- Email: [email protected]
Quo Vadis Pulmonary Autograft—The Ross Procedure in Its Second Decade: A Single-Center Experience in 645 Patients
Timo Weimar, MD, Efstratios I. Charitos, MD, Markus Liebrich, MD, Detlef Roser, MD, Ioannis Tzanavaros, MD, Nicolas Doll, MD, and Wolfgang B. Hemmer, MD Department of Cardiac Surgery, Sana Cardiac Surgery Stuttgart GmbH, Stuttgart; and Department of Cardiac and Thoracic Vascular Surgery, University of Luebeck, Luebeck, Germany
Background. The enthusiasm about the advantages of a viable autologous transplant faded with recent reports of autograft deterioration and associated reoperations after the Ross procedure. This report evaluates predictors for autograft failure and outcomes extending into the second decade after a Ross procedure. Methods. From 1995 through 2012, 645 consecutive patients (mean age, 42.3 ± 14.2 years; 76% males) underwent a Ross operation using mainly the root replacement technique (98%). They were prospectively followed up with clinical and echocardiographic evaluations. Total follow-up was 5,349 patient-years and was 96% complete. Mean follow-up duration was 8.4 ± 4.6 years (range, 0 to 17.4 years). Results. Early mortality was 0.9% (n [ 6). Cumulative survival at 15 years was 92.7% (95% confidence interval, 90.1% to 95.3%) and did not differ from the general German population (p [ 0.261). Freedom from reoperation on the autograft or the pulmonary allograft at 12 years
was 91.6% (95% confidence interval, 88.5% to 94.9%) and 95.0% (95% confidence interval, 92.8% to 97.2%), respectively. Sixty-five patients (10.1%) required a total of 78 valve-related reoperations after a Ross procedure (1.5%/ patient-year) with a reoperative hospital mortality of 3.8% (n [ 3). Forty-seven autograft reoperations were observed in 44 patients (0.87%/patient-year); 22 of them (47%) could be performed as a valve-sparing procedure. Preoperative aortic valve regurgitation and an aortic annulus of at least 26 mm were identified as predictors for autograft failure. Conclusions. In this large series, the Ross procedure resulted in excellent long-term survival rates with a low risk of valve-related morbidity and a considerably low rate of reoperations in young and middle-aged patients and should be considered as an important treatment option in this cohort.
W
considerable reoperation rates after the full root technique, challenging the raison d’etre of the Ross procedure in adults [9, 11]. However, middle-age patients requiring aortic valve replacement face a treatment dilemma: although biologic prostheses still show unsatisfactory durability, mechanical prostheses drastically change a patient’s lifestyle and thromboembolic and bleeding events are not rare [12–14]. We present one of the largest single-center Ross experiences with a follow-up extending well into the second decade, which has been assumed to be the critical mark for autograft stability, thus providing critical information for further judgment of this surgical concept.
Accepted for publication July 23, 2013. Address correspondence to Dr Weimar, Sana Cardiac Surgery Stuttgart GmbH, Herdweg 2, 70174 Stuttgart, Germany; e-mail: timo.weimar@ yahoo.de.
Ó 2014 by The Society of Thoracic Surgeons Published by Elsevier Inc
ˇ
ith its introduction 40 years ago, the Ross procedure was expected to eliminate most of the disadvantages associated with conventional aortic valve replacement [1]. Its excellent hemodynamic performance, unique reproduction of the complex function of a healthy valve creating normal transvalvular flow dynamics, avoidance of anticoagulation therapy, and allowance of unrestrained activity as well as its growth potential made the pulmonary autograft a promising treatment option for both congenital and acquired disease of the left ventricular (LV) outflow tract [2–5]. All this comes at the cost of a double-valve procedure that is technically challenging. When midterm results emerged, initial enthusiasm was followed by the recognition of a possible deterioration of the autograft with time [6, 7]. Dilatation of the neoaortic root was found to be the principal cause of autograft failure [6, 8–10], and some recent publications reported
(Ann Thorac Surg 2014;97:167–74) Ó 2014 by The Society of Thoracic Surgeons
Patients and Methods The study was approved by the institutional ethics committee, and informed consent was obtained from all patients. From February 1995 through May 2012, a Ross procedure was performed on 645 consecutive patients at Sana Cardiac Surgery Stuttgart. This reflected 6.1% of a total of 10,574 isolated aortic valve procedures performed during 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2013.07.078
ADULT CARDIAC
Quo Vadis Pulmonary Autograft—The Ross Procedure in Its Second Decade: A Single-Center Experience in 645 Patients
168 ADULT CARDIAC
WEIMAR ET AL THE ROSS PROCEDURE IN 645 PATIENTS
Ann Thorac Surg 2014;97:167–74
the same period, including 8,933 prosthetic valve replacements and 996 aortic valve reconstructions. Patients’ characteristics are shown in Table 1. Indications included severe aortic valve or root disease with the patient’s preference to avoid oral anticoagulation, contraindication for oral anticoagulation, childbearing potential in women, and an active lifestyle. Contraindications included significantly reduced LV function, more then two-vessel coronary artery disease, connective tissue or active rheumatic disorders, anomalies of the pulmonary valve, and a shortened life expectancy for other reasons.
Surgical Technique Perioperative data are shown in Table 2. The majority of patients received a freestanding aortic root replacement without (RR; n ¼ 224 [35%]; 1995–1998) or with (RRþR; n ¼ 406 [63%]; 1999–2012) autograft reinforcement. Fifteen patients (2%) were operated on with the subcoronary
Table 2. Perioperative Variables Characteristics Operative technique Root replacement Subcoronary implantation Concomitant procedures Ascending aorta replacement Ascending aorta remodeling CABG Other valve surgery Mean CPB time (min [ SD]) Mean cross-clamp time (min [ SD]) Circulatory arrest for arch surgery Autograft reinforcement Annulus Sinotubular junction Both CABG ¼ coronary artery bypass grafting; bypass; SD ¼ standard deviation.
Value 97.7% (n ¼ 630) 2.3% (n ¼ 15) 51.8% (n ¼ 334) 32.4% (n ¼ 209) 9.1% (n ¼ 59) 8.7% (n ¼ 56) 1.6% (n ¼ 10) 165 35 133 24 4.3% (n ¼ 28) 62.9% (n ¼ 406) 33.1% (n ¼ 213) 25.4% (n ¼ 164) CPB ¼ cardiopulmonary
Table 1. Patient Characteristics Characteristics
Value
Male sex Mean age (y [ SD, range]) <20 20–40 41–60 >60 Prior cardiac surgery Aortic valve repair Aortic valve replacement Other Predominant aortic hemodynamics Stenosis Insufficiency Mixed lesion Aortic valve morphology Bicuspid Prosthetic dysfunction Active endocarditis NYHA functional class I II III IV Unknown Systolic left ventricular function Good (EF >0.50) Impaired (EF 0.40–0.50) Moderate/bad (EF <0.40) Unknown Type of operation Elective Urgent EF ¼ ejection fraction; standard deviation.
75.5% (n ¼ 487) 42.3 14.2 (1–68) 10.9% (n ¼ 70) 26.7% (n ¼ 172) 58.4% (n ¼ 377) 4.0% (n ¼ 26) 14.1% (n ¼ 91) 5.3% (n ¼ 34) 6.4% (n ¼ 41) 2.5% (n ¼ 16) 31.5% (n ¼ 203) 28.8% (n ¼ 186) 33.3% (n ¼ 215) 57.8% (n ¼ 373) 6.4% (n ¼ 41) 6.0% (n ¼ 39) 31.9% 42.5% 15.7% 4.5% 5.4%
(n (n (n (n (n
¼ ¼ ¼ ¼ ¼
206) 274) 101) 29) 35)
74.9% 8.5% 0.2% 16.4%
(n (n (n (n
¼ ¼ ¼ ¼
483) 55) 1) 106)
98.1% (n ¼ 633) 1.9% (n ¼ 12)
NYHA ¼ New York Heart Association;
SD ¼
implantation technique and were excluded from analyses of freedom from autograft reoperation. Both techniques have been previously described in detail [15–17]. In the RRþR group, a polyethylene terephthalate fiber (Dacron) strip was incorporated in the proximal suture line, and a second suture line fixed the aortic wall remnant to the autograft to support the annulus (Fig 1). In most cases, the right ventricular outflow tract (RVOT) was reconstructed with a cryopreserved homograft (pulmonary, n ¼ 584 [91%]; aortic, n ¼ 8 [1%]), although 53 patients (8%) received a porcine stentless xenograft (Freestyle; Medtronic, Inc, Minneapolis, MN).
