Outcome and follow-up of aortic valve replacement with the freestyle stentless bioprosthesis

Outcome and Follow-up of Aortic Valve Replacement With the Freestyle Stentless Bioprosthesis A. Pieter Kappetein, PhD, Jerry Braun, MD, Leo H. B. Baur, PhD, Alain Prat, MD, Katinka Peels, PhD, Mark G. Hazekamp, PhD, Paul H. Schoof, MD, and Hans A. Huysmans, PhD Department of Thoracic Surgery, Leiden University Medical Center, Leiden, The Netherlands; Department of Cardiac Surgery, Center Hospitalier Regional, Lille, France; Department of Cardiology, Catharina Hospital, Eindhoven, The Netherlands; and Department of Cardiology, Leiden University Medical Center, Leiden, The Netherlands

Background. The aim of this study was to determine the morbidity, mortality, and hemodynamics after implantation of the Freestyle stentless bioprosthesis in the aortic position. Methods. A total of 280 patients were operated on from June 1993 to July 1999 as part of a multicenter investigation. Factors influencing hospital mortality and long-term survival were assessed by logistic regression and Cox proportional hazards analysis. Patients were evaluated postoperatively at discharge, at 3 to 6 months, and yearly by clinical examination and color flow Doppler echocardiography. Results. Hospital mortality in this group was relatively high (9.6%). Logistic regression analysis showed that cross-clamp time, age, myocardial infarction, diabetes, left ventricular hypertrophy, coronary artery disease, New York Heart Association class III or IV and female gender were the independent predictive factors. According to the Kaplan-Meier method, the 4-year survival for hospital survivors was 94%. In the multivariate Cox proportional hazard analysis, only coronary artery disease proved to be prognostic. During follow-up, 11 pa-

tients developed paravalvular leakage due to prosthetic dehiscence at the side of the noncoronary cusp. Performance of the prosthesis as assessed by echocardiography was excellent. Mean gradient decreased significantly between discharge and follow-up at 3 to 6 months. At 1-year follow-up trivial regurgitation was found in 6 patients (3%) and mild regurgitation in 4 (2%). Regurgitation did not increase with time. The effective orifice area increased significantly from discharge to follow-up at 3 to 6 months. Conclusions. Hospital mortality after implantation of a stentless bioprosthesis was higher compared to conventional prosthesis. A high incidence of prosthesis dehiscence at the proximal suture line was found, which was probably due to technique. Hemodynamic performance up to 3 years showed low transvalvular gradients. There is echocardiographic evidence for reduction of left ventricular hypertrophy and improvement of left ventricular function.

C

incomplete regression of hypertrophy has been shown to reduce long-term survival [3]. Homografts show excellent hemodynamic performance [4]. Clinical use of homografts however, is severely restricted by the limited availability of donor organs. Stentless aortic xenograft valves were introduced in clinical practice to overcome the problems of a stented valve and to imitate the hemodynamic performance of homografts. The lack of a rigid stent diminishes the danger of wear stress, paravalvular leakage, thromboembolism, and hemolysis [5].

urrently available mechanical and stented biological prostheses for aortic valve replacement have many disadvantages. Mechanical prostheses subject patients to lifelong anticoagulants, with risk for bleeding complications. On the other hand, stented bioprostheses have a higher failure rate. The stent may lead to nonphysiologic mechanical stress and thereby to leaflet calcification, with subsequent valve malfunction [1]. Furthermore the presence of a rigid stent contributes to a residual obstruction to transaortic flow by minimizing available flow area; this feature would inhibit complete resolution of left ventricular hypertrophy [2]. After aortic valve replacement, Presented at the Thirty-sixth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 31–Feb 2, 2000. Address reprint requests to Dr Kappetein, Department of CardioThoracic Surgery, Thoraxcenter, University Hospital Rotterdam, PO Box 2040, 3000 CA Rotterdam, The Netherlands; e-mail: kappetein@thch. azr.nl

© 2001 by The Society of Thoracic Surgeons Published by Elsevier Science Inc

(Ann Thorac Surg 2001;71:601– 8) © 2001 by The Society of Thoracic Surgeons

Financial support was provided by Medtronic (Minneapolis, MN) to the three participating departments (Leiden University Medical Center, Leiden, The Netherlands; Hoˆpital Cardiologique, Lille, France; and Catharina Hospital, Eindhoven, The Netherlands) for the clinical and echocardiographic follow-up of patients receiving a Medtronic Freestyle valve.

0003-4975/01/$20.00 PII S0003-4975(00)02519-4

602

KAPPETEIN ET AL PERFORMANCE OF FREESTYLE PROSTHESIS

The Freestyle stentless bioprosthesis is a nonstented bioprosthesis with a new antimineralization process and a new tissue fixation technique. The prosthesis can be inserted by the subcoronary, root inclusion, and total root techniques. In the present study we prospectively evaluated the hemodynamic performance of the Freestyle valve and the clinical outcome of implantation in 280 patients.

