Between October 2010 and December 2018, 26 consecutive patients were referred to our center for possible percutaneous PVL closure. We performed 33 percutaneous closure attempts in these 26 patients. Patient demographics and medical history revealed a high surgical risk: 14 patients (54%) were refused by the surgical team due to a multitude of comorbidities (four patients had ejection fraction <30%, two had marked right ventricular dysfunction and one refused blood transfusion due to religion), redo operations (four patients had more than one previous surgery) or early leak (<6 months after surgery – two patients).

All patients were symptomatic with severe heart failure and/or mechanical hemolytic anemia requiring multiple transfusions. Hemolysis alone was the indication for PVL closure in one case (3.8%), CHF alone was the indication in 18 procedures (69.2%), and a combination of hemolysis and CHF was the indication in the other seven procedures (26.9%).

Patients’ mean age was 65±13 years (range 21-81 years), and 14 (54%) were female. Regarding the leaks, 46.2% were in aortic and 53.8% were in mitral prostheses, 88.5% (23 patients) in mechanical prostheses, 7.7% in biological and 3.8% in transcatheter aortic valve implants.

Twenty-one patients had one prosthetic valve and five had two, mostly in the mitral and aortic position, except one patient with one mitral and one biological tricuspid prosthesis. Two patients had prior endocarditis and one patient had had a fistula between the left atrium and the non-coronary sinus since the first surgical procedure (Videos 1 and 2).

All mitral prostheses were studied with three-dimensional (3D) transesophageal echocardiography (TEE) (Phillips IE33, Eindhoven, Netherlands, and GE E9 and E95, GE Healthcare, Milwaukee, WI), and aortic prostheses with two-dimensional (2D)/3D TEE.

Twelve patients (46%) also underwent computed tomography angiography (CTA) on a 64-detector scanner (LightSpeed VCT, GE Healthcare, Milwaukee, WI).

PVL was defined as a regurgitant jet, demonstrated by color Doppler echocardiography, originating between the outer margin of the prosthetic sewing ring and the native tissues around the valve. The anatomic location of the leak was defined using the clock reference for precise location in 3D TEE left atrial en-face view, with the mitral-aortic curtain at 12 o’clock, the left atrial appendage at nine o’clock, the interatrial septum at 3 o’clock and the posterolateral free wall at 6 o’clock. Mitral leaks were classified as anterior (9-12 o’clock), inferior (6-9 o’clock), posterior (3-6 o’clock) and medial/septal (0-3 o’clock). For aortic leaks the anatomic location was defined with reference to the adjacent coronary sinus (left, right and non-coronary) with the non-coronary cusp between 7 o’clock and 11 o’clock, the left coronary cusp between 11 o’clock and 3 o’clock, and the right coronary cusp between 3 o’clock and 7 o’clock (Figure 1).

Figure 1.

Classification of paravalvular leaks according to reference points (12 o’clock: mitral-aortic curtain). IAS: interatrial septum; LAA: left atrial appendage.

(0.14MB).

For mitral leaks the maximum leak diameter and the anatomic regurgitant orifice were measured, and the shape of the leak was classified as round or crescent-shaped (like a comma surrounding the prosthetic sewing ring). Regurgitant jets were analyzed and mitral regurgitation classified as mild, moderate or severe in accordance with the European guidelines for regurgitant prosthetic jets.7 Twelve patients with mitral and aortic prostheses underwent CTA for better anatomic definition of the leak (Figure 2).

Figure 2.

Measurement of anatomic characteristics of the leak by three-dimensional transesophageal echocardiography and computed tomography angiography.

(0.28MB).

For aortic prostheses, 2D TEE in short-axis view, biplane images and 3D TEE of the aortic valve were used to localize the leak and measure its maximum diameter.

All procedures were performed by experts in TEE and interventional procedures under TEE guidance.