Data Acquisition All patients were prospectively followed up with annual visits obtaining clinical details and serial standardized echocardiography. All postoperative events as well as structural and nonstructural valve deterioration were defined according to the Guidelines for Reporting Morbidity and Mortality After Cardiac Valvular Interventions [18]. The database was frozen for analyses in May 2012. The total follow-up was 5,349 patient-years and was 96% complete, with a mean follow-up duration of 8.4 4.6 years (range, 0 to 17.4 years).
Statistical Analysis Continuous data were reported as mean standard deviation; categorical data were expressed as absolute numbers or proportions. The survival was compared with the age- and sex-matched general German population (www.destatis.de; Statistisches Bundesamt, Wiesbaden, Germany). Relative survival was expressed as the ratio of expected versus observed numbers of death [19]. The actuarial estimates of freedom from reoperations were calculated using the Kaplan-Meier method. The risk of reoperation is presented as the smoothed instantaneous rate of a patient requiring reoperation within an interval
WEIMAR ET AL THE ROSS PROCEDURE IN 645 PATIENTS
169
Fig 1. Annulus reinforcement (autograft invaginated into the left ventricular outflow tract for implantation). (A, B) First suture line underneath the cusps incorporating a Dacron strip. (C) Second suture line to fix the remnant aortic wall to the autograft. (D) Schematic cross section.
(t, t þ dt), provided the patient was not censored until the beginning of t [20]. To identify predictors for shorter time to reoperation, we performed univariate analyses using the Cox proportional-hazard regression method on 15 preoperative and perioperative variables. Multivariate Cox proportional-hazard models were used to confirm whether significant univariate predictors (p < 0.100) persisted in the presence of other variables. For longitudinal analyses of echocardiographic measurements, we used a random-effects model, assuming correlation between repeated follow-up measurements and a random patient effect [21]. Valve regurgitation was modeled using multinomial ordinal models, with a random patient effect and a piecewise linear relationship with follow-up time with knots at {1, 3, 5, ., 17} years. Statistical analyses were performed using R version 2.15.2 (Development Core Team, R Foundation for Statistical Computing, Vienna, Austria).
Results Patients’ Survival There were no intraoperative deaths. The 30-day mortality was 0.9% (6 of 645 patients; Table 3), caused by multiorgan failure, myocardial infarction, right ventricular failure, hepatorenal syndrome, sudden death, and arrhythmia (n ¼ 1, respectively). There were 27 late deaths. In 15 of those a cardiac cause could be demonstrated, resulting in a linearized occurrence rate (LOR) of 0.5%/patient-year. The overall cumulative survival at 15 years was 92.7% (95% confidence interval [CI], 90.1% to 95.3%) and did not differ from the general German population (Fig 2A). The subcohort of those younger than 50 years of age revealed 13 late deaths (expected, 8.3; p ¼ 0.100), whereas there were 14 late deaths (expected, 18.9; p ¼ 0.260) in patients 50 years or older.
Reoperations Sixty-five patients (10.1%) underwent a total of 78 valverelated reoperations (LOR, 1.5%/patient-year). Reoperative hospital mortality was 3.8% and was caused by endocarditis with multiorgan failure in all cases (n ¼ 3). Freedom from reoperation at 12 years was 88.6% (95% CI, 85.1% to 92.2%; Fig 2B). Six patients underwent combined
autograft and homograft reoperations and are presented in the respective valve reinterventions. PULMONARY AUTOGRAFT. Forty-four patients (6.8%) required 47 reinterventions on the autograft, of which 22 (47%) were performed as a valve-sparing procedure. In 11 patients, autograft dysfunction occurred within 1 year and was not related to autograft dilatation but caused by endocarditis (n ¼ 8), rupture of the proximal suture line (n ¼ 1), covered rupture of the proximal autograft (n ¼ 1), and development of aortic root pseudoaneurysm (n ¼ 1). Thirty-four patients underwent 36 reinterventions for late (1 year) autograft failure after a mean interval of 7.5 4.4 years (median, 7.0 years; range, 1.2 to 14.9 years). The cause for autograft dysfunction was endocarditis in 4, structural valve disease in 11, and nonstructural valve disease in 21 reinterventions. Three patients received simultaneous reoperation on the autograft and the pulmonary allograft. In 18 of 36 late reinterventions (50%), the autograft could be preserved (David procedure in most cases). Freedom from reoperation on the autograft at 12 years was 91.6% (95% CI, 88.5% to 94.9%) (Fig 2C), with
Table 3. Clinical Course Characteristics Early (30 days) Mortality Reoperation for bleeding/ tamponade Myocardial infarction Deep sternal wound infection Cerebrovascular accidentsa Pacemaker implantation Procedure-related CABG Late (>30 d) Mortality Thromboembolic eventsa TIA Hemorrhagic events Endocarditis a
Value 0.9% (n ¼ 6) 4.9% (n ¼ 32) 0.15% 0.46% 1.4% 0.93% 1.4%
(n (n (n (n (n
¼ ¼ ¼ ¼ ¼
1) 3) 9) 6) 9)
4.2% 1.9% 0.93% 0.46% 2.6%
(n (n (n (n (n
¼ ¼ ¼ ¼ ¼
27) 12), 0.22%/py 6), 0.11%/py 3), 0.06%/py 17), 0.32%/py
Including transient ischemic attack (TIA).
CABG ¼ coronary artery bypass graft;
py ¼ patient-year.
ADULT CARDIAC
Ann Thorac Surg 2014;97:167–74
170 ADULT CARDIAC
WEIMAR ET AL THE ROSS PROCEDURE IN 645 PATIENTS
Ann Thorac Surg 2014;97:167–74
Fig 2. (A) Survival of Ross patients versus general German population, sex- and age-matched. (B) Freedom from autograft (AG) or pulmonary allograft (HG) reoperation. (C) Freedom from autograft reoperation. (D) Freedom from pulmonary allograft reoperation.
Table 4. Freedom From Reoperation on the Pulmonary Autografta Characteristics Aortic valve lesion Stenosis Insufficiency Mixed Preoperative aortic morphology Nonbicuspid Bicuspid Aortic annulus diameter 26 mm >26 mm Sex Female Male a
5 Years
10 Years
96.8 (94.3–99.4) 94.3 (91.0–97.8) 97.6 (95.6–99.7)
96.8 (94.3–99.4) 89.8 (85.1–94.8) 96.9 (94.5–99.4)
p Value 0.006
0.555 96.2 (93.7–98.7) 96.3 (94.2–98.4)
94.3 (91.1–97.6) 94.2 (91.3–97.1) 0.002
99.2 (98.1–100) 93.9 (91.2–96.7)
97.8 (95.7–100) 91.2 (87.7–94.8)
98.4 (96.3–100) 95.5 (93.6–97.5)
98.4 (96.3–100) 93.0 (90.3–95.7)
0.129
Values show percentages (95% confidence intervals). Log rank probability values are reported.
no difference between the RR and the RRþR groups (p ¼ 0.781). Freedom from autograft reoperation depending on anatomic morphology is shown in Table 4. Univariate analysis identified pure aortic insufficiency (hazard ratio [HR], 2.8; 95% CI, 1.4 to 5.3; p < 0.001) and annulus diameter (HR, 1.1/mm; 95% CI, 1.0 to 1.2; p ¼ 0.040) as an independent predictor of reoperation. Multivariate analysis revealed an annulus diameter greater than 26 mm as a significant predictor of shorter time to autograft reoperation (HR, 2.8; 95% CI, 1.1 to 6.8; p ¼ 0.028), which persisted in the presence of pure aortic insufficiency. RIGHT VENTRICULAR OUTFLOW TRACT ALLOGRAFT. Twenty-nine patients (4.5%) required 37 reinterventions on the RVOT after a mean interval of 4.8 3.5 years (median, 3.7 years; range, 0.1 to 12.9 years). Stenosis was the main cause for replacement (n ¼ 30). Homograft endocarditis was the cause for reoperation in 3 patients; 2 of them had double-valve endocarditis. From 53 implanted xenografts (mean follow-up, 4.3 0.8 years), 7 (13%) had to be replaced mainly as a result of excessive tissue ingrowth at the proximal anastomosis
WEIMAR ET AL THE ROSS PROCEDURE IN 645 PATIENTS
171
Fig 3. (A, B) Postoperative probability of a patient being in each valve insufficiency grade. (C) Average pressure gradients across the right ventricular outflow tract (RVOT). (D) Longitudinal changes in sinotubular junction (STJ) and aortic annulus diameters. (CI ¼ confidence interval.)
causing stenosis. Freedom from reoperation on the RVOT (including xenografts) at 12 years was 95.0% (95% CI, 92.8% to 97.2%; Fig 2D). Cox regression analysis identified stentless xenograft (HR, 14.3; 95% CI, 4.5
to 33.3; p < 0.001), younger donor age (HR, 1.04/year; 95% CI, 1.01 to 1.02; p < 0.001), and younger patient age (HR, 1.06/year; 95% CI, 1.08 to 1.10; p < 0.001) as independent predictors of reoperation.
