet is best visualized is selected. Then, the multiplanar reconstruction tool allows the selection of the 2D plane that best shows the regurgitant orifice (B). Finally, the length, width, and area of the colored surface of the selected 2D plane, which represents the 3D color ERO, are measured using planimetry (C). In this case, the 3D color ERO measures were slightly higher than the 3D ARO measures. echocardiographer was unaware of PVL dimensions by 3D ARO and 3D color ERO approaches. Analysis of PVL Closure Success Rates The analysis of closure success was performed for the 28 patients (34 leaks) who underwent postprocedural follow-up 3D TEE sat our center. Technical success of the procedure was defined as the correct deployment of the device or devices, a reduction in the previous degree of paravalvular regurgitation, the lack of significant residual regurgitation (degrees III and IV), and the absence of new prosthetic valve malfunction.12 To determine if PVL size influenced the probability of technical success of the closure procedure, we compared the PVL measures (calculated with the 3D color ERO approach and with 3D ARO) of the successfully closed PVLs and those of the unsuccessfully closed PVLs. Moreover, we compared the rates of technical success for PVLs in different locations. We also explored if the choice of an undersized device could result in a failed closure procedure. To do so, we

compared the rates of technical success for undersized devices (defined as those devices with length, width, or area shorter than the PVL dimensions, calculated with the 3D color ERO method and with 3D ARO) to the rates obtained with well-sized devices (those with greater or equal dimensions than the PVL). For the entire cohort of 40 patients who underwent percutaneous PVL closure, we calculated the mortality rate at last follow-up and the survival estimates with 95% confidence intervals (CIs) at 6 months and 1 year, using the Kaplan-Meier method. We also collected functional New York Heart Association functional class, plasma hemoglobin level, and the number of patients who subjectively reported clinical improvement. Statistical Analysis Categorical variables are described as number (percentage) and were compared by using c2 or the Fisher’s exact tests, as appropriate. Continuous variables are described as mean 6 SD for variables with normal distributions or as median (interquartile range)

1156 Franco et al

Journal of the American Society of Echocardiography November 2014

Table 3. Three-dimensional color ERO measurements could be performed in 58 PVLs. Of these PVLs, three (5.1%) caused degree I paravalvular regurgitation, 25 (43.1%) caused degree II paravalvular regurgitation, 15 (25.9%) caused degree III paravalvular regurgitation, and 15 (25.9%) caused degree IV paravalvular regurgitation. Three-dimensional color ERO measures correlated moderately with 3D ARO measures, with intraclass correlation coefficients of 0.67 (95% CI, 0.38–0.83) for area, 0.52 (95% CI, 0.10–0.74) length, and 0.54 (95% CI, 0.18–0.74) for width. The length (major diameter) of the PVLs was systematically overestimated using 3D ARO compared with the 3D color ERO approach (mean difference, 0.27 cm; 95% CI, 0.1–0.45; P = .003). The calculated width (minor diameter) and area of the PVLs were similar using the two methods (width: mean difference, 0.05 cm [P = .056]; area: mean difference, 0.01 cm2 [P = .659]).

Figure 3 PVL locations. AO, Aortic valve; IAS, interatrial septum; LAA, left atrial appendage. for variables not normally distributed. The Kolmogorov-Smirnov test was used to assess normality in continuous variables. Comparisons among normal continuous variables were made using Student’s t tests (two-group comparisons) or analysis of variance (for comparisons among more than two groups); for variables not normally distributed, Mann-Whitney U tests (two-group comparisons) and Kruskal-Wallis tests (for comparisons among more than two groups) were used. Bilateral P values < .05 were considered statistically significant. We performed a reliability analysis of the 3D color ERO measures, in comparison with 3D ARO measures. To do so, the intraclass correlation coefficients for PVL dimensions using both methods were calculated, and the presence of systematic over- or underestimation of PVL dimensions by any of the two methods was evaluated using Student’s paired t test. For the validation analysis of 3D color ERO to assess the severity of paravalvular regurgitation, receiver operating characteristic (ROC) curves were calculated for each evaluated echocardiographic parameter to correctly diagnose the presence of moderate or severe (degrees III and IV) paravalvular regurgitation. Optimal cutoff values for each parameter were defined as those with the shorter distance to the top left corner of the ROC graph. To compare the capacity of the evaluated echocardiographic parameters to determine the severity of paravalvular regurgitation, Spearman correlation coefficients (r) between the degree (I–IV) of paravalvular regurgitation and each echocardiographic variable were calculated. Statistical analysis was performed using SPSS PASW Statistics version 15.0 package (SPSS, Inc, Chicago, IL). The comparison of the areas under the curve of ROC curves was performed using MedCalc version 13.3.3 (MedCalc Software, Mariakerke, Belgium).