Leak closure was performed under general anesthesia. TEE was used to guide transseptal puncture and passage of the guidewire and catheter across the defect, and to analyze the position of the devices and any possible interference with the disks of the prosthesis.

Mitral valve PVLs were closed via a transvenous route with transseptal puncture in the anterograde direction. In 43% of mitral PVLs a previous arteriovenous loop was necessary to support the passage of the delivery sheath. Transapical access was also used, particularly when there were double prostheses in aortic and mitral position.

Aortic valve PVLs were crossed retrogradely with femoral access. Crossing the leak was facilitated by simultaneous use of fluoroscopy and 2D/3D TEE.

PVL closure devices were selected according to physician preference and to previous measurements by 3D TEE and CTA.

Technical success was defined as successful deployment of a closure device across the leak without interfering with the valve prosthesis and without acute conversion to surgery.

Procedural success was considered to have been achieved if there was resolution of the leak with no, trivial or mild regurgitation. Partial success was defined as the regurgitant jet decreasing by one degree, but with more than mild and less than severe residual regurgitation.

Clinical success was defined as improvement of at least one New York Heart Association (NYHA) functional class within six months and/or improvement in mechanical hemolysis.

Procedural outcome events were defined as events occurring during the procedure and in the following 24 hours. Thirty-day outcome events, as assessed by telephone or during a clinic visit, included procedural outcome events and those occurring within 30 days of the procedure, including death from any cause. All follow-up events were recorded and collected by telephone or outpatient clinic visit.

Continuous variables are presented as mean ± standard deviation and categorical variables as absolute numbers and percentages.

Baseline characteristics and outcomes were compared for categorical variables using the chi-square test or Fisher's exact test when appropriate. Pearson's correlation coefficient was used to test associations between continuous variables.

Changes in PVL grade and NYHA class at baseline and after closure were analyzed using the Wilcoxon rank test.

Survival estimates with 95% confidence intervals (CIs) were calculated using the Kaplan-Meier method.

Univariate and multivariate Cox regression was used to assess potential independent associations between outcome and death or major adverse cardiovascular events (MACE). The significance level was set at 0.05. Data were analyzed using IBM SPSS for Windows, version 19.0 (IBM SPSS Inc., Chicago, IL).

PVL closure was attempted in 26 patients in 33 procedures. The procedure was first performed at our center in 2010 and began to be performed regularly 18 months later, after approval by the ethics committee.

Patients’ clinical variables are detailed in Table 1. Mean age was 65±13 years. There were no significant differences between genders (54% female). The indications for closure were heart failure (NYHA class III or IV) in 18 patients (69.2%), heart failure and hemolysis in seven patients (26.9%) and hemolysis only in one patient (3.8%) (Table 1).

Fourteen patients presented risk factors for surgical PVL closure: four had >1 previous valve surgery, four had severe left ventricular dysfunction (one after TAVI for true low-flow, low-gradient severe aortic stenosis), two had right ventricular dysfunction (due to severe tricuspid regurgitation with hepatic dysfunction, one with a dysfunctional biological tricuspid prosthesis), two had recent valve surgery (less than six months) and two had endocarditis (one after repeated piercings and cutaneous infection, the other due to intravenous drug abuse). There was a relationship between hemolysis and >1 reoperation (p=0.004). Previous hemolysis was the only parameter that correlated with technical aspects such as the number of procedures by patient (p=0.049) and partial success/failure (p=0.046).

Three procedures (11.5%) were performed in an urgent setting.

No relationship was observed between patient characteristics and technical success. NYHA>II at baseline was associated with mortality in Cox univariate analysis (p=0.008, hazard ratio [HR] 3.334, 95% confidence interval [95% CI]: 1.375-8.087).

All patients underwent complete transthoracic and transesophageal echocardiographic assessment for diagnostic purposes. During the procedure all mitral PVLs were studied with 3D TEE, and aortic prostheses with 2D/3D TEE.

