Ross Procedure and Left Ventricular Mass Regression
Background— Return of left ventricular mass to normal is considered to be a favorable result of aortic valve replacement. The Ross procedure provides near normal hemodynamics and thus allows studies of left ventricular (LV) reverse remodeling. LV mass regression may be influenced by surgical technique (subcoronary [SC] versus root replacement [RR]).
Methods and Results— Data from the German Ross Registry were analyzed. A total of 646 patients (mean age: 43.6±12.7 years, range: 16 to 71 years; SC technique n=295, RR technique n=351) underwent a Ross procedure in 7 participating centers from 1990 to 2004. The patients underwent preoperative and postoperative echocardiographic evaluations. Mean follow-up time was 3.5±2.5 years (range 0.12 to 13.7 years). Follow-up completeness was 97%. The LV mass index (LVMI) decreased significantly during follow-up in both groups (SC: 209±53 preoperatively to 154±48 at 1-year follow-up, [P<0.01 versus preoperative values] to 149±51g/m2 at 2-year follow-up, [P=NS 1-year versus 2-year follow-up] versus RR: from 195±56 preoperatively to 144±51 at 1-year follow-up [P<0.01 versus preoperative values] to 140±49g/m2 [P=NS 1-year versus 2-year follow-up]). LVMI regression remained stagnant 1 year after the Ross procedure in most patients in both groups. On the basis of multivariate analysis, predictors for incomplete LVMI regression after the autograft procedure were high preoperative LVMI, smoking, and uncontrolled diastolic hypertension.
Conclusions— At mid-term echocardiographic follow-up, patients of both groups had favorable autograft hemodynamics. Risk factors for incomplete postoperative LVMI regression in our study were smoking and persistent diastolic hypertension. This emphasizes the importance of cessation of smoking and treatment of arterial hypertension, even in younger patients, after corrected aortic valve disease.
The use of the pulmonary autograft for aortic valve replacement was first described in 1967 by Dr Ross.1 Originally introduced as subcoronary intra-aortic implantation of the autograft, the Ross procedure was later technically modified.2,3 The most widely used technique of the Ross operation today is full root replacement (RR) with or without annular support. There is renewed interest in the subcoronary (SC) insertion of autografts, however.4
Several potential advantages of the subcoronary technique are obvious. This technique retains the native aortic wall and the aortic sinuses, which should theoretically preserve aortic root and leaflet dynamics5 and decrease the incidence of aortic root dilatation and consecutively progressive aortic regurgitation.6,7 Subcoronary insertion also eliminates the need for reimplantation of the coronary arteries, thus avoiding the danger of malalignment of coronary arteries.
The disadvantages of the subcoronary technique include prolonged cardiopulmonary bypass time and aortic cross clamp time because of the increased complexity of the procedure compared with the freestanding root technique. The intra-aortic placement of the pulmonary autograft has the potential to reduce the effective valve orifice area and therefore increase the transvalvular gradient. Mismatch of the aortic root and pulmonary autograft makes the subcoronary insertion especially challenging. Imprecise valve insertion may lead to geometric distortion of the autograft and secondary autograft regurgitation.8 Some of these factors potentially influence left ventricular (LV) afterload and mass regression.
Only a few studies exist that have evaluated left ventricular reverse remodeling after the Ross procedure. To the best of our knowledge, no study has investigated the impact of autograft implantation technique (subcoronary versus root replacement) on postoperative left ventricular mass regression after the Ross procedure. The excellent hemodynamic performance of the pulmonary autograft in the aortic position with normal pressure gradients at rest9 and exercise10 and trivial or no regurgitation in the majority of patients has been described before. This complete relief of hemodynamic burden allows studies of the determinants for regression of LV mass without the impact of residual obstruction and/or regurgitation seen with other commonly used valve substitutes, such as mechanical prostheses or stented bioprostheses. The aim of the present study was to analyze predictors of complete LV mass regression after the Ross procedure using 2 different techniques (subcoronary versus root replacement) compared with controls.
Inclusion criteria for the present study were patient age over 15 years and autograft implantation using the subcoronary technique or root replacement. Patients who underwent the cylinder inclusion technique were excluded. A total of 125 patients were excluded because of age (n=77) or operative technique (n=48). These relatively small patient groups (compared with the subcoronary and root replacement groups) were excluded to reduce the heterogeneity of the baseline data of the study groups.