RESULTS A total of 46 patients (59 mitral PVLs) were studied. Demographic and clinical characteristics of these patients are shown in Table 2. The mean time from the last mitral valve replacement surgery to mitral PVL diagnosis was 10.9 6 8.7 years. Echocardiographic measurements and mitral PVLs characteristics at diagnosis are shown in

Validation of the 3D Color ERO to Evaluate Paravalvular Regurgitation Severity Figure 4 shows the calculated ROC curves for 3D color ERO and 3D ARO measures to correctly diagnose the presence of degree III and IV paravalvular regurgitation. Table 4 indicates the area under the curve (AUC) of each parameter. Only PVL dimensions calculated with the 3D color ERO method significantly predicted the presence of degree III and IV paravalvular regurgitation, but not those obtained with the 3D ARO approach. PVL major diameter and PVL area calculated with the 3D color ERO method had the highest AUC values (0.846 and 0.806, respectively), although no significant differences between the AUC of the different 3D color ERO measures were noted (area vs length, P = .479; area vs width, P = .164; length vs width, P = .214). Table 5 shows the optimal cutoff value of each evaluated parameter to detect the presence of degree III and IV paravalvular regurgitation. Three-dimensional color ERO area $ 0.13 cm2 could predict with 83.5% positive predictive value and 81.6% negative predictive value the presence of degree III and IV paravalvular regurgitation. Three-dimensional color ERO major diameter $ 0.65 cm showed a positive predictive value of 87.1% and a negative predictive value of 94%. These results must be interpreted with caution, as the sensitivity, specificity, and predictive values were calculated in the same patients used to obtain the optimal cutoff values, which represents a ‘‘best case’’ scenario and may provide overoptimistic results. Table 6 shows the 3D color ERO dimensions obtained in PVLs with different degrees of paravalvular regurgitation. Left ventricular dimensions, atrial dimensions, and pulmonary artery systolic pressure in patients with different overall degrees of paravalvular regurgitation are also summarized. The dimensions were significantly lower for degree I and II paravalvular regurgitation than for degree III and IV (P < .001 for area and length, P = .003 for width), but none of the studied parameters could reliably distinguish between degree III and degree IV paravalvular regurgitation (Figure 5). Comparison between 3D Color ERO Measures and 3D ARO Measures to Assess the Severity of Paravalvular Regurgitation The dimensions of the PVLs, calculated either with the 3D color ERO or the 3D ARO method, were significantly correlated with the degree of paravalvular regurgitation of each PVL (Table 7). The Spearman correlation coefficients were higher for 3D color ERO dimensions than for 3D ARO measures. The PVL eccentricity index calculated

Franco et al 1157

Journal of the American Society of Echocardiography Volume 27 Number 11

Table 1 Grading of paravalvular regurgitation Mild PVR

Moderate PVR

Severe PVR

Color Doppler imaging Jet width (vena contracta) (cm) Jet length

<0.3

0.3–0.69

Short central jet

<4

4–9.9

$10

<10

10–19.9

$20

Jet area (cm2) Jet area/left atrial area (%)*

$0.7 Central jet: reaches the top of the atrium or penetrates into the pulmonary veins; eccentric jet: swirling in the left atrium

Pulsed Doppler Pulmonary vein flow

Dominant systolic wave

Dominant diastolic wave

Systolic wave reversal

PVR, Paravalvular regurgitation. When there were parameters consistent with both mild and moderate PVR, degree II (mild to moderate) was used, unless the vast majority of criteria indicated moderate regurgitation (degree III was then used). When there were parameters consistent with both moderate and severe PVR, the degree used was the one with more parameters indicating it, unless the vena contracta clearly indicated severe mitral regurgitation (degree IV was then used). *Because the vast majority of mitral prostheses have dilated left atria, we chose jet areas < 10% of left atrial area as a marker of mild regurgitation (instead of <20%) and jet areas $ 20% of left atrial area as a marker of severe regurgitation (instead of $40%).

Table 2 Patients’ demographic and clinical characteristics Variable

Value

Demographic characteristics Age at PVL diagnosis (y)

64.8 6 10.6

Men

19 (41.3%)

Medical history Arterial hypertension Diabetes mellitus Hypercholesterolemia Smokers Current smokers Former Chronic renal failure (%) Atrial fibrillation Ischemic cardiomyopathy

16 (34.8%) 9 (19.6%) 17 (37.0%) 8 (17.4%) 3 (6.5%) 5 (10.9%) 8 (17.4%) 29 (63.0%) 5 (10.9%)

Mitral valve disease history Age at first mitral valve replacement surgery (y)

47.6 6 16.4

Number of reoperated patients

22 (47.8%)

Age at current mitral valve replacement surgery (y)

53.2 6 13.2

Number of mechanical prosthesis

43 (93.5%)

Prosthesis size (mm)

