A Critical Reappraisal of the Ross Operation
Renaissance of the Subcoronary Implantation Technique?
Background— The autograft procedure, an option in aortic valve replacement, has undergone technical evolution. A considerable debate about the most favorable surgical technique in the Ross operation is still ongoing. Originally described as a subcoronary implant, the full root replacement technique is now the most commonly used technique to perform the Ross principle.
Methods and Results— Between June of 1994 and June of 2005, the original subcoronary autograft technique was performed in 347 patients. Preoperative, perioperative, and follow-up data were collected and analyzed. Mean patient age at implantation was 44±13 years (range 14 to 71 years; 273 male, 74 female). Bicuspid valve morphology was present in 67%. The underlying valve disease was aortic regurgitation in 111 patients, stenosis in 46 patients, combined lesion in 188 patients, and active endocarditis in 22 patients (in 2 patients without stenosis or regurgitation). Concomitant procedures were performed in 130 patients. Clinical and echocardiographic follow-up visits were obtained annually (mean follow up 3.9±2.7 years, 1324 patient-years; completeness of follow-up 99.4%). The in-hospital mortality rate was 0.6% (n =2), and the late mortality was 1.7% (n=6), with 5 noncardiac deaths (4 cancer, 1 multiorgan failure after noncardiac surgery) and 1 cardiac death (sudden death). At last follow-up, 94% of the surviving patients were in New York Heart Association class I. Ross procedure–related valvular reoperations were necessary in 9 patients: Three received autograft explants, 5 received homograft explants, and 1 received a combined auto- and homograft explant. At last follow-up visit, autograft/homograft regurgitation grade II was present in 5/10 patients and grade III in 4/0. Maximum/mean pressure gradients were 7.4±6.2/3.7±2.1 mm Hg across the autograft and 15.3±9.4/7.6±5.0 mm Hg across the right ventricular outflow tract, respectively. Aortic root dilatation was not observed. Freedom from any valve-related intervention was 95% at 8 years (95% confidence interval 91% to 99%).
Conclusion— Midterm follow-up of autograft procedures according to the original Ross subcoronary approach proves excellent clinical and hemodynamic results, with no considerable reoperation rates. Revival of the original subcoronary Ross operation should be taken into account when considering the best way to install the Ross principle.
The autograft procedure for aortic valve replacement, also known as the Ross procedure, has been performed since 1967.1 After having been practiced worldwide for over 30 years, the Ross procedure has reached maturity.2 Advantages of this therapeutic option are the use of the patients’ own valves with favorable hemodynamic characteristics, low risk of endocarditis, avoidance of anticoagulant therapy, low thrombogenicity, and the potential to grow in children.3–5 Factors contributing to a limited acceptance are the complexity of the operation and the necessity of replacing both the aortic and pulmonary valves. In addition, little clinical long-term information is available regarding the durability of the autograft in the aortic position and the durability of the substitute of the right ventricular outflow tract. Several national and international reports are available focusing on mortality, morbidity, valve-related complications, and valvular hemodynamics. Although clear good midterm follow-up data of a considerable number of patients have been published, the overall acceptance of this surgical option in aortic valve disease is low.
The original Ross procedure involved a subcoronary replacement of the aortic valve, a surgical approach that was nearly pushed into the background over time. The present study defines the midterm results of a large consecutive monocenter series with the original subcoronary Ross operation in patients with aortic valve disease.
From June 1994 through June 2005, aortic valve replacement according to the original Ross subcoronary pulmonary autograft procedure was performed in 347 patients, with a mean age of 44.0±12.7 years (range 14 to 71 years).6–7 Within the last 5 years, our center performed at an average 395 aortic valve interventions annually (artificial valves in 13%, bioprostheses in 70%, valve reconstructions in 6%, and autograft procedures in 11%). The preoperative characteristics of the patients are reported in Table 1. The general indications for the autograft procedure were in most cases isolated aortic valve disease with the patient’s preference to avoid oral anticoagulation, contraindications for oral anticoagulation, child-bearing potential in women, and very active lifestyle (sports, profession). Exclusion criteria for the autograft procedure were severe calcifications of the aortic root, significantly reduced left ventricular systolic function, more than 2-vessel coronary artery disease, apparent connective tissue disease, and anatomical or structural defects of the pulmonary valve. Clinical and echocardiographic follow-up was performed in a prospective manner annually and was 99.4% complete. The last inquiry was in August of 2005. Informed consent was obtained from all patients.