Data from the German Ross Registry database were analyzed. Participation in the registry is voluntary and at present includes patient data from 7 departments of cardiac surgery in Germany: The Sana Herzchirurgische Klinik Stuttgart (SC technique, n=0; RR technique, n=334), the University Hospitals of Frankfurt (SC technique, n=7; RR technique, n=0), Jena (SC technique, n=0; RR technique, n=4), Lübeck (SC technique, n=276; RR technique, n=0) and Tübingen (pediatric patients only), and the German Heart Centers of Berlin (SC technique, n=6; RR technique, n=13) and Munich (SC technique, n=6; RR technique, n=0). The Ross operation was performed according to the following selection criteria: Young age, patient’s objection or medical contraindication to coumadin (eg, because of occupational hazards), being of childbearing age (women), and participation in competitive sports. The surgical technique was determined by the primary surgeon at each center. The surgical policies were based on personal experience. The 2 main contributing centers (610 of the total 646 patients) were the University Hospital of Lübeck and the Sana Herzchirurgische Klinik, Stuttgart. In Lübeck, the implanting surgeon has a preference for the subcoronary technique to avoid the risk of aortic root aneurysm formation. In Stuttgart, the surgeons perform the Ross procedure using the most common technique worldwide, the root replacement technique, which reduces the procedure length and complexity.
A total of 646 patients met the inclusion criteria. Preoperative data are listed in Table 1. The patients groups (SC group, n=295; RR group, n=351) were comparable regarding sex, age, body surface area, New York Heart Association functional class, and most associated conditions. The number of patients with pure aortic stenosis was significantly higher in the RR group, whereas the numbers of smokers and of patients with impaired renal function and arterial hypertension were significantly greater in the SC group. The Ross procedures were performed between 1990 and 2004. All patients were evaluated clinically and echocardiographically at annual intervals by each institution. The mean follow-up time was 3.2±2.2 years in the SC group and 3.9±2.7 years in the RR group (P<0.001). Follow-up completeness for clinical and echocardiographic variables was 97%.
Intraoperative data are summarized in Table 2. All Ross procedures were performed using standard cardiopulmonary bypass with mild to moderate hypothermia (33° to 26°C nasopharyngeal temperature) or normothermia (exclusively RR group). Blood cardioplegia was administered selectively every 20 minutes. The pulmonary autograft was dissected in a scalloped fashion, leaving only a 2-mm rim of right ventricular muscle in both groups.
After a J-shaped aortic incision into the noncoronary sinus, the proximal anastomosis of the autograft to the annulus was performed with multiple interrupted 4-0 polyfilament sutures. Three running 5-0 prolene sutures were used for the distal anastomosis of the autograft to the left coronary, right coronary, and noncoronary sinus. Further details of the surgical technique have been previously described.4
The proximal anastomosis of the autograft to the aortic root was performed with 3 running 3-0 prolene sutures. An external Dacron strip for annular support was added in all patients after 1998. A running prolene suture was also used for the distal anastomosis. The technical details have been described elsewhere.3
In all patients of both groups, a cryopreserved pulmonary homograft (mean diameter 25.5±2.3 mm, range 19 to 33 mm) was used for right ventricular outflow tract reconstruction. The distal and proximal anastomoses were completed with running monofilament sutures during cross clamp time.
Concomitant procedures are listed in Table 2. In both groups, any dilated ascending aorta was largely reduced or replaced with a prosthetic Dacron tube to normalize the aortic dimension at the sinotubular junction. The number of ascending aorta replacements was significantly greater in the RR group, whereas more left ventricular outflow tract myectomies and mitral valve repairs were performed in the SC group.