27 6 1.96

Data are expressed as mean 6 SD or as number (percentage).

with the 3D color ERO approach, but not with the 3D ARO approach, also achieved significant correlation. There were no significant differences between the degree of paravalvular regurgitation of each PVL depending on its location (P = .343). Utility of the 3D Color ERO to Evaluate the Success of Percutaneous PVL Closure Procedures Forty patients, 15 men (37.5%), underwent PVL closure procedures. The mean age at PVL closure was 64.8 6 10.6 years, and the median

European System for Cardiac Operative Risk Evaluation score was 13.9 (interquartile range, 8.2–23.6). The indications for PVL closure were heart failure (six patients [15%]), hemolytic anemia (seven patients [17.5%]), and a combination of the two (27 patients [67.5%]). No procedure related deaths occurred. Closure procedures were attempted for 52 leaks; in one PVL, two simultaneous closure devices were implanted (with immediate technical success); in three PVLs, after the deployment of a closure device, the persistence of significant regurgitation led to the implant of other device; and in four PVLs, after an unsuccessful first closure procedure with a single device, another procedure was scheduled. Attending to the first closure procedure of these 52 leaks, immediate technical success was achieved in 40 leaks (76.9%). The causes of the unsuccessful closures were persistence of degree III (nine PVLs) or IV (one PVL) paravalvular regurgitation and inability to deploy the closure device (two PVLs). No interference with prosthetic valve function was noticed. In the successfully closed defects, no residual regurgitation (10 PVLs), degree I regurgitation (18 PVLs), or degree II regurgitation (14 PVLs) was detected. Three patients were lost to follow-up. For the remaining 37 patients, after a median of 7.4 months (interquartile range, 3.2–15.7 months) of follow-up, the survival rate was 75.7%. Three patients (8.1%) died of noncardiac causes, and six patients (16.2%) died of cardiac causes (three patients of sudden cardiac death and three patients of refractory heart failure). The estimates for 6-month and 1year survival were, respectively, 78.8% (95% CI, 64.4%–93.2%) and 69.5% (95% CI, 51.9%–87.1%) (Figure 6). Twenty-eight patients remained alive at last follow-up, 20 of them (71.4%) reporting clinical improvement. New York Heart Association functional class was I in seven patients (7.1%), II in 15 patients (53.6%), III in 10 patients (35.7%), and IV in one patient (3.6%). The mean hemoglobin level nonsignificantly increased from 9.54 6 1.83 g/dL (before the PVL closure procedure) to 10.56 6 1.75 g/dL. One patient developed mitral prosthetic endocarditis 5 months after the PVL closure procedure and received surgical treatment. For the 28 patients (34 leaks) who underwent follow-up 3D TEE at our center, technical success of the procedure at follow-up was achieved for 23 leaks (67.6%). Follow-up 3D TEE was carried out a median of 5.3 months (interquartile range, 2.0–7.0 months) after the

1158 Franco et al

Journal of the American Society of Echocardiography November 2014

Table 3 Two-dimensional echocardiographic characteristics and 3D transesophageal echocardiographic PVL measurements at PVL diagnosis Variable

Value

2D echocardiographic measures End-diastolic thickness of the septum (cm)

1.29 6 0.29

End-diastolic thickness of the posterior wall (cm)

1.21 6 0.32

Left ventricular end-diastolic diameter (cm)

5.02 6 0.74

Left ventricular end-systolic diameter (cm)

3.16 6 0.74

Left ventricular end-diastolic volume (mL)

123.9 6 45.6

Left ventricular end-systolic volume (mL)

49.5 6 34.3

LVEF (%)

61.9 6 14.1

Left atrial dimension (cm)

5.86 6 1.1

Left atrial area (cm2)

39.3 6 12.6

Pulmonary artery systolic pressure (mm Hg) (n = 26)*

54.1 6 14.4

Number of leaks Patients with 1 PVL

31 (67.4%)

Patients with 2 PVLs

12 (26.1%)

Patients with 3 PVLs

3 (6.5%)

Patients with >3 PVLs

0

Figure 4 ROC curves for 3D color ERO measures and 3D ARO measures to correctly diagnose the presence of moderate or severe (degrees III and IV) paravalvular regurgitation.

