Spectrum and Outcome of Reoperations After the Ross ProcedureClinical Perspective
Background—Proposed advantages to the Ross procedure included presumed increased freedom from reoperation and simpler reoperation for pulmonary conduit replacement if needed. It is increasingly apparent, however, that reoperations are frequent after the Ross procedure and that when required, they may be more complex than previously thought.
Methods and Results—Between September 1991 and August 2008, 56 patients underwent reoperation at our institution after a Ross procedure performed by ourselves (n=13) or elsewhere (n=43). Median age at first reoperation at our institution was 26 years (range 1 to 69 years). The 4 most common indications for reoperation were isolated autograft (neoaortic) regurgitation in 11 cases (20), isolated pulmonary conduit regurgitation/stenosis in 9 (16), combined autograft regurgitation/dilatation in 8 (14), and combined autograft regurgitation and pulmonary conduit regurgitation/stenosis in 6 (11). A total of 144 procedures were performed in these 56 patients during first reoperation at our institution. The autograft valve required replacement in 21 cases (38) and aortic root replacement in 21 (38), with ascending aortic/arch reconstruction in 13 (23) and mitral valve surgery in 5 (9). The pulmonary valve was replaced in 33 cases (59) and the tricuspid valve was repaired/replaced in 10 (18). Early mortality was 1.8 (1 of 56 patients), and morbidity included 6 patients with respiratory failure and 3 who required postcardiotomy extracorporeal membrane oxygenation. There were 4 late deaths during the median follow-up of 8 months (range 1 to 179 months).
Conclusions—A broad spectrum of complex reoperations may be required after the Ross procedure. Patients and family members considering the procedure should be informed of the potential for associated morbidity should reoperation be necessary.
The Ross procedure was introduced in 1967 as a valuable option for both congenital and acquired disease of the left ventricular outflow tract1,–,6 with the argument that the pulmonary autograft offered growth potential in addition to excellent hemodynamics, freedom from anticoagulation and hemolysis, and durability that should be superior to that of a homograft.7,–,11 Furthermore, in addition to decreasing the need for further reoperation, when required, the reoperation was presumed to be simpler, because it would most likely entail replacement only of the pulmonary homograft.
Editorial see p 1139
Clinical Perspective on p 1158
In recent years, however, there has been increasing recognition that reoperation after the Ross operation is not infrequent. Furthermore, these repeat operations are required not only for replacement of right ventricular outflow conduits12 but also for autograft dilatation and dysfunction,13 as well as several other indications (eg, mitral and/or tricuspid disease and coronary abnormalities). Some have reported that reoperation after the Ross procedure may be required more often in patients with bicuspid aortic valve and aortic valve regurgitation as the indication for the Ross procedure, which suggests that perhaps the procedure should be avoided or modified in these individuals.14,15 Regardless, to make an informed decision about offering or recommending the Ross procedure, one needs to understand not only the frequency of reoperation but also the full spectrum of pathology and complexity of procedures that may be required at that reoperation. In the interest of clarifying the latter, we reviewed our growing experience with reoperation after the Ross procedure.
The Mayo Foundation Institutional Review Board approved this study, and all patients or their families gave written informed consent. From September 1991 to August 2008, 56 patients (38 males) who had previously undergone the Ross procedure had reoperative cardiac surgery at our institution. In 13 of these patients, the Ross procedure had been performed at our institution, and in 43, it had been performed elsewhere. During this same time interval, we performed the Ross procedure in 39 patients; however, the focus of the present review is on the nature and outcome of reoperation(s) after the Ross procedure.
The patients in the present series underwent their Ross procedures between September 1989 and October 2005 at a median age of 21 years (range 3 months to 67 years). The indications for the Ross procedure in the present patient cohort are listed in Table 1; the 3 most common indications were bicuspid aortic valve in 26 patients, congenital aortic stenosis in 18, and isolated aortic regurgitation in 5. The Ross procedure was the patient's first operation in 37 cases, the second in 11, the third in 4, the fourth in 3, and the fifth in 1. Among the 19 patients who had at least 1 operation before the Ross, a total of 32 operations had been performed.
Nine patients had 1 reoperation and 1 patient had 2 reoperations elsewhere after their Ross procedure but before their first reoperation at our institution, the indications for which are listed in Table 2. Among these 10 patients, 11 surgical procedures were performed (Table 3), and 8 had attempted percutaneous interventions, including successful pulmonary conduit stenting in 4 and successful balloon dilation of the native pulmonary artery and conduit stenosis in 1 patient each.