Fig 4. (A) Body surface area (BSA) indexed left ventricular (LV) mass preoperative and at latest follow-up (FU); p < 0.001. (B) Longitudinal changes in left ventricular mass. Dotted lines denote 160 g/m2 (mean þ one standard deviation of a normotensive control population) [22].
ADULT CARDIAC
Ann Thorac Surg 2014;97:167–74
172 ADULT CARDIAC
WEIMAR ET AL THE ROSS PROCEDURE IN 645 PATIENTS
Ann Thorac Surg 2014;97:167–74
Echocardiographic and Functional Status
reoperation at 12 years was 91.0% (95% CI, 87.3% to 94.8%), and freedom from aortic insufficiency of at least grade II was 93.4% (95% CI, 90.1% to 96.7%).
After excluding patients who had died, had reoperated valves, or had not completed 1 year of follow-up by the time of the database freeze, echocardiographic follow-up was available in 545 patients. Freedom from aortic insufficiency of at least grade II at 12 years was 91.4% (95% CI, 88.2% to 94.7%). The longitudinal probability of valve insufficiency occurrence as well as the development of RVOT gradients is shown in Figure 3. At last follow-up, mean and maximum pressure gradients across the autograft were 3.3 1.4 mm Hg (range, 1 to 11 mm Hg) and 6.3 2.7 mm Hg (range, 2 to 20 mm Hg), respectively, with 95% of patients presenting with aortic insufficiency of at least grade I. The mean and maximum pressure gradients across the RVOT were 9.4 6.7 mm Hg (range, 0 to 71 mm Hg) and 17.2 10.9 mm Hg (range, 0 to 122 mm Hg), respectively, with 90% showing a regurgitation of at least grade I. Longitudinal analyses revealed a small albeit statistically significant increase in sinotubular junction (STJ) diameters and annulus diameters (3.0 mm/decade and 0.1 mm/decade, respectively; p < 0.001) with no clinical significance or implication (Fig 3D). No effect of the applied technique (RR versus RRþR) on autograft diameters was observed (p > 0.2 for all comparisons). There was a significant LV mass regression (mean, 15.4 1.2 g/m2 per year; p < 0.001), which was mostly complete within 2 years after surgery (Fig 4). At latest follow-up, 76.1% of patients were within one standard deviation of the LV mass of normotensive individuals [22], and 95% of patients presented in New York Heart Association functional class I.
Endocarditis The LOR of endocarditis requiring medical or surgical therapy was 0.1%/patient-year and 0.2%/patient-year for the autograft and RVOT allograft, respectively. Most patients with autograft endocarditis (12 of 13 patients) required reoperation (mean interval, 3.5 5.25 years), whereas only 50% of patients with homograft endocarditis (3 of 6 patients) were in need of reoperation (mean interval, 0.15 0.10 years). None of the patients who had endocarditis as an indication for the initial Ross operation exhibited recurrent late endocarditis.
Thromboembolic or Hemorrhagic Events The LOR for thromboembolic (including transient ischemic attacks) or major bleeding events was 0.24%/ patient-year. The permanent stroke rate was 0.07%/ patient-year.
Follow-Up Exceeding 10 Years A total of 269 patients (41%) had been operated on before January 01, 2002, thus potentially having a follow-up of more than 10 years. Patients operated on before January 01, 2002, who experienced reoperation or were deceased before reaching a 10-year follow-up were also included. Mean follow-up was 12.6 2.6 years (range, 0.5 to 17.3 years; total of 3,377 patient-years). Freedom from any
Comment When reviewing reports of mortality and morbidity associated with the Ross procedure, it becomes conspicuous that results are quite variable. In a meta-analysis, Tackenberg and colleagues [11] published mortality rates ranging from 0.6% to 6.8%. Although many centers have reported excellent results and low reoperation rates irrespective of the applied technique, others revealed an unacceptable occurrence of autograft deterioration as well as mortality and morbidity [4, 6–8, 16, 17, 23, 24]. This dissension seems to reflect the complexity of this procedure. However, there is wide consensus that the pulmonary autograft is unique and advantageous in its ability to change physiologically in shape and size during the cardiac cycle and to respond extremely well to various hemodynamic stimuli [25]. This series demonstrated an excellent survival as late as 15 years after the Ross procedure, equal to that of the general population. This is remarkable considering the presence of severe valvular disease with related structural alterations in the myocardium and the connective tissue. Young adults receiving other types of aortic valve replacement, whether biologic or mechanical, have especially shown a substantially lower life expectancy than the general population and seem to have an inferior outcome with the Ross procedure with regard to survival [26]. Patient selection may have had a positive bias. Nevertheless, about 25% of patients presented with reduced LV function and about 10% received concomitant procedures. Moreover, our results revealed excellent postoperative hemodynamics that led to LV mass regression and a low occurrence rate of neurologic events, all of which may have a favorable impact on patients’ survival [2–4, 16, 22]. The early mortality in our series was 0.9%. This is acceptable considering the fact that this included 49 reoperations and 31 cases of active endocarditis reflecting subgroups with a higher risk. The risk of reoperation after the Ross procedure remains a matter of concern and is critically discussed. Progressive root dilatation is one of the major fears and has been reported especially when using the root replacement technique [6, 7, 9]. It is unclear whether a predominant dilated aortic annulus would be a marker of connective tissue disorder that would also affect the autograft, eventually predicting autograft failure [23]. This study confirmed preoperative aortic valve insufficiency being an independent risk factor for reoperation on the autograft. However, patients presenting with an aortic insufficiency as a result of annulus dilatation would likely receive a David operation at present time and would not be considered primary candidates for a Ross procedure. Taking into account that the pulmonary wall, although structurally identical to the aortic wall at the time of birth, changes within the first postnatal months and shows progressive fragmentation of its elastic fibers with a
thinning of the media portion, we consider a systolic blood pressure control of 120 mm Hg or lower during the first 3 months as critical [7, 27]. This allows the autograft to adapt to increased and distending pressure by progressive deposition of collagenous tissue in the adventitia [4]. In this study, freedom from reoperation on the neoaortic root was 92% at 12 years. Furthermore, freedom from any reoperation in patients beyond a follow-up time of 10 years was 91%. Longitudinal analysis identified no systematic increase of autograft insufficiency grade with time. This has been reported for the subcoronary technique but contradicts what has been assumed for the full root technique [6, 7, 9, 28]. We could not observe a significant difference in need for reoperation in patients with or without annulus reinforcement such as that reported by others [7, 28]. This might be owing to our RR technique of keeping the autograft as short as possible and preserving its adventitia. Nevertheless, because of the knowledge we gained from the German Dutch Ross Registry, we do recommend, and continue to apply, annulus reinforcement [3, 7]. Furthermore, we believe that the importance of stabilizing the STJ might have been underestimated in the past. Luciani and associates [6] emphasized the importance of the STJ on root geometry by showing that diameters of the sinotubular ridge and proximal ascending aorta tend to equalize with the sinus of Valsalva, thereby realizing a distally pronounced root remodeling. Therefore, we lowered the critical threshold for replacement of the ascending aorta in case of ectasia or aorta ascendens–STJ mismatch to stabilize the STJ. Early autograft failure occurred in 1.7%. This was mostly related to technical errors or endocarditis and was not a result of the disease seen in late dilatation. However, Ross-related reoperations are risk-carrying procedures, and their complexity should be taken into account when considering this technique. In a recent report, Stulak and coworkers [24] revealed a broad spectrum of reoperations that might be required. Interestingly, those complex reoperations were performed with a remarkable low mortality of 1.8%. However, in our series reoperative mortality was higher at 3.8%, as all patients presented in critical state owing to active endocarditis and required urgent or emergency surgery. No mortality was observed in patients scheduled for elective reoperation. Deterioration of the RVOT conduit as a function of time was partly expected and emphasizes the problem that a single-valve disease is treated with a double-valve replacement. However, freedom from RVOT reoperation was 95% at 12 years. Homograft replacement was mainly the result of stenosis and could be performed with no mortality. The homograft cusps were generally less affected by fibroses than the wall was. Histologic findings were nonspecific, and neither rejection nor infection was seen in these cases. Carr-White and colleagues [29] rejected a correlation between the development of HLA antibodies and homograft stenosis, so the impact of allogeneic immune response as a cause of homograft stenosis remains uncertain. In 2007 and 2008, we