All operations were performed by 1 surgeon (H.H.S.) under standard cardiopulmonary bypass conditions with the use of moderate systemic hypothermia (26°C nasopharyngeal temperature). In the first 5 years, crystalloid cardioplegia was used, and thereafter cold blood cardioplegia was used. The autograft was trimmed at the base in a scalloped fashion to match the proximal suture line with the remnants of the leaflet attachments of the patient’s excised valve as close as possible. The proximal anastomosis was performed with single 4–0 polyfilament and the distal anastomosis with running 5–0 monofilament sutures after excising part of the left and right coronary sinus to keep the suture lines upstream of the coronary ostia. Additionally, a 5–0 monofilament U-suture was used to secure the commissures at the wall of the recipient valve. The noncoronary sinus was left intact. Care was taken to implant the autograft while preserving its geometric relatives with respect to the distance of the commissures and the height of the leaflets. To adjust for normal diameters, reconstruction of a dilated ascending aorta by lateral plication over a prosthetic felt strip or replacement with a prosthetic tube was performed with regard to the extent of dilatation and the macroscopic appearance of the wall. The diameter of the ascending aorta, the patient’s size (body surface area), and the underlying valve pathology were the main determinants of whether to perform ascending aorta interventions.8 Reduction annuloplasties with Dacron or a pericardial strip were performed when the annulus diameter was more than 30 mm to neutralize the size mismatch between the autograft and the aortic valve ring. For reconstruction of the right ventricular outflow tract, a cryopreserved pulmonary homograft was used in 321 patients, a decellularized pulmonary homograft (SynerGraft; CryoLife, Inc; Atlanta, Ga) in 25 patients, and a porcine xenograft in 1 patient. Operative data including concomitant procedures are listed in Table 2.
Data Acquisition and Echocardiographic Measurements
Informed consent was obtained before each follow-up visit, and these visits were performed on an outpatient basis 1, 3, 6, and 12 months after the procedure and then annually thereafter. All postoperative events were defined according to the Guidelines for Reporting Morbidity and Mortality After Cardiac Valvular Operations published in 1996 by the ad hoc Committee for Standardizing Definitions of Prosthetic Heart Valve Morbidity of The American Association for Thoracic Surgery and The Society of Thoracic Surgeons.9 Details of the echocardiographic evaluation have been reported previously.10
The combined results of the collected data were analyzed using SPSS 11.0 for Windows (SPSS Inc, Chicago, Ill) and Minitab 13.1 (State College, Pa). Continuous data were expressed as the mean±Sd; categorical variables were presented as absolute numbers and percentage. Estimation for long-term survival and freedom from morbid events were made using the Kaplan-Meier method. The survival time of a patient started at the time of surgery and ended at death (event) or at last follow-up (censoring). The long-term survival characteristics of the patient cohort were compared with survival probabilities of the general population obtained from German Life Tables 2002/2004 (Statistisches Bundesamt, Wiesbaden, Germany). The analysis of autograft and pulmonary homograft survival rates started at the time of implantation and ended with reoperation or reintervention (event) or last follow-up or patient death (censoring). Univariate and multivariate analyses with backward logistic regression were used to study potential determinants of aortic valve incompetence grade I or more during follow-up. The following variables were considered: number of annulus and ascending aorta interventions; diameters of the aortic annulus, the sinus of Valsalva, the sinotubular junction, and the ascending aorta; and hypertension. The development of mean pulmonary homograft gradient 10 mm Hg or greater during follow-up was also investigated. For the analysis, the following variables were considered: diameter of the homograft, age of the donor and recipient, ABO blood group mismatch, right ventricular outflow tract adjustment with a pericardial patch and removal of ventricular myocardium, and decellularized homograft.
The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written.
Follow-up completeness was 99.4% for all hospital survivors, with 97% completeness of echocardiographic follow-up. The 2 patients whose follow-up was incomplete were censored at the time of their last follow-up. Mean patient age at the time of the last inquiry was 47.7±13.4 years (range 14.9 to 80.2 years). Mean follow-up duration was 3.8±2.7 years (range 0.1 to 10.5 years), with a total follow-up of 1 324 patient-years.
The 30-day mortality rate was 0.6% (2 patients). One patient died from refractory ventricular arrhythmias 3 days after the operative intervention. Another patient was operated on for acute aortic valve endocarditis with severe aortic regurgitation and hemodynamic compromise. One day before discharge, he suffered from thrombembolic occlusion of the left main coronary artery with refractory cardiogenic shock after reoperation (inspection of the autograft showed no apparent pathological changes).