The median age at the patients' first reoperation at our institution was 26 years (range 1 to 77 years). This reoperation represented the patient's second operation in 32 cases, the third in 11, the fourth in 8, the fifth in 4, and the sixth in 1. For patients in whom the first reoperation at our institution represented the first reoperation after their Ross procedure, the median time interval between the 2 operations was 6.5 years (range 1 day to 12.5 years.).
On presentation for reoperation, 19 patients had preoperative right ventricular dysfunction; it was mild in 10, moderate in 6, and severe in 3. Twenty-four patients had preoperative right ventricular enlargement; mild enlargement was present in 7 patients, moderate in 11, and severe in 6. The median preoperative left ventricular ejection fraction was 58% (range 23% to 71%). Fifteen patients had left ventricular ejection fraction <50%, and 3 had left ventricular ejection fraction <30%. Ventricular function was assessed preoperatively with transthoracic echocardiography in the majority of patients, and transesophageal echocardiography was used in select patients if more precise delineation of anatomy and pathology was required. Three patients had pacemakers implanted between the Ross procedure and the first reoperation at our institution.
Demographic and other patient-related data were obtained from Mayo Clinic medical records. Follow-up information was obtained from clinic visits and written correspondence from local physicians. Continuous data were expressed either as mean±SD or median with a range; categorical data were summarized with frequencies or counts. Early mortality was defined as death that occurred within 30 days of operation or any time during the operative hospitalization.
Reoperations after the Ross procedure have increased significantly in our practice in recent years; only 8 cases were performed before the year 2000, and of the 48 performed this decade, 28 (58%) were performed in the last 3 years of the study period. Indications for surgery at the first reoperation performed at our institution are shown in Table 4.
Most patients underwent multiple concomitant procedures during the first reoperation. In all, 144 procedures were performed in these 56 patients (Table 5). Only 21 patients (38%) underwent an isolated procedure at the first reoperation, which consisted of right ventricle–to–pulmonary artery (RV-PA) conduit replacement in 7 patients, aortic valve replacement (AVR) in 7, aortic root replacement in 6, and coronary artery bypass grafting in 1. Concomitant pathology that was present and ultimately required intervention included severe tricuspid valve regurgitation in 10 patients, moderate or greater mitral valve regurgitation in 4, moderate-severe mitral valve stenosis in 1, left ventricular aneurysm in 1, and atrial septal defect in 1. Of 5 patients who required surgery for proximal coronary artery abnormalities as the major indication for surgery at the first reoperation at our institution, 1 presented with cardiac arrest from proximal left main stenosis. Coronary artery surgery was performed as a concomitant operation in the remaining 4 patients. Two patients had prior coronary artery bypass grafting of 2 arteries immediately after their original Ross procedure because of left main coronary artery compression. During the patients' first reoperation at our institution the prior vein graft to an obtuse marginal coronary artery required transection (and subsequent revision) to expose the right ventricular outflow tract. A third patient underwent mechanical AVR, during which the aortic valve prosthesis was in close proximity to a small left main coronary ostium. Refractory hypotension during separation from bypass was presumed to be due to left coronary artery compromise that resulted in left ventricular dysfunction. A vein graft was placed to the left anterior descending coronary artery. Predismissal imaging ultimately demonstrated normal flow in the left main coronary artery. A fourth patient had prior coronary artery bypass grafting of 2 arteries (left internal mammary artery to left anterior descending coronary artery, saphenous vein graft to obtuse marginal coronary artery) at the time of his Ross procedure for atherosclerotic coronary artery disease. During the patient's first reoperation at our institution, he underwent aortic root replacement and saphenous vein graft bypass for an atherosclerotic lesion in the right coronary artery.
Of the 26 patients with an original cardiac diagnosis of bicuspid aortic valve, 21 (81%) required reoperation for autograft or ascending aortic pathology; severe autograft regurgitation was present in 17 patients, autograft dilatation in 12, and supra-autograft ascending aortic aneurysm in 5. Concomitant severe RV-PA conduit stenosis/regurgitation was present in 8 patients, and isolated right-sided pathology was present in 5 patients. Aortic root replacement was performed in 12 patients, autograft valve replacement in 8 patients, and supra-autograft ascending aortic replacement in 7. The RV-PA conduit was replaced in 15 patients; replacement was performed for severe conduit stenosis/regurgitation in 13 and moderate stenosis/regurgitation in 2.