WEIMAR ET AL THE ROSS PROCEDURE IN 645 PATIENTS
173
implanted 53 stentless xenografts because homograft supply became temporarily limited. Because of a high incidence of unacceptable pressure gradients mostly related to tissue ingrowth we abandoned this alternative. Histologic findings were nonspecific, and a technical error seemed unlikely because gradients were normal at discharge. Endocarditis was a significant cause for reoperation and counted for almost 30% of all reinterventions in this series. However, none of the patients who had endocarditis as an indication for the initial Ross operation exhibited recurrent late endocarditis. Awareness that autologous and allograft material does not provide perfect protection from endocarditis and adoption of a more aggressive prophylaxis might help to significantly reduce the need of reoperation in the future. Although we suggested in previous reports that a Ross procedure might be superior in preventing recurrent endocarditis compared with other types of aortic valve replacement, our current findings reject this assumption, and we do not perform this procedure in patients with active endocarditis anymore [15]. Our study population with a mean age of 42 years represents a treatment dilemma. The few reports on bioprostheses in young and middle-age patients demonstrated a high occurrence of early structural valve deterioration with a striking risk of early reoperation [13, 14]. This was also true for new-generation bioprostheses, and the authors concluded that their implantation in middle-age patients has to be carefully considered. The risk of mechanical valve-related complications is not rare either and may drastically change a patient’s lifestyle [12, 13, 26]. Our study revealed a low morbidity associated with the Ross procedure. Compromising thromboembolic or bleeding events were infrequent with an LOR of 0.24%/patient-year. In contrast, the reported freedom from thromboembolism for mechanical prostheses after 10 years ranges from 67% to 91% [30]. However, it seems impossible to objectively balance the risk of reoperation after a Ross procedure against the fear of thromboembolism or bleeding events after implantation of a mechanical valve.
References 1. Ross DN. Replacement of aortic and mitral valves with a pulmonary autograft. Lancet 1967;2:956–8. 2. Elkins RC, Thompson DM, Lane MM, Elkins CC, Peyton MD. Ross operation: 16-year experience. J Thorac Cardiovasc Surg 2008;136:623–30, 630.e1–5. 3. Hanke T, Stierle U, Boehm JO, et al; German Ross Registry. Autograft regurgitation and aortic root dimensions after the Ross procedure: the German Ross Registry experience. Circulation 2007;116(11 Suppl):I251–8. 4. Yacoub MH, Klieverik LM, Melina G, et al. An evaluation of the Ross operation in adults. J Heart Valve Dis 2006;15:531–9. 5. Carr-White GS, Afoke A, Birks EJ, et al. Aortic root characteristics of human pulmonary autografts. Circulation 2000;102(19 Suppl 3):III15–21. 6. Luciani GB, Casali G, Favaro A, et al. Fate of the aortic root late after Ross operation. Circulation 2003;108(Suppl 1):II61–7. 7. Klieverik LM, Takkenberg JJ, Bekkers JA, RoosHesselink JW, Witsenburg M, Bogers AJ. The Ross operation: a Trojan horse? Eur Heart J 2007;28:1993–2000.
ADULT CARDIAC
Ann Thorac Surg 2014;97:167–74
174 ADULT CARDIAC
WEIMAR ET AL THE ROSS PROCEDURE IN 645 PATIENTS
8. David TE, Omran A, Ivanov J, et al. Dilation of the pulmonary autograft after the Ross procedure. J Thorac Cardiovasc Surg 2000;119:210–20. 9. Kouchoukos NT, Masetti P, Nickerson NJ, Castner CF, Shannon WD, Davila-Rom an VG. The Ross procedure: longterm clinical and echocardiographic follow-up. Ann Thorac Surg 2004;78:773–81. 10. Takkenberg JJ, van Herwerden LA, Galema TW, et al. Serial echocardiographic assessment of neo-aortic regurgitation and root dimensions after the modified Ross procedure. J Heart Valve Dis 2006;15:100–7. 11. Takkenberg JJ, Klieverik LM, Schoof PH, et al. The Ross procedure: a systematic review and meta-analysis. Circulation 2009;119:222–8. 12. Puvimanasinghe JP, Takkenberg JJ, Edwards MB, et al. Comparison of outcomes after aortic valve replacement with a mechanical valve or a bioprosthesis using microsimulation. Heart 2004;90:1172–8. 13. Chan V, Malas T, Lapierre H, et al. Reoperation of left heart valve bioprostheses according to age at implantation. Circulation 2011;124(11 Suppl):S75–80. 14. Tanaka H, Okita Y, Kasegawa H, et al. The fate of bioprostheses in middle-aged patients: The Japanese experience. J Heart Valve Dis 2010;19:561–7. 15. B€ ohm JO, Botha CA, Rein JG, Roser D. Technical evolution of the Ross operation: midterm results in 186 patients. Ann Thorac Surg 2001;71(5 Suppl):S340–3. 16. B€ ohm JO, Hemmer W, Rein JG, et al. A single-institution experience with the Ross operation over 11 years. Ann Thorac Surg 2009;87:514–20. 17. Sievers HH, Hanke T, Stierle U, et al. A critical reappraisal of the Ross operation: renaissance of the subcoronary implantation technique? Circulation 2006;114(1 Suppl):I504–11. 18. Akins CW, Miller DC, Turina MI, et al. Guidelines for reporting mortality and morbidity after cardiac valve interventions. Ann Thorac Surg 2008;85:1490–5. 19. Pohar M, Stare J. Relative survival analysis in R. Comput Methods Programs Biomed 2006;81:272–8. 20. Charitos EI, Stierle U, Hanke T, Schmidtke C, Sievers HH, Richardt D. Long-term results of 203 young and middle-aged
Ann Thorac Surg 2014;97:167–74
21.
22. 23. 24. 25. 26.
27. 28.
29.
30.
patients with more than 10 years of follow-up after the original subcoronary Ross operation. Ann Thorac Surg 2012;93:495–502. Andrinopoulou ER, Rizopoulos D, Jin R, Bogers AJ, Lesaffre E, Takkenberg JJ. An introduction to mixed models and joint modeling: analysis of valve function over time. Ann Thorac Surg 2012;93: 1765–72. Duebener LF, Stierle U, Erasmi A, et al; German Ross Registry. Ross procedure and left ventricular mass regression. Circulation 2005;112(9 Suppl):I415–22. David TE, Woo A, Armstrong S, Maganti M. When is the Ross operation a good option to treat aortic valve disease? J Thorac Cardiovasc Surg 2010;139:68–75. Stulak JM, Burkhart HM, Sundt TM 3rd, et al. Spectrum and outcome of reoperations after the Ross procedure. Circulation 2010;122:1153–8. Yacoub MH, Cohn LH. Novel approaches to cardiac valve repair: from structure to function: part I. Circulation 2004;109: 942–50. van Geldorp MW, Eric Jamieson WR, Kappetein AP, et al. Patient outcome after aortic valve replacement with a mechanical or biological prosthesis: weighing lifetime anticoagulant-related event risk against reoperation risk. J Thorac Cardiovasc Surg 2009;137:881–6, 886e1–5. Wagenvoort CA, Neufeld H, Birge RF, Caffrey JA, Edwards JE. Origin of right pulmonary artery from ascending aorta. Circulation 1961;23:84–90. Charitos EI, Takkenberg JJ, Hanke T, et al. Reoperations on the pulmonary autograft and pulmonary homograft after the Ross procedure: an update on the German Dutch Ross Registry. J Thorac Cardiovasc Surg 2012;144: 813–23. Carr-White GS, Kilner PJ, Hon JK, et al. Incidence, location, pathology, and significance of pulmonary homograft stenosis after the Ross operation. Circulation 2001;104(12 Suppl 1):I16–20. Akins CW. Results with mechanical cardiac valvular prostheses. Ann Thorac Surg 1995;60:1836–44.