Material and Methods Between June 1993 and July 1999 280 patients underwent aortic valve replacement with the Freestyle aortic root prosthesis (Medtronic, Minneapolis, MN). The study cohort was part of a multicenter trial for the Freestyle stentless xenograft from the Leiden University Medical Center (Leiden, The Netherlands), the Hoˆpital Cardiologique (Lille, France), and the Catharina Hospital (Eindhoven, The Netherlands). All patients were followed up prospectively with serial echocardiography, which was performed preoperatively, at discharge, at 3 to 6 months, at 1 year, and annually thereafter. Details of echocardiographic measurements have been previously described [6]. In summary, echocardiography was performed according to the guidelines of by the American Society of Echocardiography. All echocardiography was performed by the same physicians on the same echocardiographic equipment at the respective centers. The highest peak aortic flow velocity across the Freestyle prosthesis was measured with the continuous wave Doppler technique. Doppler measurements of at least three cardiac cycles were averaged. Aortic valve mean gradient was calculated according to the equation p ⫽ 4(Vao 2– Vlvot2) (p ⫽ pressure gradient [in mm Hg], Vao ⫽ aortic valve velocity, and Vlvot ⫽ velocity over the LVOT). Aortic valve peak gradient was calculated as p ⫽ 4 Vao 2. The effective orifice area of the Freestyle prosthesis was calculated with the continuity equation. Assessment of aortic regurgitation was done with color flow imaging and continuous flow Doppler imaging. Grading of aortic insufficiency was done according to the criteria used by Perry and colleages [7]. Left ventricular mass was calculated with M-mode measurements of wall thickness and left ventricular end-diastolic diameter. The effective orifice area (EOA) of the aortic valve was calculated by the continuity equation, and the mean transvalvular gradient at rest was derived from the simplified Bernoulli equation accounting for the flow velocity across the left ventricular outflow tract. Preoperative patient characteristics are summarized in Table 1. The surgical procedure consisted of a subcoronary, a root inclusion, or a full root technique. In all three techniques a transverse aortotomy was used. In the subcoronary technique the inflow of the porcine root was secured with simple interrupted sutures or with three running sutures in a single plane at the annulus. The porcine valve was trimmed by excision of the remnants of the aortic tissue above the orifices for the coronary sinuses, leaving the noncoronary sinus intact. The out-

Ann Thorac Surg 2001;71:601– 8

Table 1. Preoperative Patient Characteristics Characteristic Age (y) (⫾SD) Male Left ventricular hypertrophy Coronary artery disease Hypertension Myocardial infarction Congestive heart failure Carotid atherosclerotic disease New York Heart Association class III or IV Stroke Diabetes mellitus Renal insufficiency Liver dysfunction Cancer Chronic obstructive pulmonary disease Past cardiac operation Pathological condition Stenosis Regurgitation Stenosis and regurgitation Endocarditis History Active Urgent operation

No.

%

64 (⫾14.5) 158 178 66 52 11 47 8 155

... 56 64 24 19 4 17 3 55

12 12 7 8 6 17 31

4 4 3 3 2 6 11

91 20 169 25 9 16 16

33 7 60 9 3 6 6

flow suture line was performed with a single running suture of 4-0 Prolene (Ethicon, Inc, Somerville, NJ). In the root-inclusion technique the aorta was also opened with a transverse incision. A quadrangular portion of the porcine aortic sinus was excised to remove the porcine coronary artery for both the left and right coronary arteries. Inflow sutures were applied to the aortic annulus so as to maintain the porcine inflow in a single plane. Three running sutures of Prolene were used (one per cusp). Sutures were used to attach the porcine graft to the native aortic wall at the level of the commissures to guarantee the spatial orientation of the valve commissures. A double-arm polypropylene suture was used to attach the edges of the coronary ostia to the native aortic wall. A 4-0 polypropylene suture was used to suture the outflow portion of the graft to the crest of the native aorta. In the total root replacement the aorta was also transected above the sinotubular ridge. Both coronary ostia were mobilized with buttons of aortic wall. The remaining tissue of the sinus of Valsalva was excised. The inflow anastomosis was accomplished using running sutures of 4-0 polypropylene in a single plane. The coronary buttons were sewn to the corresponding sinus of Valsalva with a continuous suture of 5-0 polypropylene. The distal end of the bioprosthesis was sewn end-to-end to the aorta with a continuous suture of 4-0 polypropylene to complete the root replacement. Intraoperative variables are listed in Table 2.