There were 6 later deaths (>30 days; Figure 1). A cardiac origin of a fatal course must be assumed in 1 patient who suffered from sudden, unexplained, unexpected death without further clinical data or autopsy. A noncardiac cause of death was present in 5 patients (1 multiorgan failure after noncardiac surgery and 4 malignancies). Cumulative overall survival was 99% at 1 month (95% confidence interval [CI] 98 to 100), 99% at 1 year (95% CI 97 to 100), 97% at 5 years (95% CI 94 to 99), and 94% at 8 years (95% CI 88 to 99). Comparison of observed fatal events of the study group with expected numbers of deaths in the general German population showed a total number of 8 observed death (expected in the general population 8.8), with 6 male (expected 7.2) and 2 female (expected 1.6). The subcohort with New York Heart Association (NYHA) class I/II had 3 deaths (expected 7.2) and 5 NYHA class III/IV deaths (expected 1.6). The group with patient age <50 years revealed 3 observed deaths (expected 1.5), and there were 5 deaths (expected 7.3) in patients 50 years or older.
Structural Valve Deterioration
Structural valve deterioration was present in 9 patients. In 7 patients, a decline from functional class I to II could be verified, and in 2 patients there was a decline from class II to III. In contrast to the clinical presentation, only 4 patients revealed an impairment of valvular function on echocardiography (n =2 aortic regurgitation grade II and III, n =2 maximal pressure gradient across the pulmonary homograft). Without changes in NYHA functional capacity, 7 patients showed an increase in aortic regurgitation of at least 1 grade, and 8 patients showed an increase of the maximal pressure gradient across the homograft of a least 20 mm Hg. In 9 patients, structural valve deterioration could be verified by reoperation (see Reoperations section).
No autograft valve thrombosis was detected. In one case of homograft stenosis requiring reintervention (14 months after the primary procedure), obstruction was caused by extensive leaflet-adherent thrombi in 2 sinuses.
A total of 13 neurological events (1.0% per patient-year) occurred; 4 were permanent, 1 was a reversible ischemic neurological deficit, and 8 were transient ischemic attacks. Completed stroke occurred in 2 patients and transient ischemic attack in 6 within 30 days postoperatively. During follow-up, 2 completed strokes were observed 5.2 and 6.2 years postoperatively. Both strokes were associated with a new onset of atrial fibrillation.
Major internal or external bleeding occurred in 1 patient (0.08% per patient-year). The patient suffered 2.4 years postoperatively from epidural hematoma after head injury while taking the oral anticoagulant therapy that is necessary in chronic atrial fibrillation. Eleven patients were treated with oral anticoagulants: 5 for chronic atrial fibrillation, 3 for embolic events of vascular or cardiac origin, 2 for deep vein thrombosis, and 1 during the postoperative course after surgery of chronic peripheral occlusive artery disease.
In 1 patient, reoperation was performed because of severe aortic regurgitation after healed endocarditis. One patient had acute autograft endocarditis that required reoperation 2 months after the initial procedure. In 2 cases, medical therapy of autograft endocarditis was successful 3.3 and 4.3 years postoperatively. Homograft endocarditis occurred in 4 cases. Acute homograft endocarditis with septic pulmonary emboli required reoperation in 1 patient 2.2 years after the Ross procedure. The patient was successfully treated with implantation of a new homograft. Another 3 patients had reoperations of the homograft with intraoperative confirmation of cured endocarditis. The linearized rate for 8 valve endocarditis events was 0.62% per patient-year.
Other Cardiac Events
Three patients suffered from acute myocardial infarction. One patient died from acute thrombotic occlusion of the left main stem 1 week after the initial operation. One patient with posterior myocardial infarction (5.6 years after the initial intervention) was successfully treated with a percutaneous intervention. One case with subacute aortic regurgitation and left ventricular failure 3 years after the Ross procedure developed non—ST-elevation myocardial infarction and had coronary artery bypass grafting during the redo procedure.