Location of leaks Septal (12–3 o’clock)

12 (20.3%)

Posterior (3–6 o’clock)

8 (13.6%)

Lateral (6–9 o’clock)

20 (33.9%)

Anterior (9–12 o’clock)

19 (32.2%)

Table 4 Areas under the ROC curves of the PVL dimensions calculated with the 3D color ERO approach and with 3D ARO to correctly identify degree III and IV paravalvular regurgitation

3D transesophageal echocardiographic PVL measures: planimetry without color images

AUC

95% CI

P

3D color ERO approach

Area (cm2)

0.23 6 0.19

Length (major diameter) (cm)

1.03 6 0.62

Area

0.806

0.635–0.977

.003

Width (minor diameter) (cm)

0.39 6 0.13

Length (major diameter)

0.846

0.696–0.996

.001

Eccentricity index (length/width)

2.87 6 1.67

Width (minor diameter)

0.756

0.576–0.935

.014

Area

0.673

0.480–0.866

.097

3D planimetry without color images

3D transesophageal echocardiographic PVL measures: 3D color ERO Area (cm )

0.20 6 0.12

Length (major diameter)

0.615

0.403–0.828

.269

Length (major diameter) (cm)

0.72 6 0.30

Width (minor diameter)

0.671

0.485–0.858

.101

Width (minor diameter) (cm)

0.33 6 0.13

Eccentricity index (length/width)

2.31 6 1.13

2

LVEF, Left ventricular ejection fraction. Data are expressed as mean 6 SD or as number (percentage). *Pulmonary artery systolic pressure was estimated only in patients with tricuspid regurgitation (n = 26).

intervention. The unsuccessful closures were due to the persistence of degree III (seven PVLs) or IV (one PVL) paravalvular regurgitation or to a nonreduction in previous degree II paravalvular regurgitation (three PVLs). No interference with prosthetic valve function or inability to deploy the devices was observed in this cohort of patients. Interestingly, no significant differences between the dimensions of the PVLs with or without technically successful closure were noticed (Table 8). Also, the eccentricity index did not differ significantly. The PVL location did not influence the rate of technical success. Comparing the immediate degree of paravalvular regurgitation after the closure procedure with that on follow-up 3D TEE in

these 34 leaks, 17 leaks had the same degree of paravalvular regurgitation, whereas in 13 leaks, increases in the degree of paravalvular regurgitation were noticed, and in four leaks, decreases were found (P = .054 for the presence of a difference between the immediate and follow-up degrees of paravalvular regurgitation). As for the size of the selected devices, all devices with shorter major dimension than the 3D color ERO length resulted in unsuccessful closures. The undersizing of the device according to other 3D color ERO measures, and to all 3D ARO measures, did not result in significantly greater rates of closure technical failure (Table 9).

DISCUSSION New 3D transesophageal echocardiographic techniques are emerging to improve the echocardiographic assessment of different structures and offer guidance for various transcatheter

Franco et al 1159

Journal of the American Society of Echocardiography Volume 27 Number 11

Table 5 Best cutoff values to identify degree III and IV paravalvular regurgitation of the PVL dimensions calculated with the 3D color ERO method and with 3D ARO Cutoff value

Sensitivity (%)

Specificity (%)

PPV (%)

NPV (%)

$0.13 cm2

90.0

71.4

83.5

81.6

3D color ERO approach Area Length (major diameter)

$0.65 cm

90.0

78.6

87.1

94.0

Width (minor diameter)

$0.28 cm

75.0

71.4

80.9

63.9

3D planimetry without color images Area

$0.15 cm2

70.0

57.1

72.6

54.3

Length (major diameter)

$0.84 cm

70.0

57.1

72.6

54.3

Width (minor diameter)

$0.34 cm

75.0

50.0

70.7

55.4

NPV, Negative predictive value; PPV, positive predictive value.

Table 6 Three-dimensional color ERO dimensions in PVLs with degree I, degree II, degree III, and degree IV PVR: left ventricular dimensions, atrial dimensions, and pulmonary artery systolic pressure in patients according to their overall degree of PVR Variable

Degree I PVR

Degree II PVR

Degree III PVR

Degree IV PVR

P

3D color ERO dimensions (PVLs) Area (cm2)

0.06 6 0.02

0.15 6 0.09

0.24 6 0.11

0.27 6 0.14

.001

Length (major diameter) (cm)

0.32 6 0.31

0.59 6 0.24

0.76 6 0.14

0.98 6 0.34

<.001

Width (minor diameter) (cm)

0.19 6 0.25

0.29 6 0.11

0.40 6 0.11

0.37 6 0.16

.014

2D echocardiographic measures (patients) Left ventricular end-diastolic diameter (cm)

5.6*

4.76 6 0.68

5.15 6 0.74

5.14 6 0.81

.420

Left ventricular end-systolic diameter (cm)

3.6*

2.99 6 0.63

3.24 6 0.66

3.22 6 0.94

.764

Left ventricular end-diastolic volume (mL)

152.0*

105.5 6 32.5

126.4 6 52.0

139.7 6 50.4

.270

Left ventricular end-systolic volume (mL)

55.0*

38.9 6 20.1

43.1 6 15.6

65.4 6 50.7

.249

Left atrial dimension (cm) Pulmonary artery systolic pressure (mm Hg) (n = 26)