Ann Thorac Surg 2001;71:601– 8

KAPPETEIN ET AL PERFORMANCE OF FREESTYLE PROSTHESIS

Table 2. Surgical Characteristics of Patients Characteristic Concomitant procedures Coronary artery bypass graft Mitral valve repair/replacement Ascending aorta repair Tricuspid valve repair Aortic crossclamp time, all valves (min) (⫾SD) AVR only AVR ⫹ concomitant procedure Valve size (mm) 21 23 25 27 Implantation technique Subcoronary Root inclusion Root

No.

%

Table 3. Logistic Regression Analysis For Hospital Mortality (N ⫽ 27) Risk Factor

43 12 11 5 100 (⫾30)

15 4 4 2 ...

92 (⫾19) 127 (⫾41)

... ...

27 82 76 95

10 29 27 34

171 64 45

61 23 16

Concomitant Procedures A total of 43 patients (15%) required additional coronary artery bypass grafting. Twelve (4%) mitral valve repair or replacements, 11 (4%) ascending aorta repairs, five (2%) tricuspid valve annuloplasties, and one partial lung resection for bullouous emphysema were also performed. Morbid and fatal valve-related events were categorized as structural valve deterioration, nonstructural valve dysfunction, tromboembolism, anticoagulant-related hemorrhage, prosthetic valve endocarditis, reoperation, and valve-related mortality. After stentless valve implantation, permanent anticoagulation with warfarin was prescribed only if additional atrial fibrillation was present. Patients in the conventional group received a 3-month course of warfarin. Because there have been no problems regarding thromboembolic events, this protocol was recently changed. Currently no patient receives systemic anticoagulation therapy.

a

Crossclamp time (continuous) Age (continuous)a Myocardial infarctiona Diabetes mellitusa Left ventricular hypertrophy Coronary artery disease Concomitant cardiac procedure NYHA class III or IV Female gendera Urgent operation Left ventricular dilatation Past cardiac operation Operative technique Subcoronary Root inclusion Root Valve size (mm) 21 23 25 27 a

OR

⫾95% CL

1.02 1.07 6.11 5.33 5.14 3.50 3.17 2.50 2.39 1.37 1.29 1.00

1.01, 1.04 1.02, 1.13 1.66, 22.44 1.49, 19.04 1.51, 17.53 1.55, 7.90 1.40, 7.18 1.02, 6.11 1.05, 5.40 0.29, 6.36 0.52, 3.21 0.28, 3.55

1.00 1.07 1.60

0.40, 2.91 0.58, 4.39

1.00 1.18 0.32 0.32

0.35, 3.96 0.07, 1.40 0.08, 1.29

Significant in multivariate analysis.

CL ⫽ confidence limits; OR ⫽ odds ratio.

NYHA ⫽ New York Heart Association;

myocardial infarction, 2 of uncontrollable bleeding, and 1 of unknown cause. Logistic regression analysis found the following factors predictive for hospital mortality: crossclamp time, age, myocardial infarction, diabetes, left ventricular hypertrophy, coronary artery disease, New York Heart Association (NYHA) class III or IV, and female gender In the multivariate analysis crossclamp

Statistical Analysis Continuous variables are expressed as means ⫾ standard deviations. Repeated-measures analysis of variance was used to detect any significant changes in mean transvalvular gradient across the aortic valve prosthesis, EOA, and left ventricular mass index over time. Logistic regression analysis was used to analyze predictive factors for hospital mortality. Survival was estimated by the Kaplan-Meier method, the variability is indicated by the standard error of the mean. The Cox proportional hazards model was used to examine the impact of covariates on observed late mortality.

Results The 30-day operative mortality was 9.6% (N ⫽ 27). Nine patients died of multiorgan failure, 6 of low cardiac output, 4 of sepsis, 3 after a thromboembolic event, 2 of a

603

Fig 1. Long-term survival after implantation of aortic Freestyle bioprosthesis.

604

KAPPETEIN ET AL PERFORMANCE OF FREESTYLE PROSTHESIS

Ann Thorac Surg 2001;71:601– 8

Table 4. Causes of Late Deaths Cause

No.