Informed consent was obtained before echocardiography and all investigative procedures were performed according to institutional guidelines. At each center, transthoracic echocardiography using 2.5-MHz transducers was performed by an experienced echocardiographer during follow-up visits. The dimensions of the left ventricular outflow tract and the aortic root at different levels (annulus, sinus of Valsalva, sinotubular junction) were measured from parasternal long-axis views using the inner wall distances. Doppler measurements of transvalvular velocities for the autograft were obtained from apical standard views. Systolic pressure gradient (Δp) was calculated according to the modified Bernoulli equation (ΔP=4×v2, where v indicates maximal Doppler velocity), and effective orifice area (EOA) was calculated according to the continuity equation. Autograft regurgitation was graded by color-flow Doppler using the method of Perry and coauthors.11 LV mass (LVM) was calculated by the formula described by Devereux and Reichek.12 LVM was indexed to the patient’s body surface area. To assess the completeness of LVM index (LVMI) regression, we compared the postoperative LVM of patients to the LVM values of 10 normotensive control subjects without history of cardiac disease and with a mean age and body surface area similar to patients in the study group.
Continuous data are expressed in terms of the mean and standard deviation and were compared between the SC and RR groups using the 2-sample Student t test. Ordinal data such as aortic valve regurgitation grade were compared between the 2 operative techniques using the χ2 test. Fisher exact test was used to evaluate differences in simple proportions, as the number of events or patients with a specific characteristic was often rather small. Variables for LVM regression were tested in multivariate analysis to identify independent predictors of incomplete LVMI regression at 1-year follow-up postoperatively using multiple stepwise logistic regression (backward selection method). A cutoff point >160 g/m2 defined values outside of the normal range. Odds ratios and 95% confidence intervals were calculated for significant multivariate predictors. Two-tailed values of P<0.05 were used as the criterion for statistical significance. Analysis of the data was performed using the SPSS statistical package (version 12.0, SPSS Inc). Power analysis indicated that the numbers of patients who underwent SC implantation and full RR provided over 80% power (α = 0.05, β = 0.20) to detect a difference of 10% in the change in LVMI between the operative techniques, assuming a variability of 20% for each group (effect size =0.5) using the Student t test (version 4.0, nQuery Advisor, Statistical Solutions).
Table 3 summarizes patient outcomes. Causes of early death in the SC group were refractory arrhythmias in 1 patient and thrombotic occlusion of the left main coronary artery in another patient presenting with acute endocarditis. In the RR group, the only early death was due to low cardiac output. There were a total of 10 late deaths. Two patients died of cardiac causes during follow-up (homograft endocarditis and pulmonary bleeding and refractory ventricular arrhythmias, both in the RR group). A total of 3 sudden late deaths (SC group n=1; RR group n=2) remained unexplained, and 5 late deaths were due to noncardiac causes during follow-up. Mortality was related to cancer (RR group: n=1 gastrointestinal; SC group: n=1 larynx, n=1 gastrointestinal, n=1 pulmonary) and aortic dissection (RR group n=1). Reasons for autograft reoperation were a tear in the autograft during the early postoperative period (RR group n=1), regurgitation due to dilatation (RR group, n=4), endocarditis (RR group n=2, SC group n=1), a tethered cusp (RR group n=1), subalvular aneurysmatic autograft malformation (RR group n=1), and leaflet prolapse (RR group n=2, SC group n=3). In 4 patients undergoing the RR technique, the autograft failure was treated by valve reconstruction. Indications for homograft reoperation were homograft obstruction (n=5), combined lesion (n=1) or regurgitation (n=2) of clinical significance.
All data from echocardiographic examinations are listed in Table 4. The maximum gradient across the autograft at rest was low in both groups (SC group, 6.8±3.8 mm Hg, range 2 to 26 mm Hg, versus RR group, 5.5±3.7 mm Hg, range 1 to 33 mm Hg; P<0.001). The mean grade of postoperative autograft regurgitation was also low in both groups; however, it was higher in the RR group (0.8±0.5) compared with the SC group (0.3±0.4). The number of patients with autograft regurgitation grade 2 or higher was in fact larger in the SC group (n=9; 3.1%) than in the RR group (n=4;1.1%; P=NS). Only 1 patient (SC group) had an autograft regurgitation of grade ≥3.