5.80* *

5.75 6 0.99

6.09 6 1.06

5.83 6 1.37

.932

55.8 6 14.6

59.0 6 16.1

48.8 6 12.9

.370

PVR, Paravalvular regurgitation. Data are expressed as mean 6 SD. *Only one patient had overall degree I PVR, whereas three PVLs were classified as causing individual degree I PVR. This patient did not have tricuspid regurgitation to estimate pulmonary artery systolic pressure.

procedures.13,14 In this work, we have tested the utility of a novel 3D transesophageal echocardiographic technique in the field of mitral PVLs. On the basis of our data, 3D TEE with color-flow images is a feasible method to calculate the dimensions of PVLs, including the ERO (3D color ERO). A previous report suggested the utility of 3D TEE with color-flow images to calculate the length and width of PVLs in eight patients, but the areas of dehiscence were not calculated.15 The use of 3D color ERO, according to our results, shows the following features: good predictive value to identify the presence of degree III and IV paravalvular regurgitation, better correlation with the degree of paravalvular regurgitation than the dimensions of the PVL calculated with 3D ARO, and a possible capacity to predict the technical failure of percutaneous closure procedures when the devices are undersized according to the 3D color ERO major diameter. Three-dimensional TEE is useful for detecting and locating mitral PVLs, but color-flow images are needed to confirm the existence of PVLs.5,16 In our experience, using these color-flow loops to calculate

the dimensions of the 3D color ERO might be a better tool for PVL sizing than the 3D ARO approach. First, 3D ARO was associated with a systematic oversizing of PVL length. This may also be considered a systematic underestimation of PVL length using the 3D color ERO approach, but this parameter correlated better with the degree of paravalvular regurgitation than the length calculated with the 3D ARO method (r = 0.571 vs r = 0.384), suggesting that it better represents the actual major dimension of the PVLs. In vitro valvular regurgitation models have shown that the ERO dimensions are smaller than the corresponding ARO dimensions,17 which is consistent with our findings. This is due to the flow contraction phenomenon that blood experiences when it abruptly enters the ARO.18 Moreover, the undersizing of a closure device according to the 3D color ERO length, but not to the length of the PVL calculated with 3DARO, significantly contributed to technical failure of percutaneous closure procedures. These findings suggest, at least, that the major dimensions of PVLs might be more precisely calculated using the 3D color ERO approach.

1160 Franco et al

Journal of the American Society of Echocardiography November 2014

Table 7 Correlation between the dimensions of each PVL and its degree of paravalvular regurgitation Spearman’s r

P

0.523

<.001*

Length (major diameter)

0.571

<.001*

Width (minor diameter)

0.345

.008*

Eccentricity index

0.283

.031*

Variable

3D color ERO approach Area

3D planimetry without color images Area

0.463

.002*

Length (major diameter)

0.384

.011*

Width (minor diameter)

0.375

.007*

Eccentricity index

0.163

.309

*Statistically significant (P < .05).

Figure 6 Survival free of all-cause death.

Figure 5 Comparison between 3D color ERO dimensions in patients with degree I, II, III, and IV paravalvular regurgitation. Each box represents the interquartile range of each measure, with the median represented as the intermediate horizontal line. Vertical lines denote the minimum and maximum values, and dots represent outliers.

As for paravalvular regurgitation assessment, 3D ARO measures did not significantly predict the presence of degree III and IV paravalvular regurgitation, according to the areas under the ROC curves, whereas all 3D color ERO dimensions (length, width, and area), which obtained good AUC values, did. The best predictive values for detecting degree III and IV paravalvular regurgitation were 3D

color ERO length $ 0.65 cm (positive predictive value, 87.1%; negative predictive value, 94%) and 3D color ERO area $ 0.13 cm2 (positive predictive value, 83.5%; negative predictive value, 81.6%). These predictive values, however, may be overoptimistic, as they were calculated using the same patients used to calculate the cutoff values (‘‘best-case’’ scenario). Hence, our findings suggest that the 3D color ERO dimensions may be used in the process of quantification of paravalvular regurgitation, but prospective studies are needed to confirm our results and validate the cutoff values we have calculated, which may not apply when used prospectively. Before doing the statistical analysis, we had expected that the area of the 3D color ERO would correlate better with the degree of paravalvular regurgitation than the major dimension. We have not found a satisfactory response for this finding. Perhaps the resolution of 3D volumetric imaging, or the crescentic shapes of many PVLs, could explain this observation, but we have not been able to identify the reason. Interestingly, no measure could distinguish between degree III and degree IV paravalvular regurgitation. Current guidelines recommend PVL closure in the presence of symptoms, without establishing the need of a certain degree of paravalvular regurgitation.19 Hence, we

Franco et al 1161

Journal of the American Society of Echocardiography Volume 27 Number 11

Table 8 Influence of PVL dimensions and location on the technical success of PVL closure procedures