Congestive heart failure Reoperation for paravalvular leakage Reoperation for subvalvular fistula to right atrium Sudden death/arrhythmia Miscellaneous noncardiac Myocardial infarction Complications after resection aortic coarctation Gastrointestinal bleeding Unknown Total

Fig 2. Freedom from valve-related complications.

time, age, preoperative myocardial infarction, diabetes, and female gender proved to be independent predictive factors (Table 3). Hospital mortality for the subcoronary, root inclusion, and full root technique was 15 (9%), 6 (9%), and 6 patients (13%), respectively. The overall actuarial survival rates were 94% ⫾ 1% and 84% ⫾ 5% after 1 and 5 years, respectively (Fig 1). The causes of late deaths (determined by autopsies, death certificates, hospital records, and family communications) are listed in Table 4. Of the 20 late deaths, 13 were cardiac related. Freedom from valve-related complications was 96% ⫾ 1% at 1 year and 86% ⫾ 4% at 5 years (Figure 2). There were five (2%) documented thromboembolic events, three of which were fatal. These three occurred in the perioperative period (two strokes and one occlusion of the mesenteric artery). During follow-up there was one stroke and one transient ischemic attack. A total of 5 patients developed a clinically important, anticoagulant-related bleeding event. One patient had an upper gastrointestinal hemorrhage on postoperative day 12, which resolved with conservative treatment. Another patient required laparotomy for a bleeding ulcer 3 days after surgery. One patient had severe hemoptysis 1 day after surgery, which resolved with conservative treatment. One patient had an upper gastrointestinal hemorrhage 3 months postoperatively and required readmission and medical management. The fifth patient (in whom warfarin therapy erroneously was continued after 3 months) developed a subdural hematoma 10 months postoperatively, which required neurosurgical intervention. Estimated freedom from anticoagulant-related hemorrhage at 1 and 5 years was 98% ⫾ 1%. In the Cox proportional hazard analysis, valve-related events (relative risk [RR] 3.9, 95% confidence limits [CL] 1.3 to 11.7), coronary artery disease (RR 3.4, 95% CL 1.4 to 8.2) and concomitant cardiac procedure at time of insertion of the Freestyle prosthesis (RR 3.2, 95% CL 1.3 to 7.8) were prognostic factors (Table 5). In the multivariate

1 1 1 5 2 2 1 1 6 20

analysis only coronary artery disease proved to be prognostic. Reoperation was necessary in 12 patients. Eleven patients had a dehiscence of the proximal suture line at the site of the noncoronary cusp. Running sutures of 4-0 Prolene (Ethicon, Inc, Somerville, NJ) were used in 2 of these patients, and 3 separate running sutures of 3-0 or 4-0 Prolene were used in 9 patients. When dehiscence was diagnosed most patients were symptomatic, presenting with dyspnea (n ⫽ 8), or collapse (n ⫽ 2). Estimated freedom from reoperation for dehiscence was 100% at 1 year, 97% ⫾ 1.4% at 3 years, and 92% ⫾ 3.8% at 5 years. Broken sutures were found in 2 patients at reoperation. Adherence of the prosthetic free wall to the native aortic Table 5. Risk Factors For Late Mortality Analyzed With Cox Proportional Hazards Analysis Risk Factor

RR

⫾95% CL

Valve-related event Coronary artery diseasea Concomitant cardiac procedure History of myocardial infarction Age (continuous) Female gender Operative technique Subcoronary Root inclusion Root Preoperative congestive heart failure Left ventricular hypertrophy Left ventricular dilatation Left ventricular dysfunction Valve size (mm) 21 23 25 27 NYHA class III or IV preoperatively Cancer preoperatively

3.86 3.40 3.19 1.87 1.04 0.65

1.28, 11.67 1.41, 8.21 1.32, 7.75 0.25, 14.03 0.99, 1.09 0.26, 1.63

1.00 0.62 0.68 1.35

0.18, 2.19 0.15, 3.03 0.48, 3.80

3.13 0.89 2.17

0.92, 10.71 0.30, 2.70 0.50, 9.35

1.00 1.12 1.34 1.46 2.49

0.22, 5.79 0.25, 7.17 0.30, 7.06 0.90, 6.86

2.18

0.29, 16.35

a

Significant in multivariate analysis.

Abbreviations as in Table 3.

Ann Thorac Surg 2001;71:601– 8

KAPPETEIN ET AL PERFORMANCE OF FREESTYLE PROSTHESIS

605

Table 6. Changes in Mean Aortic Transvalvular Gradient According to Valve Size Size (mm) 21 23 25 27 Total