The postoperative dimensions of the aortic annulus, sinus of Valsalva, and sinotubular junction were significantly larger in the RR group. The EOA index (EOAI) increased significantly after the Ross procedure in both groups. In the RR group, however, a significantly larger EOAI was found at follow-up compared with the SC group (2.20±0.82 cm2/m2 versus 1.60±0.57, respectively; P<0.05). The end-diastolic diameter of the left ventricle and thickness of the interventricular septum and posterior wall decreased significantly during follow-up in both groups. Although the preoperative posterior wall thickness was significantly larger in the SC group than in the RR group (1.34±0.26 versus 1.19±0.25 cm), the postoperative percent change of posterior wall thickness was not significantly different between the groups (SC decrease to 1.18±0.17 [−22%] versus RR 1.05±0.19 [−22%]). No significant change of global left ventricular function could be demonstrated after the Ross procedure in either group (Table 4).
LVMI decreased significantly from 209±53 and 195±56 g/m2 preoperatively to 154±48 (P<0.01) and 144±51 g/m2 (P<0.01) at 1-year follow-up in the SC and RR groups, respectively (Figure). Between the 1-year and 2-year follow-up, there was a much smaller decrease (P>0.05) in LVMI to 149±51 (SC group) and 140±49 g/m2 (RR group). The LVMI in healthy controls was 130±30 g/m2. The LVMI was not significantly different from the controls in 72% of the SC group and 73% of the RR group at the 1-year follow-up (P=NS).
Completeness of LV Mass Regression
Normotensive healthy controls in our study had a LVMI of 130±30 g/m2. The cutoff point of mean plus 1 standard deviation (160 g/m2) was applied to define values outside the normal range. Applying univariate analysis, the following predictors of incomplete LVM reduction were found: Male gender, preoperative aortic stenosis, higher preoperative LVMI, higher postoperative systolic blood pressure, higher postoperative diastolic blood pressure, and higher preoperative LV ejection fraction (Table 5). Only preoperative LVMI, postoperative diastolic blood pressure, and smoking were significant predictors of incomplete LVMI regression according to the multivariate analysis. Operative technique and autograft and homograft pressure gradients were not predictors of incomplete LVM regression on the basis of univariate and multivariate analyses.
Cardiac causes of LV hypertrophy include LV outflow tract obstruction, valvular and supravalvular aortic stenosis, and aortic regurgitation. LVM regression after aortic valve replacement is considered to be proportional to valvular hemodynamics13 and has been shown to have an important impact on long-term patient survival.14 All prosthetic valves are known to have smaller EOAs than the normal native aortic valve. The relative stenosis induced by these valves (especially in sizes ≤21 mm) imposes a chronic afterload stress on the myocardium.
The pulmonary autograft is an attractive valve substitute for aortic valve replacement. Many studies have demonstrated excellent hemodynamic performance of the autograft after aortic implantation.8,10 However, very little, if any, information is available on LV reverse remodeling after the Ross procedure in a large series of patients that compares autograft implantation techniques. Because of a near physiological aortic valve flow profile, from a hemodynamic point of view, the Ross procedure can be regarded as a reference in LVM regression and LV geometric restoration studies. The present study is based on this assumption, and the primary goal of this study was to assess postoperative LV reverse remodeling and LVM regression after pulmonary autograft aortic valve replacement. Two surgical techniques were compared (subcoronary versus root replacement).
Statistical analysis revealed that patients in the SC group were subjected to significantly longer cardiopulmonary bypass and aortic cross clamp times relative to RR. This is related to the higher technical complexity of the SC insertion technique of the pulmonary autograft. This did not affect early mortality, however. In both groups, 30-day mortality was low (<1%). This is probably secondary to the fact that adequate myocardial protection was achieved by intermittent cold blood cardioplegia. Thus, the myocardial injury at the time of the procedure had no impact on intermediate postoperative global ventricular function. We also found no evidence of late deterioration of global LV function. The analysis of LVMI regression postoperatively did not indicate any correlation with prolonged intraoperative ischemic times. Late mortality was low (1.4% SC versus 1.8% RR) in both groups.