Background. The enthusiasm about the advantages of a viable autologous transplant faded with recent reports of autograft deterioration and associated reoperations after the Ross procedure. This report evaluates predictors for autograft failure and outcomes extending into the second decade after a Ross procedure. Methods. From 1995 through 2012, 645 consecutive patients (mean age, 42.3 ± 14.2 years; 76% males) underwent a Ross operation using mainly the root replacement technique (98%). They were prospectively followed up with clinical and echocardiographic evaluations. Total follow-up was 5,349 patient-years and was 96% complete. Mean follow-up duration was 8.4 ± 4.6 years (range, 0 to 17.4 years). Results. Early mortality was 0.9% (n [ 6). Cumulative survival at 15 years was 92.7% (95% confidence interval, 90.1% to 95.3%) and did not differ from the general German population (p [ 0.261). Freedom from reoperation on the autograft or the pulmonary allograft at 12 years
was 91.6% (95% confidence interval, 88.5% to 94.9%) and 95.0% (95% confidence interval, 92.8% to 97.2%), respectively. Sixty-five patients (10.1%) required a total of 78 valve-related reoperations after a Ross procedure (1.5%/ patient-year) with a reoperative hospital mortality of 3.8% (n [ 3). Forty-seven autograft reoperations were observed in 44 patients (0.87%/patient-year); 22 of them (47%) could be performed as a valve-sparing procedure. Preoperative aortic valve regurgitation and an aortic annulus of at least 26 mm were identified as predictors for autograft failure. Conclusions. In this large series, the Ross procedure resulted in excellent long-term survival rates with a low risk of valve-related morbidity and a considerably low rate of reoperations in young and middle-aged patients and should be considered as an important treatment option in this cohort.
W
considerable reoperation rates after the full root technique, challenging the raison d’etre of the Ross procedure in adults [9, 11]. However, middle-age patients requiring aortic valve replacement face a treatment dilemma: although biologic prostheses still show unsatisfactory durability, mechanical prostheses drastically change a patient’s lifestyle and thromboembolic and bleeding events are not rare [12–14]. We present one of the largest single-center Ross experiences with a follow-up extending well into the second decade, which has been assumed to be the critical mark for autograft stability, thus providing critical information for further judgment of this surgical concept.
Accepted for publication July 23, 2013. Address correspondence to Dr Weimar, Sana Cardiac Surgery Stuttgart GmbH, Herdweg 2, 70174 Stuttgart, Germany; e-mail: timo.weimar@ yahoo.de.
Ó 2014 by The Society of Thoracic Surgeons Published by Elsevier Inc
ˇ
ith its introduction 40 years ago, the Ross procedure was expected to eliminate most of the disadvantages associated with conventional aortic valve replacement [1]. Its excellent hemodynamic performance, unique reproduction of the complex function of a healthy valve creating normal transvalvular flow dynamics, avoidance of anticoagulation therapy, and allowance of unrestrained activity as well as its growth potential made the pulmonary autograft a promising treatment option for both congenital and acquired disease of the left ventricular (LV) outflow tract [2–5]. All this comes at the cost of a double-valve procedure that is technically challenging. When midterm results emerged, initial enthusiasm was followed by the recognition of a possible deterioration of the autograft with time [6, 7]. Dilatation of the neoaortic root was found to be the principal cause of autograft failure [6, 8–10], and some recent publications reported
(Ann Thorac Surg 2014;97:167–74) Ó 2014 by The Society of Thoracic Surgeons
Patients and Methods The study was approved by the institutional ethics committee, and informed consent was obtained from all patients. From February 1995 through May 2012, a Ross procedure was performed on 645 consecutive patients at Sana Cardiac Surgery Stuttgart. This reflected 6.1% of a total of 10,574 isolated aortic valve procedures performed during 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2013.07.078
ADULT CARDIAC
Quo Vadis Pulmonary Autograft—The Ross Procedure in Its Second Decade: A Single-Center Experience in 645 Patients
168 ADULT CARDIAC
WEIMAR ET AL THE ROSS PROCEDURE IN 645 PATIENTS
Ann Thorac Surg 2014;97:167–74
the same period, including 8,933 prosthetic valve replacements and 996 aortic valve reconstructions. Patients’ characteristics are shown in Table 1. Indications included severe aortic valve or root disease with the patient’s preference to avoid oral anticoagulation, contraindication for oral anticoagulation, childbearing potential in women, and an active lifestyle. Contraindications included significantly reduced LV function, more then two-vessel coronary artery disease, connective tissue or active rheumatic disorders, anomalies of the pulmonary valve, and a shortened life expectancy for other reasons.
Surgical Technique Perioperative data are shown in Table 2. The majority of patients received a freestanding aortic root replacement without (RR; n ¼ 224 [35%]; 1995–1998) or with (RRþR; n ¼ 406 [63%]; 1999–2012) autograft reinforcement. Fifteen patients (2%) were operated on with the subcoronary
Table 2. Perioperative Variables Characteristics Operative technique Root replacement Subcoronary implantation Concomitant procedures Ascending aorta replacement Ascending aorta remodeling CABG Other valve surgery Mean CPB time (min [ SD]) Mean cross-clamp time (min [ SD]) Circulatory arrest for arch surgery Autograft reinforcement Annulus Sinotubular junction Both CABG ¼ coronary artery bypass grafting; bypass; SD ¼ standard deviation.
Value 97.7% (n ¼ 630) 2.3% (n ¼ 15) 51.8% (n ¼ 334) 32.4% (n ¼ 209) 9.1% (n ¼ 59) 8.7% (n ¼ 56) 1.6% (n ¼ 10) 165 35 133 24 4.3% (n ¼ 28) 62.9% (n ¼ 406) 33.1% (n ¼ 213) 25.4% (n ¼ 164) CPB ¼ cardiopulmonary
Table 1. Patient Characteristics Characteristics
Value
Male sex Mean age (y [ SD, range]) <20 20–40 41–60 >60 Prior cardiac surgery Aortic valve repair Aortic valve replacement Other Predominant aortic hemodynamics Stenosis Insufficiency Mixed lesion Aortic valve morphology Bicuspid Prosthetic dysfunction Active endocarditis NYHA functional class I II III IV Unknown Systolic left ventricular function Good (EF >0.50) Impaired (EF 0.40–0.50) Moderate/bad (EF <0.40) Unknown Type of operation Elective Urgent EF ¼ ejection fraction; standard deviation.
75.5% (n ¼ 487) 42.3 14.2 (1–68) 10.9% (n ¼ 70) 26.7% (n ¼ 172) 58.4% (n ¼ 377) 4.0% (n ¼ 26) 14.1% (n ¼ 91) 5.3% (n ¼ 34) 6.4% (n ¼ 41) 2.5% (n ¼ 16) 31.5% (n ¼ 203) 28.8% (n ¼ 186) 33.3% (n ¼ 215) 57.8% (n ¼ 373) 6.4% (n ¼ 41) 6.0% (n ¼ 39) 31.9% 42.5% 15.7% 4.5% 5.4%
(n (n (n (n (n
¼ ¼ ¼ ¼ ¼
206) 274) 101) 29) 35)
74.9% 8.5% 0.2% 16.4%
(n (n (n (n
¼ ¼ ¼ ¼
483) 55) 1) 106)
98.1% (n ¼ 633) 1.9% (n ¼ 12)
NYHA ¼ New York Heart Association;
SD ¼
implantation technique and were excluded from analyses of freedom from autograft reoperation. Both techniques have been previously described in detail [15–17]. In the RRþR group, a polyethylene terephthalate fiber (Dacron) strip was incorporated in the proximal suture line, and a second suture line fixed the aortic wall remnant to the autograft to support the annulus (Fig 1). In most cases, the right ventricular outflow tract (RVOT) was reconstructed with a cryopreserved homograft (pulmonary, n ¼ 584 [91%]; aortic, n ¼ 8 [1%]), although 53 patients (8%) received a porcine stentless xenograft (Freestyle; Medtronic, Inc, Minneapolis, MN).