Preoperative

Discharge

3– 6 Months

1 Year

2 Years

3 Years

4 Years

62.3 ⫾ 18.7 42.8 ⫾ 13.2 44.0 ⫾ 14.4 48.2 ⫾ 21.6 49.5 ⫾ 17.3

11.4 ⫾ 8.0 10.7 ⫾ 4.2 8.9 ⫾ 4.5 7.0 ⫾ 4.2 8.0 ⫾ 6.5

8.3 ⫾ 4.2 5.8 ⫾ 3.1 4.6 ⫾ 3.2 3.9 ⫾ 2.7 5.8 ⫾ 3.7

7.7 ⫾ 4.3 5.9 ⫾ 3.7 4.3 ⫾ 2.6 4.7 ⫾ 2.4 5.3 ⫾ 3.6

8.2 ⫾ 4.2 5.9 ⫾ 3.1 4.2 ⫾ 2.1 4.5 ⫾ 3.2 5.0 ⫾ 4.5

6.5 ⫾ 3.1 4.8 ⫾ 2.1 3.4 ⫾ 2.2 4.6 ⫾ 3.3 4.8 ⫾ 3.2

6.4 ⫾ 3.1 3.1 ⫾ 2.4 3.9 ⫾ 2.7 5.8 ⫾ 4.0 5.2 ⫾ 3.9

wall was not found in any patient. In 3 patients the valve was replaced with a Freestyle bioprosthesis, in 4 with another device (stented bioprosthesis or mechanical valve), and in 3 patients refixation of the Freestyle valve took place at the site of dehiscence. In 1 patient who died of myocardial infarction, suture dehiscence was found at autopsy. In this patient signs of endocarditis were found and confirmed by cultures. During follow-up another 2 patients had positive blood cultures with Staphylococcus epidermidis. Although no vegetations were seen on echocardiography, these patients were empirically treated on the assumption of prosthetic endocarditis. One patient underwent reoperation for a subvalvular fistula to the right atrium. New-onset AV block was seen in 11 patients postoperatively. In 6 patients, regular conduction was completely restored after a maximum of 10 days. The remaining 5 patients required permanent pacemaker implantation before discharge. Four of these 5 patients had heavily calcified aortic annuli requiring extensive decalcification. At discharge 79% were in stable sinus rhythm. At follow-up all patients were clinically improved, and tolerated more physical activities with little or no dyspnea. The mean aortic transvalvular gradient and EOA according to valve size and for all patients at discharge, at 6 months, and at yearly intervals up to 3 and 4 years are summarized in Tables 6 and 7, respectively. As shown, mean transvalvular gradient decreased significantly after AVR and was reduced further at 3 to 6 months; thereafter the gradient remained relatively stable. There was also a difference in transvalvular gradient among the three implantation techniques for subcoronary, root-inclusion, and full root technique. The mean gradients were 11.2 ⫾ 6.2, 9.5 ⫾ 6.2, and 6.2 ⫾ 4.4 at 4 weeks (p ⫽ 0.007) and were 7.2 ⫾ 4.8, 6.2 ⫾ 5.1, and 4.8 ⫾ 4.1 (p ⫽ 0.02) at 3 to 6 months, remaining relatively stable thereafter. Correspondingly there was a significant increase in EOA in the immediate postoperative period and again at 3 to 6 months. At last follow-up, 91% of patients were in NYHA

class I or II, with a mean of 1.3 (vs 3.23 ⫾ 0.9 preoperatively, p ⫽ 0.001). Left ventricular mass, indexed to body surface area was 161 ⫾ g/m2 1 month after aortic valve replacement and decreased within 3 to 6 months to 141 ⫾ 59 g/m2 and remained constant thereafter. At 1 year of follow-up, trivial regurgitation was found in 6 patients (3%) and mild regurgitation in 4 (2%). Regurgitation did not increase with time during follow-up.

Comment The less obstructive profile of the Freestyle stentless was confirmed by this study by the excellent hemodynamic performance of the prosthesis. The improvement of ventricular function after aortic valve replacement became evident soon after the operation and during follow-up. Most patients were postoperatively in NYHA class I or II. Clinical studies have shown that impaired left ventricular function and incomplete regression of left ventricular hypertrophy are associated with a residual gradient across the aortic valve [8]. The transvalvular gradients of the Freestyle valve are lower than in stented prostheses [9] and LV mass index decreases over time. The improvement in left ventricular function should have a beneficial effect on long-term survival. David and colleagues [10], in a case match study of aortic valve replacement with stentless and stented porcine aortic valves, found that the stentless valve was an independent predictor of late survival. The lower gradients of the valve are probably caused by larger effective orifice areas and the fact that a larger prosthesis can be inserted in a smaller patient [9]. Especially in patients with poor ventricular function, rapid recovery may occur because ventricular function in these patients is particularly sensitive even to mild obstructive gradients [11]. Gradients regressed during the first year after implantation. This phenomenon has been ascribed to expansion and remodeling of the porcine root to fit the native aortic sinuses, together with

Table 7. Changes in Effective Orifice Area According to Valve Size Size (mm)