Variable

PVLs with PVLs with successful unsuccessful closure (n = 23) closure (n = 11)

P

3D color ERO approach Area (cm2)

0.18 6 0.10

0.20 6 0.17

.466

Length (major diameter) (cm)

0.67 6 0.25

0.73 6 0.42

.854

Width (minor diameter) (cm)

0.32 6 0.10

0.34 6 0.19

.288

Eccentricity index

2.18 6 0.68

2.42 6 1.36

.454

Area (cm2)

0.20 6 0.19

0.27 6 0.27

.696

Length (major diameter) (cm)

1.06 6 0.69

1.12 6 0.72

.639

Width (minor diameter) (cm)

0.36 6 0.14

0.41 6 0.14

.701

Eccentricity index

3.19 6 2.10

2.58 6 1.31

.490

Anterior

5 (62.5%)

3 (37.5%)

.987

Septal

2 (66.7%)

1 (33.3%)

Posterior

9 (69.2%)

4 (30.8%)

7 (70%)

3 (30%)

3D planimetry without color images

PVL location

Lateral

Quantitative data are expressed as mean 6 SD. PVL location is shown as number (percentage).

do not consider the impossibility to distinguish between degree III or IV paravalvular regurgitation to be a major limitation. Some readers may be concerned about the fact that a ‘‘moderate to severe’’ degree of paravalvular regurgitation was not included in our study, such that there was a wide range of severity between degrees III and IV. The reason we did not consider such a degree is based on the fact that PVL closure is indicated on the basis of clinical findings. Symptomatic patients with moderate (and ‘‘more than moderate’’) paravalvular regurgitation probably have the origin of their symptoms related to the paravalvular regurgitation, whereas in patients with symptoms and ‘‘less than moderate’’ paravalvular regurgitation, other causes of symptoms might be more probable and should be carefully ruled out before indicating a closure procedure. Hence, we thought that it would be more useful to distinguish between mild to moderate and moderate paravalvular regurgitation than between moderate and moderate to severe paravalvular regurgitation. In our series, 40 patients (52 leaks) underwent percutaneous closure procedures. The immediate technical success rate of these procedures was similar to the rate observed in other large series.20 Ruiz et al.21 reported a higher immediate technical success rate, but this can be explained because their series did not consider residual regurgitation as a marker of procedural failure. Our success rate was, however, a bit higher than the success rates of other series.22 Nevertheless, the estimate for 1-year survival (68.3%) in our series was slightly lower.20,23 If we consider the technical success rate in our series at 3D transesophageal echocardiographic follow-up, it decreases to 67.6%, which may be due to the nonsignificant trend toward an increase in the degree of paravalvular regurgitation that was found from the closure procedure to last follow-up. No differences in the technical success rate of the procedure according to the PVLs location were

noted, whereas previous series had reported higher success rates for posterior PVLs and worse results in other locations.22 According to our results, the dimensions of the PVLs did not influence the technical success rate of the closure procedures. The variety of available sizes of the Vascular Plug III devices may be responsible for this finding. According to previous reports, the use of Vascular Plug III devices (instead of other types) may be associated with higher technical success rates.24 The eccentricity index of the PVLs was also unhelpful to predict closure success, contrasting with a previous series that had proposed that crescent-shaped (i.e., eccentric) PVLs could be more difficult successfully close.6 However, that series used Amplatzer Patent Ductus Arteriosus occluders (St Jude Medical) instead of Vascular Plug III devices, which have a more elongated shape. Anyway, the development of novel devices specifically designed to close PVLs may improve technical success rates and would be of great interest. Finally, we explored whether the undersizing of the closure devices according to PVL dimensions was associated with the technical failure of the procedures. All devices smaller than the 3D color ERO length (significantly) or the 3D color ERO area (nonsignificant trend) resulted in failed closures. Undersizing according to 3D ARO measures, however, was not associated with closure success rate. These findings suggest that 3D color ERO length may be a useful tool to safely select the closure device size and improve technical success rates. However, the limited sample size of our analysis warrants careful interpretation of our results. Prospective studies are needed to confirm this hypothesis. Limitations The main limitation of our study is that we included only patients who had undergone 3D TEE at our center. These were highly selected patients who had been referred for evaluation of the need for intervention on PVLs because of symptomatic regurgitation. Hence, our findings cannot be extrapolated to other populations. Also, this was a retrospective study. Hence, prospective studies are needed to confirm our findings. Another limitation, as in all clinical studies on mitral regurgitation, is the lack of a gold standard to compare the results of different methods. Prospectively designed studies with predefined echocardiographic criteria are required to establish valid severity markers. Moreover, the current 3D color Doppler imaging techniques offer lower frame rates than conventional 2D echocardiography. In this study, 3D color ERO was calculated using a single systolic frame, assuming that it corresponds to the time of maximum regurgitant flow. This approach can be limited by the lower temporal resolution of 3D color Doppler, so the selected ERO may not necessarily be the absolute largest. To minimize the potential effect of low temporal resolution, several loops were analyzed in each patient, looking for the largest ERO in each loop. Finally, the results of the analysis of PVL closure success rates need careful interpretation, because of the limited available sample size.