Data Acquisition All patients were prospectively followed up with annual visits obtaining clinical details and serial standardized echocardiography. All postoperative events as well as structural and nonstructural valve deterioration were defined according to the Guidelines for Reporting Morbidity and Mortality After Cardiac Valvular Interventions [18]. The database was frozen for analyses in May 2012. The total follow-up was 5,349 patient-years and was 96% complete, with a mean follow-up duration of 8.4 4.6 years (range, 0 to 17.4 years).
Statistical Analysis Continuous data were reported as mean standard deviation; categorical data were expressed as absolute numbers or proportions. The survival was compared with the age- and sex-matched general German population (www.destatis.de; Statistisches Bundesamt, Wiesbaden, Germany). Relative survival was expressed as the ratio of expected versus observed numbers of death [19]. The actuarial estimates of freedom from reoperations were calculated using the Kaplan-Meier method. The risk of reoperation is presented as the smoothed instantaneous rate of a patient requiring reoperation within an interval
WEIMAR ET AL THE ROSS PROCEDURE IN 645 PATIENTS
169
Fig 1. Annulus reinforcement (autograft invaginated into the left ventricular outflow tract for implantation). (A, B) First suture line underneath the cusps incorporating a Dacron strip. (C) Second suture line to fix the remnant aortic wall to the autograft. (D) Schematic cross section.
(t, t þ dt), provided the patient was not censored until the beginning of t [20]. To identify predictors for shorter time to reoperation, we performed univariate analyses using the Cox proportional-hazard regression method on 15 preoperative and perioperative variables. Multivariate Cox proportional-hazard models were used to confirm whether significant univariate predictors (p < 0.100) persisted in the presence of other variables. For longitudinal analyses of echocardiographic measurements, we used a random-effects model, assuming correlation between repeated follow-up measurements and a random patient effect [21]. Valve regurgitation was modeled using multinomial ordinal models, with a random patient effect and a piecewise linear relationship with follow-up time with knots at {1, 3, 5, ., 17} years. Statistical analyses were performed using R version 2.15.2 (Development Core Team, R Foundation for Statistical Computing, Vienna, Austria).
Results Patients’ Survival There were no intraoperative deaths. The 30-day mortality was 0.9% (6 of 645 patients; Table 3), caused by multiorgan failure, myocardial infarction, right ventricular failure, hepatorenal syndrome, sudden death, and arrhythmia (n ¼ 1, respectively). There were 27 late deaths. In 15 of those a cardiac cause could be demonstrated, resulting in a linearized occurrence rate (LOR) of 0.5%/patient-year. The overall cumulative survival at 15 years was 92.7% (95% confidence interval [CI], 90.1% to 95.3%) and did not differ from the general German population (Fig 2A). The subcohort of those younger than 50 years of age revealed 13 late deaths (expected, 8.3; p ¼ 0.100), whereas there were 14 late deaths (expected, 18.9; p ¼ 0.260) in patients 50 years or older.
Reoperations Sixty-five patients (10.1%) underwent a total of 78 valverelated reoperations (LOR, 1.5%/patient-year). Reoperative hospital mortality was 3.8% and was caused by endocarditis with multiorgan failure in all cases (n ¼ 3). Freedom from reoperation at 12 years was 88.6% (95% CI, 85.1% to 92.2%; Fig 2B). Six patients underwent combined
autograft and homograft reoperations and are presented in the respective valve reinterventions. PULMONARY AUTOGRAFT. Forty-four patients (6.8%) required 47 reinterventions on the autograft, of which 22 (47%) were performed as a valve-sparing procedure. In 11 patients, autograft dysfunction occurred within 1 year and was not related to autograft dilatation but caused by endocarditis (n ¼ 8), rupture of the proximal suture line (n ¼ 1), covered rupture of the proximal autograft (n ¼ 1), and development of aortic root pseudoaneurysm (n ¼ 1). Thirty-four patients underwent 36 reinterventions for late (1 year) autograft failure after a mean interval of 7.5 4.4 years (median, 7.0 years; range, 1.2 to 14.9 years). The cause for autograft dysfunction was endocarditis in 4, structural valve disease in 11, and nonstructural valve disease in 21 reinterventions. Three patients received simultaneous reoperation on the autograft and the pulmonary allograft. In 18 of 36 late reinterventions (50%), the autograft could be preserved (David procedure in most cases). Freedom from reoperation on the autograft at 12 years was 91.6% (95% CI, 88.5% to 94.9%) (Fig 2C), with
Table 3. Clinical Course Characteristics Early (30 days) Mortality Reoperation for bleeding/ tamponade Myocardial infarction Deep sternal wound infection Cerebrovascular accidentsa Pacemaker implantation Procedure-related CABG Late (>30 d) Mortality Thromboembolic eventsa TIA Hemorrhagic events Endocarditis a
Value 0.9% (n ¼ 6) 4.9% (n ¼ 32) 0.15% 0.46% 1.4% 0.93% 1.4%
(n (n (n (n (n
¼ ¼ ¼ ¼ ¼
1) 3) 9) 6) 9)
4.2% 1.9% 0.93% 0.46% 2.6%
(n (n (n (n (n
¼ ¼ ¼ ¼ ¼
27) 12), 0.22%/py 6), 0.11%/py 3), 0.06%/py 17), 0.32%/py
Including transient ischemic attack (TIA).
CABG ¼ coronary artery bypass graft;
py ¼ patient-year.
ADULT CARDIAC
Ann Thorac Surg 2014;97:167–74
170 ADULT CARDIAC
WEIMAR ET AL THE ROSS PROCEDURE IN 645 PATIENTS
Ann Thorac Surg 2014;97:167–74
Fig 2. (A) Survival of Ross patients versus general German population, sex- and age-matched. (B) Freedom from autograft (AG) or pulmonary allograft (HG) reoperation. (C) Freedom from autograft reoperation. (D) Freedom from pulmonary allograft reoperation.
Table 4. Freedom From Reoperation on the Pulmonary Autografta Characteristics Aortic valve lesion Stenosis Insufficiency Mixed Preoperative aortic morphology Nonbicuspid Bicuspid Aortic annulus diameter 26 mm >26 mm Sex Female Male a
5 Years
10 Years
96.8 (94.3–99.4) 94.3 (91.0–97.8) 97.6 (95.6–99.7)
96.8 (94.3–99.4) 89.8 (85.1–94.8) 96.9 (94.5–99.4)
p Value 0.006
0.555 96.2 (93.7–98.7) 96.3 (94.2–98.4)
94.3 (91.1–97.6) 94.2 (91.3–97.1) 0.002
99.2 (98.1–100) 93.9 (91.2–96.7)
97.8 (95.7–100) 91.2 (87.7–94.8)
98.4 (96.3–100) 95.5 (93.6–97.5)
98.4 (96.3–100) 93.0 (90.3–95.7)
0.129
Values show percentages (95% confidence intervals). Log rank probability values are reported.
no difference between the RR and the RRþR groups (p ¼ 0.781). Freedom from autograft reoperation depending on anatomic morphology is shown in Table 4. Univariate analysis identified pure aortic insufficiency (hazard ratio [HR], 2.8; 95% CI, 1.4 to 5.3; p < 0.001) and annulus diameter (HR, 1.1/mm; 95% CI, 1.0 to 1.2; p ¼ 0.040) as an independent predictor of reoperation. Multivariate analysis revealed an annulus diameter greater than 26 mm as a significant predictor of shorter time to autograft reoperation (HR, 2.8; 95% CI, 1.1 to 6.8; p ¼ 0.028), which persisted in the presence of pure aortic insufficiency. RIGHT VENTRICULAR OUTFLOW TRACT ALLOGRAFT. Twenty-nine patients (4.5%) required 37 reinterventions on the RVOT after a mean interval of 4.8 3.5 years (median, 3.7 years; range, 0.1 to 12.9 years). Stenosis was the main cause for replacement (n ¼ 30). Homograft endocarditis was the cause for reoperation in 3 patients; 2 of them had double-valve endocarditis. From 53 implanted xenografts (mean follow-up, 4.3 0.8 years), 7 (13%) had to be replaced mainly as a result of excessive tissue ingrowth at the proximal anastomosis
WEIMAR ET AL THE ROSS PROCEDURE IN 645 PATIENTS
171
Fig 3. (A, B) Postoperative probability of a patient being in each valve insufficiency grade. (C) Average pressure gradients across the right ventricular outflow tract (RVOT). (D) Longitudinal changes in sinotubular junction (STJ) and aortic annulus diameters. (CI ¼ confidence interval.)
causing stenosis. Freedom from reoperation on the RVOT (including xenografts) at 12 years was 95.0% (95% CI, 92.8% to 97.2%; Fig 2D). Cox regression analysis identified stentless xenograft (HR, 14.3; 95% CI, 4.5
to 33.3; p < 0.001), younger donor age (HR, 1.04/year; 95% CI, 1.01 to 1.02; p < 0.001), and younger patient age (HR, 1.06/year; 95% CI, 1.08 to 1.10; p < 0.001) as independent predictors of reoperation.