Discharge

3– 6 Months

1 Year

2 Years

3 Years

21 23 25 27 Total

1.15 ⫾ 0.37 1.48 ⫾ 0.27 1.86 ⫾ 0.49 2.01 ⫾ 0.58 1.54 ⫾ 0.47

1.59 ⫾ 0.36 1.92 ⫾ 0.33 2.06 ⫾ 0.62 2.52 ⫾ 0.63 1.90 ⫾ 0.61

1.50 ⫾ 0.34 1.89 ⫾ 0.31 2.24 ⫾ 0.59 2.77 ⫾ 0.78 2.44 ⫾ 0.70

1.53 ⫾ 0.40 1.90 ⫾ 0.41 2.22 ⫾ 0.49 2.80 ⫾ 0.86 2.24 ⫾ 0.70

1.62 ⫾ 0.55 2.02 ⫾ 0.36 2.27 ⫾ 0.57 2.87 ⫾ 0.73 2.48 ⫾ 0.84

606

KAPPETEIN ET AL PERFORMANCE OF FREESTYLE PROSTHESIS

resolution of both perivalvular hematoma and inflammatory reaction [12]. The stentless valve was designed to mimic the hemodynamic performance of the homograft. The firmness of the tissue and the Dacron (C.R. Bard, Haverhill, PA) reinforcement of the suture lines makes the implantation technique for the Freestyle easier compared with the pliable homograft. Postoperative regurgitation seems to be even less than in other reported series of stentless valves and homografts, and is usually less than the normal closing volume observed in mechanical valves [13–15]. The physiologic concept of stentless valves and the excellent midterm performance might lead to performance of these valves equaling that of aortic homografts in long-term follow-up [16]. The 9.6% operative mortality in this series was relatively high compared to that in other series [1, 15, 17]. Logistic regression analysis showed that crossclamp time, age, history of myocardial infarction, diabetes, and female gender were independent predictive factors for hospital mortality. Patients with a concomitant cardiac procedure had a higher crossclamp time, which was also a significant risk factor in the univariate analysis but was lost in the multivariate analysis. Although implantation of the valve is somewhat easier than that of a homograft, implantation of a stentless valve is technically more demanding and needs longer ischemic times compared with traditional stented valves. Especially in patients who have additional risk factors and need additional procedures such as coronary artery bypass grafting or mitral valve repair, ischemic time may become crucial. Other investigators [15, 17, 18] also reported a longer aortic crossclamp time for stentless prostheses, but could not demonstrate an increased mortality compared to that of stented prostheses. However none of these reports performed a logistic regression analysis to analyze risk factors. The 1- and 5-year actuarial survival rates of 94% and 84% are comparable to the 96% and 80% rates reported by Yun and colleagues [15] for the Freestyle and to the 94% and 87% rates of David and colleagues [18] for the Toronto SPV valve (93% at 5 years). Coronary artery disease was the independent prognostic factor. Age also was not a significant prognostic factor in the univariate analysis. This shows that older patients also benefit from replacement of a diseased aortic valve and restoration of left ventricular function. The Freestyle bioprosthesis is versatile and can be used for the subcoronary implantation, root inclusion, and full root techniques. Transvalvular gradients were somewhat lower with the root inclusion and full root techniques than with the subcoronary technique. More consistent maintenance of the sinotubular junction and commissural geometry of the Freestyle valve may be achieved with the root replacement or the inclusion cylinder techniques and might lead to a longer durability [19 –21]. The full root and root-inclusion technique are technically more demanding and need longer cross-clamp time. When extensive concomitant procedures must be performed, a subcoronary or even a

Ann Thorac Surg 2001;71:601– 8

stented bioprosthesis should be used. In the case of a very small calcified aortic root it seems difficult to apply the root inclusion technique, which therefore should probably be avoided [19]. The high hospital mortality rate of 19.4% for the root technique found by Cartier and colleagues [22] could not be confirmed. Logistic regression analysis failed to demonstrate implantation technique as an independent risk factor. The incidence of valvular dehiscence shows reoperation rates for dehiscence and periprosthetic leakage similar to those for homograft implants [14, 23]. Endocarditis is one of the most important risk factors for valvular dehiscence but was only suspected to be the cause in 1 patient. Dehiscence was not related to the implantation mode or the surgeon. Most dehiscence’s were seen at the site of the noncoronary sinus. We hypothesized that excessive suture traction took place at this site because blood invaded the space between the native aortic wall and the wall of the prosthesis. Together with the use of running sutures, this could have led to the dehiscence at this site. We abandoned the running technique and now use interrupted nonabsorbable sutures for the inflow suture line [24] No dehiscences have been noted since this change in technique. There were 3 early and 2 late neurologic events. Severe stroke occurred in 2 patients with known carotid atherosclerotic disease. One patient who had extensive atherosclerotic disease died postoperatively due to occlusion of the mesenteric artery. The number of anticoagulantrelated hemorrhages was low (n ⫽ 5). Four of these patients were on warfarin. A definitive diagnosis of valve endocarditis could not be documented by echocardiography. In 2 patients positive blood cultures were found, and in 1 patient with valvular dehiscence a positive culture was obtained at autopsy. In summary, implantation of the Freestyle valve is technically more demanding, but the hemodynamics are comparable to those of homografts or autografts. The incidence of thromboembolic events and endocarditis is low. The high incidence of dehiscence at the proximal suture line is probably due to technique. The question of durability compared with conventional stented bioprostheses remains unanswered. Longer follow-up is mandatory to determine the durability of the Freestyle valve and its potential hemodynamic benefits. We thank Heleen Jansen, Lidwien Dijkman, and Lambert Muhlenberg for their contribution in making the database complete.