Reoperation on the pulmonary autograft or the pulmonary homograft was required in 9 patients. The time interval between the initial procedure and the reoperation ranged from 2 to 85 months (mean 24.8±24.4 months; median 16 months). Three patients underwent reoperation on the pulmonary autograft, 1 patient on the autograft and pulmonary homograft, and 5 patients on the pulmonary homograft. The indication for replacement of the 4 pulmonary autografts was regurgitation of the neoaortic valve in all patients due to endocarditis (n =2) and leaflet prolapse (n =2). Aneurysmatic changes of the aortic root or the ascending aorta were not present in any case. All procedures to replace the pulmonary autograft were performed from 2 to 36 months (mean 17.3±14.0 months; median 15.5 months) after the initial operation. Among the 6 patients who required replacement of the pulmonary homograft, 2 had pure stenosis (maximal gradients 59 and 41 mm Hg, respectively), 3 had pure regurgitation (grade III in 2; grade IV in 1), and 1 had combined stenosis and regurgitation in the clinical setting of acute homograft endocarditis. The procedures were performed from 13 to 85 months (mean 28.2±28.2 months; median 15.5 months) after the initial Ross operation. All patients survived reoperation and were all alive at the date of the last follow-up. Freedom from any autograft and homograft related intervention was 100% at 1 month, 99% (95% CI 99 to 100) at 1 year, 97% (95% CI 95 to 99) at 5 years, and 95% (95% CI 91 to 99) at 8 years postoperatively (Figure 2). The event-free survival rate with freedom from death, reoperation, thromboembolism, and endocarditis was 99% (95% CI 98 to 100) at 1 month, 98% (95% CI 97 to 100) at 1 year, 91% (95% CI 88 to 95) at 5 years, and 84% (95% CI 77 to 92) at 8 years postoperatively (Figure 3).
At the last follow-up visit, 337 patients (97.1%) were alive with their pulmonary autograft in place (8 patients died and 2 were lost to follow-up). Of those, 315 (94%) had no cardiac symptoms and were in New York Heart Association functional class I, 15 (4%) were in functional class II, and 3 (1%) were in class III.
Autograft and Homograft Function at Last Follow-Up
Results are displayed subdivided in follow-up groups <5 years and >5 years in Table 3.
Univariate and Multivariate Analyses of Valve Dysfunction
Univariately, there is significant evidence of an increase in time to aortic regurgitation grade I or more with increasing diameters of the aortic sinus (P=0.041) and the sinotubular junction (P=0.014). In a multivariate Cox model, both variables are not independently predictive. These 2 variables are highly correlated (r=0.77). It appears from its probability value that the diameter of the sinotubular junction is the more predictive variable.
Univariate analysis indicated that patient age (P<0.0001), number of right ventricular outflow tract adjustment interventions (P=0.03), and the implantation of decellularized homografts (SynerGraft, P=0.002) were predictive of late homograft stenosis with a mean gradient of 10 mm Hg or more. Multivariately, all 3 variables were independent risk factors associated with at least a moderate homograft obstruction.
Our 11-year experience with the original subcoronary implantation technique clearly confirms previous reports documenting the safety of the Ross operation in selected adults with a low prevalence of valve-related complications.5,11,12 For example, when comparing the annual stroke rate of the study group receiving the Ross procedure with the incidence rates of stroke from a community-based study of stroke in Germany, the event rate of 1.57 events/year per 1000 patients was similar to the expected event rate from the community study (annual incidence rate 1.82 events/year per 1000 patients).13 The survival rate is excellent, and rate of reoperation of the autograft in the present subcoronary implantation cohort is low, similar to other operative techniques in the midterm follow-up.11,14 The mortality rate of patients undergoing the Ross procedure was slightly better than was expected using the German Life Table figures. With the small number of fatal events in the Ross group, it will be difficult to detect genuine effects of the operative procedure. Therefore, large-scale clinical conclusions cannot be drawn from these data.
Autograft failure necessitating reoperation occurred in 4 patients with severe aortic regurgitation. There was no autograft stenosis. Another 4 patients with echocardiographic evidence of moderate to severe aortic regurgitation were carefully followed, because they probably will require aortic valve replacement in the near future. No risk factors indicating autograft valve failure could be derived from this series. Diameters of the aortic sinus and the sinotubular ridge may play a role in the long term, but multivariate factor analysis at present failed to identify these dimensions as independent risk factors.