Independent of implantation technique, the peak autograft gradients were within the physiological range (well below 10 mm Hg). The EOA was higher in the RR group than in the SC group. Even in the SC group, however, the indexed EOA dimension (1.60±0.57 cm2/m2) was significantly higher than 0.85 cm2/m2. This value is largely believed to represent the threshold for prosthesis size-patient mismatch.15,16
The mean grade of autograft regurgitation and the incidence of clinically relevant autograft regurgitation were low in both groups. Our finding is different from several reports demonstrating that patients with a larger aortic annulus are more prone to clinically relevant autograft regurgitation after the Ross operation.2,17 This might be related to the fact that in the majority of patients in the RR group, the aortic annulus was effectively supported by an external Dacron strip. The large EOAs might explain the good clinical results in both groups and the excellent exercise capacity of patients after the Ross procedure.10,16
The major intention of this study was to evaluate LVMI change after the Ross procedure. Independent of surgical technique, we found a significant decrease in LVMI 1 year after the Ross procedure (SC −26.4% versus RR −26.2%; follow-up versus preoperative values, P<0.001), but not from the first to second year after the procedure (SC −3.3% versus RR −3.8%). The decrease of LVMI was paralleled by significant reductions in LV end-diastolic dimensions and interventricular septum and posterior wall thickness. It has been well described before that most of the regression of LV hypertrophy occurs early after aortic valve replacement. Further LV reverse remodeling occurs in the second year after surgery, although these changes did not reach statistical significance in the present study.
Our results of percent change in LVMI are well in keeping with findings from a study by Schmid et al,18 who evaluated LV remodeling after pulmonary autograft aortic valve replacement by echocardiography and magnetic resonance imaging. They found a 21% decrease in LVMI 12 months postoperatively in 27 patients using the RR technique for autograft insertion.
Although our study only included patients over 15 years of age, excellent LVM regression can also be expected in pediatric patients after the Ross procedure. Brown et al19 described LV dimensions and mass indexes after the Ross procedure in children (age range 1 month to 18 years). A 35% regression of LVMI was observed 3 years after the procedure.
Comparison to Other Aortic Valve Substitutes
Many reports on LVM regression after aortic valve replacement with stentless and stented bioprostheses or mechanical valves have been published.20–23 In most studies, a significant regression of LVM is reported. According to these reports, LV mass index decreased by 17% to 25% 1 year after prosthetic aortic valve replacement. Some studies on prosthetic aortic valve replacement showed less decrease in LVM, whereas the results of others are comparable to those of our study. There are studies showing that LVM regression is dependent on valve type. Pibarot et al21 concluded in their study of LVM after bioprosthetic aortic valve replacement that the superior hemodynamic performance of stentless bioprostheses might have some benefits with regard to LVM regression. Our study demonstrated that the pulmonary autograft in the aortic position offers physiological pressure gradients that are superior even to the hemodynamic performance of stentless xenografts. The significance of these benefits in terms of prognosis, however, remains to be determined. Because of the near physiological postoperative aortic valve performance, the Ross procedure allows studies of LVM regression after aortic valve replacement. The near physiological hemodynamics may establish the autograft valve as a reference valve substitute for LVM regression studies after aortic valve replacement.
Completeness of LV Mass Index Regression
Normalization of LVMI after surgery and risk factors for incomplete LVMI regression were studied by univariate and multivariate analysis. Normal LVMI was defined by a control group using the same echocardiographic technique. At 1-year follow-up, LVMI normalized (LVMI ≤160 g/m2) in the majority of patients in both groups. In a univariate analysis, male sex, preoperative LV ejection fraction, preoperative aortic stenosis, and arterial blood pressure (systolic and diastolic) were found to be risk factors for incomplete LVMI regression (LVMI >160 g/m2). In the multivariate model, independent predictors for incomplete LVMI reduction were preoperative LVMI, smoking, and higher postoperative diastolic blood pressure. This is probably one of the most interesting findings of this study. Although we are able to demonstrate normal pressure gradients across the autograft valve with a low incidence of relevant autograft regurgitation in virtually all patients, there is a significant number of patients with incomplete LV regression. As suggested by multivariate analysis, this may be related to extra-cardiac causes of LV hypertrophy, ie, diastolic arterial hypertension. At follow-up, 187 (29%) of all patients who had undergone the Ross procedure were found to be hypertensive. Preoperatively, there were a total of 181 patients (SC group, n=100, RR group, n=81) with the diagnosis of arterial hypertension. Practically all of these patients were taking antihypertensive drugs. Approximately 72 patients (40%), however, had ineffective control of arterial hypertension, whereas the remaining 60% were normotensive at last clinical follow-up. In addition, a significant number of patients (n=115) without the preoperative “label” arterial hypertension were also found to be hypertensive at clinical follow-up.