Interventions were also required on the mitral and tricuspid valves during the current reoperation. Tricuspid valve replacement was required in 2 patients, in 1 instance owing to severe annular and severe right ventricular dilatation and in the other to dysplastic leaflets tethered by preexisting defibrillator leads. Mitral valve replacement was required in 2 patients because of leaflet/annular calcification in 1 patient and chordal/leaflet shortening and fibrosis in the other.
Subsequent to the first reoperation at our institution, further reoperation has been required in 7 patients to date. AVR was performed in 4 patients; coronary artery bypass grafting in 1; combined ascending aortic replacement, AVR, and pulmonary valve replacement in 1; and combined aortic root replacement, pulmonary valve replacement with right ventricular outflow tract reconstruction, mitral valve replacement, and tricuspid valve annuloplasty in 1.
Median total cardiopulmonary bypass time for all reoperations was 125 minutes (range 29 to 367 minutes); 45 patients required aortic cross-clamping, with a median cross-clamp time of 83 minutes (range 22 to 143 minutes). Three patients required deep hypothermic circulatory arrest with a median arrest time of 17 minutes (range 13 to 20 minutes) for aortic arch surgery.
The single early death occurred in a patient who underwent the Ross procedure at 39 years of age for bicuspid aortic valve with severe aortic regurgitation and required concomitant coronary artery bypass grafting of 2 arteries at the time of the Ross procedure for left main coronary artery compression. The indication for surgery at the first reoperation was RV-PA conduit stenosis/regurgitation and severe tricuspid regurgitation with severe right ventricular dysfunction and enlargement. He underwent combined pulmonary valve replacement, tricuspid valve annuloplasty, and revision of prior coronary artery bypass grafting. Postoperatively, the patient had adult respiratory distress syndrome and died of multisystem organ failure after 69 days in the hospital.
Despite the young patient age, the median intensive care unit stay after reoperation was 2 days (range 1 to 28 days). Early significant nonfatal morbidity included respiratory failure in 6 patients, transient atrial arrhythmias in 5, extracorporeal membrane oxygenation in 3, renal failure in 2, and pneumonia in 2. In all 3 cases, extracorporeal membrane oxygenation was instituted for hemodynamic instability with separation from cardiopulmonary bypass. After a period of recovery on extracorporeal membrane oxygenation support (a mean of 62 hours), all 3 patients were weaned successfully and progressed to hospital dismissal. No new permanent pacemakers were required.
Median follow-up since the original Ross operation in all patients was 7.6 years (range 17 to 160 months), and median follow-up was 8 months after first reoperation at our institution (range 1 to 179 months) and was available for 49 of 55 early survivors. There have been 4 late deaths in the follow-up period. Two were due to unknown causes, and the other 2 were operative deaths during subsequent complex reoperations for multivalvular pathology.
Taken together, the “reoperation burden” after the initial Ross procedure in these 56 patients totaled 168 surgical procedures extending to the last follow-up. This included 10 patients who underwent 11 procedures in the interval between their original Ross procedure and reoperation at our institution, 56 who underwent 144 procedures at our institution, and 7 who underwent 13 subsequent procedures during follow-up after having reoperation at our institution. An additional 8 patients have undergone subsequent percutaneous intervention for pathology of their RV-PA conduit. Importantly, in the present patient cohort of 56 patients, 44 patients (79%) ultimately required AVR or aortic root replacement of the failed autograft during the relatively short median follow-up period after their Ross procedure.
The results of the present study demonstrate that reoperations after the Ross procedure, when required, may be complex, frequently involve multiple structures, and carry significant morbidity. Furthermore, reoperations are often multiple. This is particularly significant given the young age of the patient population and the many years during which they will be at risk. In our experience, the autograft itself is often lost as well. These observations should be considered when patients and surgeons select among the many options for replacement of the aortic valve. The Ross pulmonary autograft procedure was introduced with the hope that it would represent the ideal solution for aortic valve disease, combining freedom from reoperation and freedom from anticoagulation with optimum hemodynamics. It was believed that the autograft would function normally, perhaps indefinitely, and that the pulmonary homograft would be particularly durable given the lower hemodynamic stresses of the pulmonary circulation. If reoperation was needed, it was thought that it would likely be a straightforward pulmonary conduit replacement with low operative risk. In addition, in the specific case of the growing child, it was anticipated that the autograft would grow. Despite the theoretical advantages of the Ross procedure, there is an increasing experience with reoperation(s) after the Ross procedure, with overall left- or right-sided valve–related intervention rates ranging from 14% to 37% extending up to 16 years.16,17 Autograft dilatation17,–,19 and pulmonary homograft dysfunction16,–,18 have respectively brought 17% to 34% and 18% to 21% of patients to reoperations at 10 to 16 years. The present data do not permit us to assess the frequency of reoperation, because many of these patients underwent their Ross operation elsewhere, nor was that the aim of the present study. In addition to the numeric frequency of the operation, however, an overall assessment of the morbidity associated with the operation must also consider the magnitude and complexity of some of the reoperative procedures that may be required. Despite a low early mortality in the present series (1.8%), morbidity was considerable, particularly when contrasted with a more straightforward reoperation for rereplacement of a dysfunctional bioprosthesis.