Fig 4. (A) Body surface area (BSA) indexed left ventricular (LV) mass preoperative and at latest follow-up (FU); p < 0.001. (B) Longitudinal changes in left ventricular mass. Dotted lines denote 160 g/m2 (mean þ one standard deviation of a normotensive control population) [22].
ADULT CARDIAC
Ann Thorac Surg 2014;97:167–74
172 ADULT CARDIAC
WEIMAR ET AL THE ROSS PROCEDURE IN 645 PATIENTS
Ann Thorac Surg 2014;97:167–74
Echocardiographic and Functional Status
reoperation at 12 years was 91.0% (95% CI, 87.3% to 94.8%), and freedom from aortic insufficiency of at least grade II was 93.4% (95% CI, 90.1% to 96.7%).
After excluding patients who had died, had reoperated valves, or had not completed 1 year of follow-up by the time of the database freeze, echocardiographic follow-up was available in 545 patients. Freedom from aortic insufficiency of at least grade II at 12 years was 91.4% (95% CI, 88.2% to 94.7%). The longitudinal probability of valve insufficiency occurrence as well as the development of RVOT gradients is shown in Figure 3. At last follow-up, mean and maximum pressure gradients across the autograft were 3.3 1.4 mm Hg (range, 1 to 11 mm Hg) and 6.3 2.7 mm Hg (range, 2 to 20 mm Hg), respectively, with 95% of patients presenting with aortic insufficiency of at least grade I. The mean and maximum pressure gradients across the RVOT were 9.4 6.7 mm Hg (range, 0 to 71 mm Hg) and 17.2 10.9 mm Hg (range, 0 to 122 mm Hg), respectively, with 90% showing a regurgitation of at least grade I. Longitudinal analyses revealed a small albeit statistically significant increase in sinotubular junction (STJ) diameters and annulus diameters (3.0 mm/decade and 0.1 mm/decade, respectively; p < 0.001) with no clinical significance or implication (Fig 3D). No effect of the applied technique (RR versus RRþR) on autograft diameters was observed (p > 0.2 for all comparisons). There was a significant LV mass regression (mean, 15.4 1.2 g/m2 per year; p < 0.001), which was mostly complete within 2 years after surgery (Fig 4). At latest follow-up, 76.1% of patients were within one standard deviation of the LV mass of normotensive individuals [22], and 95% of patients presented in New York Heart Association functional class I.
Endocarditis The LOR of endocarditis requiring medical or surgical therapy was 0.1%/patient-year and 0.2%/patient-year for the autograft and RVOT allograft, respectively. Most patients with autograft endocarditis (12 of 13 patients) required reoperation (mean interval, 3.5 5.25 years), whereas only 50% of patients with homograft endocarditis (3 of 6 patients) were in need of reoperation (mean interval, 0.15 0.10 years). None of the patients who had endocarditis as an indication for the initial Ross operation exhibited recurrent late endocarditis.
Thromboembolic or Hemorrhagic Events The LOR for thromboembolic (including transient ischemic attacks) or major bleeding events was 0.24%/ patient-year. The permanent stroke rate was 0.07%/ patient-year.
Follow-Up Exceeding 10 Years A total of 269 patients (41%) had been operated on before January 01, 2002, thus potentially having a follow-up of more than 10 years. Patients operated on before January 01, 2002, who experienced reoperation or were deceased before reaching a 10-year follow-up were also included. Mean follow-up was 12.6 2.6 years (range, 0.5 to 17.3 years; total of 3,377 patient-years). Freedom from any
Comment When reviewing reports of mortality and morbidity associated with the Ross procedure, it becomes conspicuous that results are quite variable. In a meta-analysis, Tackenberg and colleagues [11] published mortality rates ranging from 0.6% to 6.8%. Although many centers have reported excellent results and low reoperation rates irrespective of the applied technique, others revealed an unacceptable occurrence of autograft deterioration as well as mortality and morbidity [4, 6–8, 16, 17, 23, 24]. This dissension seems to reflect the complexity of this procedure. However, there is wide consensus that the pulmonary autograft is unique and advantageous in its ability to change physiologically in shape and size during the cardiac cycle and to respond extremely well to various hemodynamic stimuli [25]. This series demonstrated an excellent survival as late as 15 years after the Ross procedure, equal to that of the general population. This is remarkable considering the presence of severe valvular disease with related structural alterations in the myocardium and the connective tissue. Young adults receiving other types of aortic valve replacement, whether biologic or mechanical, have especially shown a substantially lower life expectancy than the general population and seem to have an inferior outcome with the Ross procedure with regard to survival [26]. Patient selection may have had a positive bias. Nevertheless, about 25% of patients presented with reduced LV function and about 10% received concomitant procedures. Moreover, our results revealed excellent postoperative hemodynamics that led to LV mass regression and a low occurrence rate of neurologic events, all of which may have a favorable impact on patients’ survival [2–4, 16, 22]. The early mortality in our series was 0.9%. This is acceptable considering the fact that this included 49 reoperations and 31 cases of active endocarditis reflecting subgroups with a higher risk. The risk of reoperation after the Ross procedure remains a matter of concern and is critically discussed. Progressive root dilatation is one of the major fears and has been reported especially when using the root replacement technique [6, 7, 9]. It is unclear whether a predominant dilated aortic annulus would be a marker of connective tissue disorder that would also affect the autograft, eventually predicting autograft failure [23]. This study confirmed preoperative aortic valve insufficiency being an independent risk factor for reoperation on the autograft. However, patients presenting with an aortic insufficiency as a result of annulus dilatation would likely receive a David operation at present time and would not be considered primary candidates for a Ross procedure. Taking into account that the pulmonary wall, although structurally identical to the aortic wall at the time of birth, changes within the first postnatal months and shows progressive fragmentation of its elastic fibers with a
thinning of the media portion, we consider a systolic blood pressure control of 120 mm Hg or lower during the first 3 months as critical [7, 27]. This allows the autograft to adapt to increased and distending pressure by progressive deposition of collagenous tissue in the adventitia [4]. In this study, freedom from reoperation on the neoaortic root was 92% at 12 years. Furthermore, freedom from any reoperation in patients beyond a follow-up time of 10 years was 91%. Longitudinal analysis identified no systematic increase of autograft insufficiency grade with time. This has been reported for the subcoronary technique but contradicts what has been assumed for the full root technique [6, 7, 9, 28]. We could not observe a significant difference in need for reoperation in patients with or without annulus reinforcement such as that reported by others [7, 28]. This might be owing to our RR technique of keeping the autograft as short as possible and preserving its adventitia. Nevertheless, because of the knowledge we gained from the German Dutch Ross Registry, we do recommend, and continue to apply, annulus reinforcement [3, 7]. Furthermore, we believe that the importance of stabilizing the STJ might have been underestimated in the past. Luciani and associates [6] emphasized the importance of the STJ on root geometry by showing that diameters of the sinotubular ridge and proximal ascending aorta tend to equalize with the sinus of Valsalva, thereby realizing a distally pronounced root remodeling. Therefore, we lowered the critical threshold for replacement of the ascending aorta in case of ectasia or aorta ascendens–STJ mismatch to stabilize the STJ. Early autograft failure occurred in 1.7%. This was mostly related to technical errors or endocarditis and was not a result of the disease seen in late dilatation. However, Ross-related reoperations are risk-carrying procedures, and their complexity should be taken into account when considering this technique. In a recent report, Stulak and coworkers [24] revealed a broad spectrum of reoperations that might be required. Interestingly, those complex reoperations were performed with a remarkable low mortality of 1.8%. However, in our series reoperative mortality was higher at 3.8%, as all patients presented in critical state owing to active endocarditis and required urgent or emergency surgery. No mortality was observed in patients scheduled for elective reoperation. Deterioration of the RVOT conduit as a function of time was partly expected and emphasizes the problem that a single-valve disease is treated with a double-valve replacement. However, freedom from RVOT reoperation was 95% at 12 years. Homograft replacement was mainly the result of stenosis and could be performed with no mortality. The homograft cusps were generally less affected by fibroses than the wall was. Histologic findings were nonspecific, and neither rejection nor infection was seen in these cases. Carr-White and colleagues [29] rejected a correlation between the development of HLA antibodies and homograft stenosis, so the impact of allogeneic immune response as a cause of homograft stenosis remains uncertain. In 2007 and 2008, we
WEIMAR ET AL THE ROSS PROCEDURE IN 645 PATIENTS
173
implanted 53 stentless xenografts because homograft supply became temporarily limited. Because of a high incidence of unacceptable pressure gradients mostly related to tissue ingrowth we abandoned this alternative. Histologic findings were nonspecific, and a technical error seemed unlikely because gradients were normal at discharge. Endocarditis was a significant cause for reoperation and counted for almost 30% of all reinterventions in this series. However, none of the patients who had endocarditis as an indication for the initial Ross operation exhibited recurrent late endocarditis. Awareness that autologous and allograft material does not provide perfect protection from endocarditis and adoption of a more aggressive prophylaxis might help to significantly reduce the need of reoperation in the future. Although we suggested in previous reports that a Ross procedure might be superior in preventing recurrent endocarditis compared with other types of aortic valve replacement, our current findings reject this assumption, and we do not perform this procedure in patients with active endocarditis anymore [15]. Our study population with a mean age of 42 years represents a treatment dilemma. The few reports on bioprostheses in young and middle-age patients demonstrated a high occurrence of early structural valve deterioration with a striking risk of early reoperation [13, 14]. This was also true for new-generation bioprostheses, and the authors concluded that their implantation in middle-age patients has to be carefully considered. The risk of mechanical valve-related complications is not rare either and may drastically change a patient’s lifestyle [12, 13, 26]. Our study revealed a low morbidity associated with the Ross procedure. Compromising thromboembolic or bleeding events were infrequent with an LOR of 0.24%/patient-year. In contrast, the reported freedom from thromboembolism for mechanical prostheses after 10 years ranges from 67% to 91% [30]. However, it seems impossible to objectively balance the risk of reoperation after a Ross procedure against the fear of thromboembolism or bleeding events after implantation of a mechanical valve.