CONCLUSIONS Three-dimensional color ERO quantification of mitral PVLs is a feasible technique that offers a more precise approach to PVL assessment than 3D ARO. Three-dimensional color ERO dimensions were superior in predicting the presence of degree III and IV paravalvular regurgitation, with good areas under of the ROC curves.

1162 Franco et al

Journal of the American Society of Echocardiography November 2014

Table 9 Influence of undersized closure devices in the technical failure of PVL closure procedures Variable

Number of devices

Devices with successful closure

Devices with unsuccessful closure

Device area $ 3D color ERO area

32

23 (71.9%)

9 (28.1%)

Device area < 3D color ERO area

2

P

3D color ERO approach

Device length $ 3D color ERO length

30

Device length < 3D color ERO length

4

0 (0) 23 (76.7%) 0 (0)

.098

2 (100%) 7 (23.3%)

.007

4 (100%)

Device width $ 3D color ERO width

22

15 (68.1%)

7 (31.8%)

Device width < 3D color ERO width

12

8 (66.7%)

4 (33.3%)

27

18 (66.7%)

9 (33.3%) 2 (28.6%)

1.0

3D planimetry without color images Device area $ 3D ARO area Device area < 3D ARO area

7

5 (71.4%)

Device length $ 3D ARO length

26

19 (73.1%)

Device length < 3D ARO length

8

4 (50%)

7 (26.9%)

1.0 .388

4 (50%)

Device width $ 3D ARO width

18

13 (72.2%)

5 (27.8%)

Device width < 3D ARO width

16

10 (62.5%)

6 (37.5%)

.545

Data are expressed as number (percentage). One PVL was closed using two devices (the remaining PVLs were closed with one device); the length, width, and area of these devices have been summed, so the two devices have been considered one device.

Furthermore, the undersizing of closure devices according to 3D color ERO length was associated with failed closure procedures; thus, 3D color ERO dimensions may be useful to select closure devices size. 8.

REFERENCES 1. Chen YT, Kan MN, Chen JS, Lin WW, Chang MK, Hu WS, et al. Detection of prosthetic mitral valve leak: a comparative study using transoesophageal echocardiography, transthoracic echocardiography, and auscultation. J Clin Ultrasound 1990;18:557-61. 2. Zamorano JL, Badano LP, Bruce C, Chan KL, Goncalves A, Hahn RT, et al. EAE/ASE recommendations for the use of echocardiography in new transcatheter interventions for valvular heart disease. Eur Heart J 2011; 32:2189-214. 3. Hamilton-Craig C, Boga T, Platts D, Walters DL, Burstow DJ, Scalia G. The role of 3D transoesophageal echocardiography during percutaneous closure of paravalvular mitral regurgitation. JACC: Cardiovasc Imaging 2009;2:771-3. 4. Becerra JM, Almeria C, de Isla LP, Zamorano J. Usefulness of 3D transoesophageal echocardiography for guiding wires and closure devices in mitral perivalvular leaks. Eur J Echocardiogr 2009;10:979-81. 5. Hagler DJ, Cabalka AK, Sorajja P, Cetta F, Mankad SV, Bruce CJ, et al. Assessment of percutaneous catheter treatment of paravalvular prosthetic regurgitation. J Am Coll Cardiol Imaging 2010;3:88-91. 6. Garcia-Fernandez MA, Cortes M, Garcia-Robles JA, Gomez de Diego JJ, Perez-David E, Garcia E. Utility of real-time three-dimensional transoesophageal echocardiography in evaluating the success of percutaneous transcatheter closure of mitral paravalvular leaks. J Am Soc Echocardiogr 2010; 23:26-32. 7. Zoghbi WA, Chambers JB, Dumesnil JG, Foster E, Gottdiener JS, Grayburn PA, et al. Recommendations for evaluation of prosthetic valves with echocardiography and Doppler ultrasound: a report from the American Society of Echocardiography’s Guidelines and Standards Committee and the Task Force on Prosthetic Valves, developed in conjunction with the American College of Cardiology Cardiovascular Imaging Committee, Cardiac Imaging Committee of the American Heart Association, the European Association of Echocardiography, a registered branch of the European Society of Cardiology, the Japanese Society of

9.

10.

11.

12.

13.

14.

15.

16.

17.