In the long term, autograft function and reoperations on the autograft may depend on the maintenance of the anatomy, structural integrity, and intact 3-dimensional geometry of the aortic root. Therefore, operative techniques and their impact on aortic root dimensions will play a predominant role in the further development of autograft dysfunction with the need of reoperation.14
Almost universally, the procedure is now being performed as a root replacement. There was evidence and conviction that the initial inaugurated subcoronary implantation technique has higher early and midterm failure rates when adopted and practiced by many surgeons, each with an individual learning curve.15 The long-term fate of the pulmonary autograft when used as a free-standing root or inclusion cylinder is largely unknown, particularly in adult patients. A progressive dilatation of the native aorta adjacent to the pulmonary autograft (anastomosis site) may result in neoaortic regurgitation. The clinical and hemodynamic results of the present consecutive series with subcoronary implants do not support the initial misgivings that this complex technique led to a less consistently competent autograft valve compared with the root replacement technique.16 The main reason for this might be the fact that the present series is a 1 center—1 surgeon experience, with the advantage of only 1 learning curve. Aortic root dimensions remained stable over time and no progressive autograft incompetence was detected during intensive surveillance over the years. There is growing evidence that remodeling of the full root autograft is a morphological prerequisite for a progressive deterioration in valvular hemodynamics.17–19 However, the exact magnitude of this phenomenon and its clinical consequences are still controversial, and some surgeons perform additional stabilizing techniques with support of the aortic annulus and the sinotubular junction, with good clinical and echocardiographic results in the midterm.11,14,16,19,20 The long-term effects of these surgical measures have to be determined in the future. Although attention has been primarily dedicated to the annular and sinus level, the study by Luciani et al17 clearly showed that diameters at the sinotubular ridge and proximal ascending aorta tend to equalize with the sinus of Valsalva, thereby realizing a distally pronounced root remodeling process. On the contrary, in patients undergoing surgery with root-sparing techniques (subcoronary implantation, root inclusion technique) a significant decrease in diameter in the sinus, the sinotubular ridge, and ascending aorta was identified (reverse remodeling17). The data of the present study support the concept of reverse remodeling shown by Luciani et al.17 To maintain physiological root geometry or to support a reverse remodeling process, a substantial subset of patients were treated with aortic valve annuloplasty and ascending aorta interventions in our cohort (ascending aorta replacement with vascular grafts, reductive aortoplasty). These adjunctive procedures contribute considerably to the significant decrease of the diameters within the aortic root and the ascending aorta. These changes in geometry are unlikely related to the intra-aortic position of the autograft valve. Whether this translates into a hemodynamic and clinical benefit has to be determined in the long term.
On the basis of the results of the present series, a major concern is the threat of infective endocarditis causing valve dysfunction with the need of reoperation. Considering both the aortic and the pulmonary valve as possible sites of infection, we observed 6 cases of endocarditis (2 acute, 4 cured) in 9 reoperation procedures. Long-term studies have to define the clinical impact of this important finding and clarify the number of unreported cases. In the meantime, liberal strategies regarding prophylaxis of infective endocarditis are recommended.
An additional important contributing factor might be structural changes in the wall of the autograft and in the valve leaflets. Focal interruption of the media of the vessel wall with total absence of elastin fibers and loss of smooth muscle cells has been reported in patients with root replacement and progressive root dilatation.14,21,22 It has been speculated that these changes may be common in all pulmonary autograft roots that are exposed to systemic pressure for a longer period of time. These findings could represent a preexisting abnormality (eg, in bicuspid valves), or could also be a transmural ischemic injury to the wall of the pulmonary autograft root caused by division of the vasa vasorum.23 The long-term implications of these findings are unknown, especially of the combined effect of ischemic injury and long-term systemic pressure load.
Right ventricular outflow tract reconstruction was routinely done with a cryopreserved homograft. In 25 patients, a cryopreserved decellularized homograft was used (SynerGraft). The implantation of decellularized homografts was associated with a shorter time until mean homograft gradient 10 mm Hg or more was observed, indicating no clear benefit of this new device in clinical practice. Acute infective endocarditis of the homograft was the indication of reoperation in 1 patient, and structural valve failure of the homograft required reintervention in 4 patients (cured endocarditis in 3 patients). Another 26 patients currently have remarkable mean gradients (> 15 mm Hg) over the right ventricular outflow tract, indicating patients at risk for further hemodynamic deterioration. Pulsed Doppler studies indicated that the obstruction was located directly at the homograft annular level and not at the tubular part or the distal anastomosis. On the basis of computed tomographic angiographic studies,24 we are inclined to think that a circular shrinkage process will cause obstruction at the valvular level. Therefore, we have changed our policy and are going to remove all muscular parts adjacent of the cryopreserved homograft. For further stabilization, a Gore Tex (W.L. Gore & Associates, Newark, Del) or pericardial strip is inserted between the recipient right ventricle and the homograft (called right ventricular outflow tract adjustment). With the use of right ventricular outflow tract adjustment interventions, a small but evident benefit was achieved to reduce the hazard of homograft obstruction.