CTA was performed in 12 patients for additional characterization of the leak and TEE and CTA results were integrated for these patients. For mitral PVLs, 3D TEE gave precise anatomic details concerning the size, area and shape of the leak. Additionally, in a patient with double mitral and aortic mechanical prostheses, TEE revealed the presence of a fistula between the non-coronary sinus and left atrium that was not diagnosed by CTA. However, for aortic prostheses TEE did not provide sufficient information concerning the anatomic area of the leak and other anatomic details like outpouching (small pseudoaneurysms) of the left ventricular outflow tract, which was masked by the struts of mechanical prostheses. CTA provided more information in these cases, however the anatomic area and maximum diameter of the leak appeared oversized by CTA (Table 2).

Regurgitant jets were severe in 20 patients (10 mitral and 10 aortic), mild in one patient with a mitral prosthesis (hemolysis only) and moderate to severe in five (two aortic and three mitral).

No echocardiographic or CTA parameter was significantly associated with technical success. There was also no relationship between these parameters and mortality and MACE by univariate Cox proportional hazard regression.

Procedural variables are shown in Table 3. The majority of PVLs were around mechanical valves (88.5%), and mitral PVLs accounted for 53.8% of the cases. Devices were successfully implanted in 81% patients (n=21). The approach was anterograde for mitral prostheses in 11 patients, but in six patients it was necessary to perform a arteriovenous loop between the right femoral vein, left ventricle, aorta and left femoral artery beforehand. For aortic leaks the approach was retrograde (11 patients); an apical approach was adopted in three patients (two for mitral and one for aortic prostheses) and 11% of the patients required more than one procedure.

The devices chosen are listed in Table 3. The number of devices used per patient ranged from one to three. The Amplatzer Vascular Plug III (AVP III) accounted for 57% of the devices used. This is because the PVLs, though complex, were long and crescent-shaped in most of the series, which gives this oblong device an advantage. Occlutech has two types of device, one oblong and the other square. However, our experience with these devices is limited (two cases). Our data, comparing technical success between the AVP III and the other devices, were in favor of the AVP III (p=0.027). Procedures with more than one attempt were associated with partial success or failure (p=0.003).

There was one acute tamponade as an intra-procedural complication, solved using a percutaneous approach, with no adverse consequences. There were no procedure-related deaths. The only case of 30-day mortality was due to sepsis and infection, in a patient who had been in the intensive care unit for 15 days, so the procedure was urgent.

One-year mortality (one death) was due to infection (a case of endocarditis three months after leak closure).

Follow-up for patients surviving more than one year was 44±24 months (7-97 months).

Clinical improvement, defined as a reduction in NYHA functional class, was significant and associated with reduction of the leak (p=0.0001) (Figure 3). Cardiac-related events in follow-up (new hospital admissions, new cardiac surgery, need for transfusion) were more frequent in patients with partially successful or unsuccessful closure (p=0.012). Univariate Cox regression analysis revealed that NYHA>II at follow-up (p=0.03, HR 0.17 [95% CI 0.034-0.843]), residual PVL (p=0.046, HR 0.257 [95% CI 0.68-0.961]) and cardiac-related events (p=0.029, HR 0.95 [95% CI 0.11-0.791]) were predictors of a worse prognosis. In multivariate Cox analysis only cardiac-related events were independent predictors of mortality in follow-up (p=0.029, HR 0.095 [95% CI 0.11-0.791]). Kaplan-Meier analysis showing all-cause mortality and MACE comparing patients with successful versus partially successful closure is presented in Figure 4.

Figure 3.

(A) New York Heart Association functional class before and after percutaneous closure attempt; (B) severity of paravalvular leak (Wilcoxon signed rank test). NYHA: New York Heart Association.

(0.17MB).

Figure 4.

Kaplan-Meier cumulative survival plots according to closure success and severity of residual paravalvular leak at follow-up (A) and closure success and cardiac-related events in follow-up (B).

(0.17MB).