The principal theoretical downside to the Ross procedure is that in addition to preexisting left-sided pathology, the Ross procedure itself creates right-sided pathology. Indications for percutaneous intervention or surgery on the right ventricular outflow tract after the Ross procedure most commonly include RV-PA conduit stenosis or regurgitation.12,20 In the present series, 33 patients underwent reoperation on the RV-PA for pulmonary conduit stenosis/regurgitation; however, this was an isolated procedure in only 7 patients (21%). Furthermore, 5 of these individuals had incurred severe right ventricular dysfunction. Although discrete RV-PA conduit stenosis can be treated in some settings with percutaneous stenting and excellent midterm outcomes (but at the expense of the creation of pulmonary regurgitation),21 only 8 patients in the present series had undergone percutaneous transcatheter interventions for stenosis in the RV-PA conduit and native pulmonary artery stenosis, and only 6 of these interventions were successful. The use of percutaneous pulmonary valve replacement will have greater applicability in this setting, particularly when the obstruction is limited to the valvar (homograft) level.
It is clear that it is in the pediatric population, particularly young children, that the Ross procedure has its greatest role. Reoperation is inevitable in the majority of these children, regardless of prosthesis type, because of somatic growth or prosthetic valve dysfunction. The specter of early bioprosthetic valve deterioration and the challenges of warfarin anticoagulation make the Ross procedure an appealing choice when AVR is necessary. In our pediatric practice, the Ross operation is our procedure of choice when AVR is required; among adults, however, we are highly selective. Compared with a mechanical prosthesis, the Ross is likely significantly less durable. Although Yacoub et al22 reported an overall freedom from pulmonary homograft reoperation of 94.9% and freedom from autograft reoperation of 92.9% at 10 years, certain populations have been shown to fare significantly worse than this. This includes patients with rheumatic heart disease,23 preoperative aneurysm,24 and predominant preoperative aortic regurgitation.25 A recent report from David et al26 demonstrated that the Ross procedure provided suboptimal results in male patients with aortic regurgitation. The best outcomes were achieved in female patients, those with aortic stenosis, and those with an aortic annulus <27 mm. The technique of implantation did not affect late outcome. Importantly, those with a dilated aortic annulus did not fare well, despite annular reduction procedures being applied at the time of the Ross procedure. Although there is an abundance of literature reporting reoperation rates after the Ross procedure, these reoperation rates alone do not reflect the valve disease that patients may be harboring who have not yet come to reoperation.
The preoperative evaluation of the patient who requires reoperation after the Ross procedure is critical. In addition to conducting a review of prior operative reports, we obtain preoperative transthoracic echocardiograms (selective transesophageal echocardiography) and computerized tomography or magnetic resonance imaging to obtain right ventricular volume and ejection fraction, anatomy of the aortic root, and coronary artery anatomy. Cardiac catheterization to assess hemodynamics or coronary angiography is performed selectively when other imaging modalities are equivocal. The dilated autograft or the right coronary artery button can be in close proximity to the sternum, and both are vulnerable to injury during resternotomy. Previously placed vein grafts to the anterior descending or obtuse marginal coronary arteries may be at risk during resternotomy or during pulmonary valve replacement.