A Plea for the Subcoronary Technique
In our view, there is a clear need for a randomized controlled trial comparing the subcoronary and root techniques. On reviewing the available reports in literature, 2 facts emerge. First, subcoronary valve transfer does not give rise to postoperative root dilatation with progressive valve regurgitation, provided there is no malinsertion of the valve and the aortic root is normal or treated with additional interventions. Second, when we are dealing with a root replacement, many reports about potential and real problems are available, all of which may influence one in favor of the subcoronary mode.6,7 We submit that subcoronary surgery, by retaining the patient’s own aortic wall, the natural sinuses, and coronary artery egress, means that only the cusps are transferred. This approach is a technical challenge to the surgeon that can be best overcome by adhering to the described guidelines.6 The exceptional good survival and favorable hemodynamic results of the present series are probably influenced by our patient selection criteria, the absence of concomitant cardiac morbidities (eg, coronary artery disease), and unimpaired overall left ventricular function. The usefulness of the Ross operation as a root replacement especially in children is proven, and many of these patients are not candidates for an intra-aortic Ross operation. However, there are possible drawbacks of the full root technique, primarily the potential long-term dilatation of the root especially if no reinforcement techniques are used at the sites of the anastomoses. Possible limitations of the use of the subcoronary implantation technique are the expertise of the surgeon, extensive root calcification, especially calcification around the coronary ostia, and a marked distortion of the annulus.
In conclusion, although the freestanding autograft root is clearly well established with good midterm results (especially in children), we firmly believe that the subcoronary or cylinder transfers merit widespread clinical and scientific reappraisal. Our present experience with a large consecutive series with subcoronary implants has confirmed the suitability and safety of this approach with low mortality and morbidity rates and good functional results. It must be pointed out that in the present study, long-term follow-up observations are limited and echocardiographic data suggest a mild increase in the incidence of autograft valve regurgitation. At midterm, this is of no considerable clinical impact.
The authors thank Katrin Meyer for her excellent technical assistance and data management at the Registry Site at the Department of Cardiac Surgery, University Hospital Schleswig-Holstein, Campus Luebeck.
Presented at the American Heart Association Scientific Sessions, Dallas, Tex, November 13–16, 2005.
Chambers JC, Somerville J, Stone S, Ross DN. Pulmonary autograft procedure for aortic valve disease. Long-term results of the pioneer series. Circulation. 1997; 96: 2206–2214.
Simon P, Aschauer C, Moidl R, Marx M, Keznickl FP, Eigenbauer E, Wolner E, Wollenek G. Growth of the pulmonary autograft after the Ross operation in childhood. Eur J Cardiothorac Surg. 2001; 19: 118–121.
Ross DN. The subcoronary technique versus the root technique for autograft surgery. J Heart Valve Dis. 2002; 12: 553–558.
Duebener LF, Stierle U, Erasmi A, Bechtel MF, Zurakowski D, Boehm JO, Botha CA, Hemmer W, Rein JG, Sievers HH. Ross procedure and left ventricular mass regression. Circulation. 2005; 112 (supp I): I-415–I-422.
Kolominsky-Rabas PL, Sarti C, Heuschmann PU, Graf Ch, Siemonse S, Neundoerfer B, Katalinic A, Lang E, Gassmann KG, Ritter von Stockert T. A prospective community-based study of stroke in Germany: The Erlangen Stroke Project (ESPro): incidence and case fatality at 1, 3, and 12 months. Stroke. 1998; 29: 2501–2506.
Luciani GB, Casali G, Favaro A, Prioli MA, Barozzi L, Santini F, Mazzucco A. Fate of the aortic root late after Ross operation. Circulation. 2003; 108 (suppl II): II-61–II-67.
Takkenberg JJ, Dossche KME, Hazekamp MG, Nijveld, A, Jansen EWL, Waterbolk TW, Bogers AJJC; the Dutch Ross Registry Study Group. Report of the Dutch experience with the Ross procedure in 343 patients. Eur J Cardiothor Surg. 2002; 22: 70–77.
Elkins RC. Invited commentary. Ann Thorac Surg. 1999; 67: 553–554.
Bechtel JFM, Gellissen J, Erasmi AW, Petersen M, Hiob M, Stierle U, Sievers HH. Mid-term findings on echocardiography and computed tomography after RVOT-reconstruction: comparison of decellularized (SynerGraft) and conventional homografts. Eur J Cardiothor Surg. 2005; 27: 410–415.