References 1. Ross DN. Replacement of aortic and mitral valves with a pulmonary autograft. Lancet 1967;2:956–8. 2. Elkins RC, Thompson DM, Lane MM, Elkins CC, Peyton MD. Ross operation: 16-year experience. J Thorac Cardiovasc Surg 2008;136:623–30, 630.e1–5. 3. Hanke T, Stierle U, Boehm JO, et al; German Ross Registry. Autograft regurgitation and aortic root dimensions after the Ross procedure: the German Ross Registry experience. Circulation 2007;116(11 Suppl):I251–8. 4. Yacoub MH, Klieverik LM, Melina G, et al. An evaluation of the Ross operation in adults. J Heart Valve Dis 2006;15:531–9. 5. Carr-White GS, Afoke A, Birks EJ, et al. Aortic root characteristics of human pulmonary autografts. Circulation 2000;102(19 Suppl 3):III15–21. 6. Luciani GB, Casali G, Favaro A, et al. Fate of the aortic root late after Ross operation. Circulation 2003;108(Suppl 1):II61–7. 7. Klieverik LM, Takkenberg JJ, Bekkers JA, RoosHesselink JW, Witsenburg M, Bogers AJ. The Ross operation: a Trojan horse? Eur Heart J 2007;28:1993–2000.
ADULT CARDIAC
Ann Thorac Surg 2014;97:167–74
174 ADULT CARDIAC
WEIMAR ET AL THE ROSS PROCEDURE IN 645 PATIENTS
8. David TE, Omran A, Ivanov J, et al. Dilation of the pulmonary autograft after the Ross procedure. J Thorac Cardiovasc Surg 2000;119:210–20. 9. Kouchoukos NT, Masetti P, Nickerson NJ, Castner CF, Shannon WD, Davila-Rom an VG. The Ross procedure: longterm clinical and echocardiographic follow-up. Ann Thorac Surg 2004;78:773–81. 10. Takkenberg JJ, van Herwerden LA, Galema TW, et al. Serial echocardiographic assessment of neo-aortic regurgitation and root dimensions after the modified Ross procedure. J Heart Valve Dis 2006;15:100–7. 11. Takkenberg JJ, Klieverik LM, Schoof PH, et al. The Ross procedure: a systematic review and meta-analysis. Circulation 2009;119:222–8. 12. Puvimanasinghe JP, Takkenberg JJ, Edwards MB, et al. Comparison of outcomes after aortic valve replacement with a mechanical valve or a bioprosthesis using microsimulation. Heart 2004;90:1172–8. 13. Chan V, Malas T, Lapierre H, et al. Reoperation of left heart valve bioprostheses according to age at implantation. Circulation 2011;124(11 Suppl):S75–80. 14. Tanaka H, Okita Y, Kasegawa H, et al. The fate of bioprostheses in middle-aged patients: The Japanese experience. J Heart Valve Dis 2010;19:561–7. 15. B€ ohm JO, Botha CA, Rein JG, Roser D. Technical evolution of the Ross operation: midterm results in 186 patients. Ann Thorac Surg 2001;71(5 Suppl):S340–3. 16. B€ ohm JO, Hemmer W, Rein JG, et al. A single-institution experience with the Ross operation over 11 years. Ann Thorac Surg 2009;87:514–20. 17. Sievers HH, Hanke T, Stierle U, et al. A critical reappraisal of the Ross operation: renaissance of the subcoronary implantation technique? Circulation 2006;114(1 Suppl):I504–11. 18. Akins CW, Miller DC, Turina MI, et al. Guidelines for reporting mortality and morbidity after cardiac valve interventions. Ann Thorac Surg 2008;85:1490–5. 19. Pohar M, Stare J. Relative survival analysis in R. Comput Methods Programs Biomed 2006;81:272–8. 20. Charitos EI, Stierle U, Hanke T, Schmidtke C, Sievers HH, Richardt D. Long-term results of 203 young and middle-aged
Ann Thorac Surg 2014;97:167–74
21.
22. 23. 24. 25. 26.
27. 28.
29.
30.
patients with more than 10 years of follow-up after the original subcoronary Ross operation. Ann Thorac Surg 2012;93:495–502. Andrinopoulou ER, Rizopoulos D, Jin R, Bogers AJ, Lesaffre E, Takkenberg JJ. An introduction to mixed models and joint modeling: analysis of valve function over time. Ann Thorac Surg 2012;93: 1765–72. Duebener LF, Stierle U, Erasmi A, et al; German Ross Registry. Ross procedure and left ventricular mass regression. Circulation 2005;112(9 Suppl):I415–22. David TE, Woo A, Armstrong S, Maganti M. When is the Ross operation a good option to treat aortic valve disease? J Thorac Cardiovasc Surg 2010;139:68–75. Stulak JM, Burkhart HM, Sundt TM 3rd, et al. Spectrum and outcome of reoperations after the Ross procedure. Circulation 2010;122:1153–8. Yacoub MH, Cohn LH. Novel approaches to cardiac valve repair: from structure to function: part I. Circulation 2004;109: 942–50. van Geldorp MW, Eric Jamieson WR, Kappetein AP, et al. Patient outcome after aortic valve replacement with a mechanical or biological prosthesis: weighing lifetime anticoagulant-related event risk against reoperation risk. J Thorac Cardiovasc Surg 2009;137:881–6, 886e1–5. Wagenvoort CA, Neufeld H, Birge RF, Caffrey JA, Edwards JE. Origin of right pulmonary artery from ascending aorta. Circulation 1961;23:84–90. Charitos EI, Takkenberg JJ, Hanke T, et al. Reoperations on the pulmonary autograft and pulmonary homograft after the Ross procedure: an update on the German Dutch Ross Registry. J Thorac Cardiovasc Surg 2012;144: 813–23. Carr-White GS, Kilner PJ, Hon JK, et al. Incidence, location, pathology, and significance of pulmonary homograft stenosis after the Ross operation. Circulation 2001;104(12 Suppl 1):I16–20. Akins CW. Results with mechanical cardiac valvular prostheses. Ann Thorac Surg 1995;60:1836–44.