References 1. Sidiropoulos A, Hotz H, Tschesnow J, Konertz W. Stentless porcine bioprostheses for all types of aortic root pathology. Eur J Cardiothorac Surg 1997;11:917–21. 2. Cohen G, Christakis GT, Buth KJ, et al. Early experience with stentless versus stented valves. Circulation 1997;96(Suppl):II82. 3. He GW, Grunkemeier GL, Gately HL, Furnary AP, Starr A. Up to 30-year survival after aortic valve replacement in the small aortic root. Ann Thorac Surg 1995;59:1056– 62.

Ann Thorac Surg 2001;71:601– 8

4. O’Brien MF, McGiffin DC, Stafford EG. Allograft aortic valve implantation: techniques for all types of aortic valve and root pathology. Ann Thorac Surg 1989;48:600–9. 5. Sabbah HN, Hamid MS, Stein PD. Mechanical stresses on closed cusps of porcine bioprosthetic valves: correlation with sites of calcification. Ann Thorac Surg 1986;42:93– 6. 6. Baur LH, Jin XY, Houdas Y, et al. Echocardiographic parameters of the freestyle stentless bioprosthesis in aortic position: the European experience. J Am Soc Echocardiogr 1999; 12:729–35. 7. Perry GJ, Helmcke F, Nanda NC, Byard C, Soto B. Evaluation of aortic insufficiency by Doppler color flow mapping. J Am Coll Cardiol 1987;9:952–9. 8. Lindblom D, Lindblom U, Qvist J, Lundstrom H. Long-term relative survival rates after heart valve replacement [See comments]. J Am Coll Cardiol 1990;15:566–73. 9. Dumesnil JG, Leblanc MH, Cartier PC, et al. Hemodynamic features of the freestyle aortic bioprosthesis compared with stented bioprosthesis. Ann Thorac Surg 1998;66:S130 –3. 10. David TE, Puschmann R, Ivanov J, et al. Aortic valve replacement with stentless and stented porcine valves: a case-match study. J Thorac Cardiovasc Surg 1998;116:236– 41. 11. Collinson J, Henein M, Flather M, Pepper JR, Gibson DG. Valve replacement for aortic stenosis in patients with poor left ventricular function: comparison of early changes with stented and stentless valves. Circulation 1999;100(Suppl):II1–5. 12. Westaby S, Amarasena N, Long V, et al. Time-related hemodynamic changes after aortic replacement with the freestyle stentless xenograft. Ann Thorac Surg 1995;60:1633– 8. 13. Del Rizzo DF, Goldman BS, Christakis GT, David TE. Hemodynamic benefits of the Toronto stentless valve. J Thorac Cardiovasc Surg 1996;112:1431– 45. 14. Kirklin JK, Smith D, Novick W, et al. Long-term function of cryopreserved aortic homografts. A ten-year study. J Thorac Cardiovasc Surg 1993;106:154– 65.

KAPPETEIN ET AL PERFORMANCE OF FREESTYLE PROSTHESIS

607

15. Yun KL, Sintek CF, Fletcher AD, et al. Aortic valve replacement with the freestyle stentless bioprosthesis: five-year experience. Circulation 1999;100(Suppl 19):II17–23. 16. Gross C, Harringer W, Mair R, et al. Aortic valve replacement: is the stentless xenograft an alternative to the homograft? Early results of a randomized study. Ann Thorac Surg 1995;60:S418 –21. 17. Van Nooten G, Caes F, Francois K, van Belleghem Y, Taeymans Y. Stentless or stented aortic valve implants in elderly patients? Eur J Cardiothorac Surg 1999;15:31– 6. 18. David TE. The Toronto SPV bioprosthesis: clinical and hemodynamic results at 6 years. Ann Thorac Surg 1999; 68(Suppl 3):S9 –13. 19. Huysmans HA, Kappetein AP, Baur LHB. Root inclusion with stentless bioprosthesis. In: Huysmans HA, David TE, Westaby S, eds. Stentless bioprosthesis. Oxford: Isis Medical Media, 1999:121–5. 20. Hazekamp MG, Goffin YA, Huysmans HA. The value of the stentless biovalve prosthesis. An experimental study. Eur J Cardiothorac Surg 1993;7:514–9. 21. Angell WW, Pupello DF, Bessone LN, Hiro SP, Brock JC. Effect of stent mounting on tissue valves for aortic valve replacement. J Card Surg 1991;6(Suppl):595–9. 22. Cartier PC, Dumesnil JG, Metras J, et al. Clinical and hemodynamic performance of the Freestyle aortic root bioprosthesis. Ann Thorac Surg 1999;67:345–9. 23. Jones EL, Shah VB, Shanewise JS, et al. Should the freehand allograft be abandoned as a reliable alternative for aortic valve replacement? Ann Thorac Surg 1995;59:1397– 403. 24. Schoof PH, Baur LH, Kappetein AP, Hazekamp MG, RijkZwikker GL, Huysmans HA. Dehiscence of the Freestyle stentless bioprosthesis. Semin Thorac Cardiovasc Surg 1999; 11(Suppl 1):133– 8.