Echocardiography and the Canadian Society of Echocardiography, endorsed by the American College of Cardiology Foundation, American Heart Association, European Association of Echocardiography, a registered branch of the European Society of Cardiology, the Japanese Society of Echocardiography, and Canadian Society of Echocardiography. J Am Soc Echocardiogr 2009;22:975-1014. Yildiz M, Duran NE, Gokdeniz T, Kaya H, Ozkan M. The value of real-time three-dimensional transoesophageal echocardiography in the assessment of paravalvular leak origin following prosthetic mitral valve replacement. Turk Kardiyol Dern Ars 2009;37:371-7. Meloni L, Ricchi A, Cirio E, Falchi S, Abbruzzese PA, Aru GM, et al. Echocardiographic assessment of aortic valve replacement with stentless porcine xenografts. Am J Cardiol 1995;76:294-6. De Cicco G, Russo C, Moreo A, Beghi C, Fucci C, Gerometta P, et al. Mitral valve periprosthetic leakage: anatomical observations in 135 patients from a multicenter study. Eur J Cardiothorac Surg 2006;30:887-91. Zoghbi WA, Enriquez-Sarano M, Foster E, Grayburn PA, Kraft CD, Levine RA, et al. Recommendations for evaluation of the severity of native valvular regurgitation with two-dimensional and Doppler echocardiography. J Am Soc Echocardiogr 2003;16:777-802. Kliger C, Eiros R, Isasti G, Einhorn B, Jelnin B, Cohen H, et al. Review of surgical prosthetic paravalvular leaks: diagnosis and catheter-based closure. Eur Heart J 2013;34:638-48. Hahn RT, Khalique O, Williams MR, Koss E, Paradis JM, Daneault B, et al. Predicting paravalvular regurgitation following transcatheter valve replacement: utility of a novel method for three-dimensional echocardiographic measurements of the aortic annulus. J Am Soc Echocardiogr 2013;26:1043-52. Goncalves A, Almeria C, Marcos-Alberca P, Feltes G, HernandezAntolin R, Rodriguez E, et al. Three-dimensional echocardiography in paravalvular aortic regurgitation assessment after transcatheter aortic valve implantation. J Am Soc Echocardiogr 2012;25:47-55. Biner S, Kar S, Siegel RJ, Rafique A, Shiota T. Value of colour Doppler threedimensional transoesophageal echocardiography in the percutaneous closure of mitral prosthesis paravalvular leak. Am J Cardiol 2010;105:984-9. Kim MS, Casserly IP, Garcia JA, Klein AJ, Salcedo EE, Carroll JD. Percutaneous transcatheter closure of prosthetic mitral paravalvular leaks: are we there yet? JACC Cardiovasc Interv 2009;2:81-90. Mascherbauer J, Rosenhek R, Bittner B, Binder J, Simon P, Maurer G, et al. Doppler echocardiographic assessment of valvular regurgitation severity by measurement of the vena contracta: an in vitro validation study. J Am Soc Echocardiogr 2005;18:999-1006.

Journal of the American Society of Echocardiography Volume 27 Number 11

18. Yoganathan AP, Cape EG, Sung HW, Williams FP, Jimoh A. Review of hydrodynamic principles for the cardiologist: applications to the study of blood flow and jets by imaging techniques. J Am Coll Cardiol 1988;12: 1344-53. 19. Vahanian A, Alfieri O, Andreotti F, Antunes MJ, Bar on-Esquivias G, Baumgartner H, et al. Guidelines on the management of valvular heart disease (version 2012): the Joint Task Force on the Management of Valvular Heart Disease of the European Society of Cardiology (ESC) and the European Association for Cardio-Thoracic Surgery (EACTS). Eur J Cardiothorac Surg 2012;42:S1-44. 20. Sorajja P, Cabalka AK, Hagler DJ, Rihal CS. Percutaneous repair of paravalvular prosthetic regurgitation: acute and 30-day outcomes in 115 patients. Circ Cardiovasc Interv 2011;4:314-21.

Franco et al 1163

21. Ruiz CE, Jelnin V, Kronzon I, Dudiy Y, Del Valle-Fernandez R, Einhorn BN, et al. Clinical outcomes in patients undergoing percutaneous closure of periprosthetic paravalvular leaks. J Am Coll Cardiol 2011;58:2210-7. 22. Cortes M, Garcia E, Garcia-Fernandez MA, Gomez JJ, Perez-David E, Fernandez-Aviles F. Usefulness of transoesophageal echocardiography in percutaneous transcatheter repairs of paravalvular mitral regurgitation. Am J Cardiol 2008;101:382-6. 23. Sorajja P, Cabalka AK, Hagler DJ, Rihal CS. Long-term follow-up of percutaneous repair of paravalvular prosthetic regurgitation. J Am Coll Cardiol 2011;58:2218-24. 24. Palacios IF, Arzamendi D. Structural heart intervention. Beyond transcatheter valve therapy. Rev Esp Cardiol (Engl Ed) 2012;65(Suppl 1):4-11.