In patients with a diastolic arterial blood pressure <80 mm Hg at follow-up, the LVMI regressed to 139±41 g/m2, whereas in patients with a diastolic pressure >100 mm Hg, the LVMI only regressed to 153±41 g/m2 (P<0.05). This underlines the importance of strict hypertension control after the Ross procedure. Although the ideal for the patient would be a drug-free life after the Ross operation, this may have an adverse effect on LVM regression in a significant proportion of patients. Because LVM reduction is closely related to long-term outcome, frequent monitoring of arterial blood pressure and intensive treatment of hypertension seem to be of utmost importance. In patients who have undergone the RR technique, this should also decrease the strain of the autograft and may prevent or postpone relevant autograft dilatation.
This observational study had a retrospective design. The study compared the Ross procedure results from experienced heart centers in Germany that voluntarily submitted data to the German Ross Registry. The observed event rate for mortality and morbidity was small and therefore did not allow definitive statistical analysis of risk factors. For bias-free comparison of the 2 surgical techniques, a randomized, prospective trial would be necessary. This would require surgeons with experience in both subcoronary implantation and full root replacement.
In summary, both the subcoronary technique and the root replacement technique resulted in excellent outcomes with low mortality, morbidity, and reoperation rates. The major difference between the groups was a larger autograft dimension after full root replacement despite annular support. The regression of left ventricular mass was excellent in most patients and independent of autograft implantation technique. Predictors for incomplete LVM regression were smoking and diastolic arterial hypertension. This emphasizes the importance of cessation of smoking and effective treatment of arterial hypertension, even in younger patients, after corrected aortic valve disease.
Registry Participants: Drs Lange and Hörer at the German Heart Center in Munich; Drs Moritz and Simon at the Johann-Wolfgang-Goethe-University Hospital in Frankfurt; Drs Wahlers, Franke, and Madershabian at the Friedrich-Schiller-University Hospital in Jena; Drs Hetzer and Hübler at the German Heart Center in Berlin; and Drs Ziemer, Walker, and Kosan at the Eberhard-Karls-University Hospital in Tübingen.
The authors thank Katrin Meyer and Siegurd Küster for the excellent data management at the Registry Site in the Department of Cardiac Surgery, University Hospitals Schleswig-Holstein, Campus Lübeck. The manuscript was written on behalf of the German Ross Registry. The authors thank all participants.
Presented in part at the 75th Scientific Sessions of the American Heart Association, New Orleans, La, November 9, 2004.
Elkins RC. The Ross operation: a 12-year experience. Ann Thorac Surg. 1999; 68: S14–S18.
Simon-Kupilik N, Bialy J, Moidl R, Moidl R, Kasimir MT, Mittlbock M, Seebacher G, Wolner E, Simon P. Dilatation of the autograft root after the Ross operation. Eur J Cardiothorac Surg. 2002; 21: 470–473.
Devereux RB, Reichek N. Echocardiographic determination of left ventricular mass in man: anatomic validation of the method. Circulation. 1977; 55: 613–618.
De Paulis R, Sommariva L, Colagrande L, De Matteis GM, Fratini S, Tomai F, Bassano C, Penta de Peppo A, Chiariello L. Regression of left ventricular hypertrophy after aortic valve replacement for aortic stenosis with different valve substitutes. J Thorac Cardiovasc Surg. 1998; 116: 590–598.
Koren MJ, Devereux RB, Casale PN, Savage DD, Laragh JH. Relation of left ventricular mass and geometry to morbidity and mortality in uncomplicated essential hypertension. Ann Intern Med. 1990; 114: 345–352.
Gaudino M, Glieca F, Luciani N, Cellini C, Morelli M, Girola F, Guarini G, Possati G. Left ventricular mass regression after aortic valve replacement for aortic stenosis: time course and determinants. J Heart Valve Dis. 2004; 13: S55–S58.
Monrad ES, Hess OM, Murakami T, Nonogi H, Corin WJ, Krayenbuehl HP. Time course of regression of left ventricular hypertrophy after aortic valve replacement. Circulation. 1988; 77: 1345–1355.