DISCUSSION DR R. SCOTT MITCHELL (Stanford, CA): Doctor Murray, Dr Matloff, members, and guests. I would like to congratulate Dr Kappetein on an excellent presentation and thank him for forwarding to me the manuscript well in advance. He has presented a very nicely analyzed series of Freestyle stentless valves implanted at three Western European centers. Postoperative gradients were admirably low, and decreased even further over time. The majority of patients returned to functional class 1 or 2, and left ventricular mass regressed toward normal. However, as happens so frequently in cardiac surgery, all these benefits were not achieved without some significant costs. Ischemic times were prolonged over that required for a stented valve, and operative mortality was a rather high 9%. Actuarial 5-year survival was 84%, which is not dissimilar from that achieved by third generation stented valves as reported by Tirone David and others. Whether this purported improved long-term survival with regression of left ventricular mass will indeed prove to be true is still conjectural. I have three somewhat general questions. Given the increased complexity of this procedure, with prolonged ischemic times and some additional measures of technical finesse, for which patients are these additional risks warranted? It seems in fact that the patient in whom we would most like to implant this valve, the elderly female patient with the small root, is perhaps the last person that you would like to implant it in because of the necessity of increased ischemic and bypass times.

Along this same line, what has been your experience with these very young patients? Are the curves for freedom from structural deterioration similar to those in older patients? And finally, you mentioned that two thirds of your late deaths were cardiac related. Was there any evidence that suture lines, especially beneath the right and left coronary ostia, were these suture lines perhaps responsible for any of these late deaths secondary to coronary occlusive disease? I would like to thank the Society for the privilege of discussing this very elegant manuscript. DR JOSEPH E. BAVARIA (Philadelphia, PA): I, too, applaud your wonderful paper; 250 cases is a nice series. Regarding this valve, I am becoming more of a “full root” surgeon as I agree with Francis Robicsek whose motion picture, “AVR, Is It a Half-Operation?” the other night was quite good. I wanted to ask the author a couple of questions. First, was the morbidity and mortality different when comparing the full root subgroup versus the subcoronary subgroup? And second, were there any proximal suture line dehiscences in the full root group? It seems to me that the full root procedure may be a better procedure with this valve. And in my particular experience, I’ve come to the point where I can do a full root just as fast as I can do the subcoronary implant; the cross-clamp times are similar. Thank you very much. The information in this paper was very important.

608

KAPPETEIN ET AL PERFORMANCE OF FREESTYLE PROSTHESIS

DR KAPPETEIN: I thank Dr Mitchell and Dr Bavaria for their questions. The complexity of the procedure is really an issue. What we have shown is that cross-clamp time is longer than in a stented bioprosthesis, and that there are several risk factors that make it more likely that people could die from this operation. In our institution we perform about 90 valve replacements per year, and about 25 of these are stentless valves. The stentless valves are implanted by only three surgeons who also have a large experience in implanting homografts. So thereby we try to combine these patients under three specific surgeons who go through the learning curve one time. The implantation technique used was the root inclusion, the full root, and the subcoronary technique. We did not see any difference. We saw a difference in cross-clamp time for the root inclusion technique: it was much longer, and for the subcoronary technique and the root technique it was shorter. So if you

Ann Thorac Surg 2001;71:601– 8

have to choose a technique in a patient who has additional risk factors, you should prefer the subcoronary technique. Especially in the female patient with a small aortic size, one should weigh the risks of implanting the freestyle valve versus a stented bioprosthesis. In an elderly patient, from about 80 years, you can debate if she will benefit from implanting a stentless valve, because a stentless valve will perform better in the long-term survival and not in the short-term. What about suture lines? We did not see any dehiscence of the suture line in the full root technique. We only saw it in the root inclusion technique and in the subcoronary implantation technique. There was a difference in gradients among the three techniques. The subcoronary technique showed a little bit higher mean transfer of the gradient compared to the root inclusion and the root technique, but there was no difference in morbidity or mortality among these three techniques.