The timing of reoperation(s) usually revolves around pathology of the pulmonary autograft, as well as the right ventricular outflow tract and pulmonary homograft valve. Autograft regurgitation with or without dilatation and pulmonary conduit stenosis/regurgitation are the most common lesions, although concomitant abnormalities such as supra-autograft ascending aortic dilatation, mitral and tricuspid valve regurgitation, and coronary artery abnormalities may also be present. In addition to the standard indications for intervention for neoaortic (autograft) valve regurgitation,27 class I indications for reintervention in patients with RV-PA conduit bioprosthetic pulmonary valve dysfunction include severe stenosis (peak gradient >50 mm Hg), conduit regurgitation, or any of the following: (1) Decreased exercise capacity, (3) depressed right ventricular function, (3) at least moderately enlarged right ventricular end-diastolic size, and (4) at least moderate tricuspid valve regurgitation.28
Although the strategy of prosthesis selection for these patients at the time of reoperation is individualized, the following is our general algorithm. If an isolated pulmonary conduit replacement is required, we prefer a tissue valve, either stented or stentless bioprosthesis. We have used the Contegra valved bovine conduit (Medtronic, Minneapolis, Minn); however, we rarely use a pulmonary homograft in the adult. All of these bioprosthesis should allow placement of a percutaneous pulmonary valve in the future. We do believe there is a role for mechanical pulmonary valve replacement (albeit uncommonly), particularly in this patient population. We consider mechanical pulmonary valve replacement in the patient who has previously undergone the Ross procedure when mechanical AVR is being performed, when there have been multiple previous sternotomies, or when there is another indication for warfarin anticoagulation. For isolated autograft regurgitation without autograft dilatation (particularly if the autograft was placed within the native aorta at the time of the Ross procedure), isolated AVR may be performed. The choice between mechanical and biological prostheses is based on numerous considerations, including patient age and surgeon/patient preference. Mechanical prostheses have a greater role in patients who have undergone multiple prior operations and those who require concomitant replacement of other heart valves in whom a mechanical prosthesis is selected. In the presence of autograft dilatation (>4.5 cm), aortic root replacement with a mechanical or bioprosthetic composite graft with coronary reimplantation is performed. Valve-sparing root replacement can be considered in highly selected patients (those with isolated aortic root disease or minimal additional pathology requiring straightforward intervention). The ideal patient has an aortic root ≤5 cm, mild or less autograft valve regurgitation, and good left and right ventricular function29,30 without other concomitant valve or coronary problems.
The present study was a case series and is subject to all limitations inherent in this study design. Importantly, we cannot comment on the overall risk of the need for reoperation after the Ross procedure. However, the objective of the present study was to outline the nature of problems that may occur after the Ross procedure, not to define the incidence of reoperation. This study represents, to the best of our knowledge, the largest reported experience with reintervention (percutaneous and surgical) after the Ross procedure. Although this is a highly selected patient population representing those at 1 end of a spectrum, it is important for clinicians and patients interested in the Ross procedure to be aware of these issues when selecting an aortic prosthesis when AVR is necessary.
- Received July 27, 2009.
- Accepted July 12, 2010.
- © 2010 American Heart Association, Inc.
- Stelzer P,
- Jones DJ,
- Elkins RC
- Elkins RC,
- Knott-Craig CJ,
- Ward KE
- Luciani GB,
- Casali G,
- Favaro A,
- Prioli MA,
- Barozzi L,
- Santini F,
- Mazzucco A
- Bonow RO,
- Carabello BA,
- Chatterjee K,
- de Leon AC Jr.,
- Faxon DP,
- Freed MD,
- Gaasch WH,
- Lytle BW,
- Nishimura RA,
- O'Gara PT,
- O'Rourke RA,
- Otto CM,
- Shah PM,
- Shanewise JS
- Warnes CA,
- Williams RG,
- Bashore TM,
- Child JS,
- Connolly HM,
- Dearani JA,
- del Nido P,
- Fasules JW,
- Graham TP Jr.,
- Hijazi ZM,
- Hunt SA,
- King ME,
- Landzberg MJ,
- Miner PD,
- Radford MJ,
- Walsh EP,
- Webb GD
Although this manuscript deals with a topic that is surgical in nature, it also provides valuable information to cardiologists, because they are involved in the preoperative consultation and the long-term follow-up of patients being considered for the Ross procedure. The Ross procedure is performed for aortic valve disease and is offered to patients in all age groups: infants, children, and adults. The operation is palliative, and the need for reintervention (percutaneous or surgical) over a lifetime is inevitable for the majority of patients. This study represents the largest reported series of patients undergoing reoperation after the Ross procedure to date. Although literature is available that quantifies the incidence of reoperation after the Ross procedure, there are no published reports that address the qualitative nature and risks associated with reoperation when it is required. This manuscript provides that “missing” information so the clinician can properly counsel patients not only about the potential need for reoperation but also about the nature of the reoperation when it is necessary.