Autograft Reinforcement to Preserve Autograft Function After the Ross Procedure
A Report From the German-Dutch Ross Registry
Background— Autograft reinforcement interventions (R) during the Ross procedure are intended to preserve autograft function and improve durability. The aim of this study is to evaluate this hypothesis.
Methods and Results— 1335 adult patients (mean age:43.5±12.0 years) underwent a Ross procedure (subcoronary, SC, n=637; root replacement, Root, n=698). 592 patients received R of the annulus, sinotubular junction, or both. Regular clinical and echocardiographic follow-up was performed (mean:6.09±3.97, range:0.01 to 19.2 years). Longitudinal assessment of autograft function with time was performed using multilevel modeling techniques. The Root without R (Root−R) group was associated with a 6× increased reoperation rate compared to Root with R (Root+R), SC with R (SC+R), and without R (SC-R; 12.9% versus 2.3% versus 2.5%.versus 2.6%, respectively; P<0.001). SC and Root groups had similar rate of aortic regurgitation (AR) development over time. Root+R patients had no progression of AR, whereas Root−R had 6 times higher AR development compared to Root+R. In SC, R had no remarkable effect on the annual AR progression. The SC technique was associated with lower rates of autograft dilatation at all levels of the aortic root compared to the Root techniques. R did not influence autograft dilatation rates in the Root group.
Conclusions— For the time period of the study surgical autograft stabilization techniques preserve autograft function and result in significantly lower reoperation rates. The nonreinforced Root was associated with significant adverse outcome. Therefore, surgical stabilization of the autograft is advisable to preserve long-term autograft function, especially in the Root Ross procedure.
The Ross operation is an alternative to conventional aortic valve replacement in selected patients. It can be performed with low mortality, and provides excellent hemodynamics and low rates of thromboembolism, avoiding the need for anticoagulation therapy and its consequences.1–5 After a renewal of interest in this procedure in the early 90s, long-term results of these procedures are beginning to emerge, and it is well established that autograft function may deteriorate over time eventually requiring replacement.1–8 Several teams have used autograft reinforcement interventions (R) to treat underlying abnormalities and to stabilize the components of the aortic root8–11 in an attempt to prevent autograft failure. These reports include small number of patients and provide short follow-up, which in combination with the low incidence of autograft failure itself may be insufficient to evaluate the effects of such additional procedures.
Using data from the adult population of the German-Dutch Ross Registry, we sought to investigate the effect of R on postoperative outcome, the need for reoperation, and the longitudinal autograft performance over time.
Patients and Methods
Study Population and Operative Data
Data from the German-Dutch Ross Registry were analyzed. The registry includes data from 12 departments of cardiac surgery since 1988. Follow-up data from each center were entered in the database and a systematic prospective registry was started in January 2002 (Clinical trial ID NCT 00708409). The used surgical technique was according to the surgeon’s preference, with more or less each center having adopted the one or the other technique. The operative technique (SC/Root) was specific for each institution and remained the same throughout the time period of the study. The vast majority of SC procedures were performed in one center. In the Root technique, one center performed no reinforcement procedures at all, whereas the incidence of prophylactically performed reinforcement procedures increased with time in all other centers performing the Root Ross procedure. Thirty patient operated with the root inclusion technique were included in the subcoronary group (SC), to create a group with all native root preserving procedures. A total of 1335 patients were entered in the registry as of January 2008. Patients’ preoperative characteristics as well as operative technique (subcoronary, SC; root replacement, Root) and presence (+R) or absence (−R) of R are summarized in Tables 1 and 2⇓, respectively.
Indications and contraindications for the Ross procedure have been described in detail elsewhere.4,9,12 As R was regarded any additional procedure performed at the aortic annulus, sinotubular junction, or both. Usually, at the level of the annulus, a 4-mm wide strip of pericardium, Dacron, or a 2/0 GoreTexR suture was placed between donor and recipient tissues to stabilize or to prevent dilatation. In the Root group, R with mainly a Dacron strip, was used in almost all patients in the last 8 years. In the SC group, as R, a 2/0 GoreTexR suture was incorporated in the annulus suture line, if the annulus diameter exceeded 28 to 30 mm as measured before autograft implantation. In the Root group, autograft R consisted also of an additional second suture line fixating circumferentially the remnants of the wall of the native aortic root to the autograft, 4 mm distal to the proximal suture line. In both groups (Root, SC), R of the sinotubular junction was performed by suturing a Dacron prosthesis directly distal to the commissures, if an ascending aorta replacement was indicated.
Informed consent was obtained from all patients. The study was approved by the local ethics committee. All authors had full access to and take full responsibility for the integrity of the data.
Clinical and Echocardiographic Follow-Up
Clinical and echocardiographic follow-up was performed at discharge and on a yearly basis. The standardized echocardiographic data acquisition protocol of the registry has been published elsewhere.3 Autograft dimensions were measured at 3 levels (annulus, sinus of Valsalva, sinutubular junction) as described by Roman et al.13 Aortic regurgitation (AR) was graded on a scale from 0 to 4 according to Perry et al.14 Mean duration of follow-up was 6.09±3.97 years (median 5.6 years; range 0.01 to 19.2 years; 8205 patient-years). Follow-up completeness was 93%. The 7% missing follow up visits were evenly distributed across the groups. Classification of the mode of valve failure has been performed according to the latest guidelines for reporting outcome after valve interventions.15 All indications for autograft reoperations were in accordance with the ACC/AHA guidelines.16 In 2 patients subvalvular aortic aneurysms were the primary indication for reoperation.
Frequencies are given as absolute numbers and percentages. Continuous data are expressed as mean±SD. Patients were classified according to the operative technique (SC, Root) and the presence (+R) or absence (−R) of R. Comparisons between the groups were performed using the Mann-Whitney U test and the Fisher exact test. Actuarial estimates of survival and freedom from autograft reoperation were accomplished with Kaplan-Meier methods. Survival curves were compared using the log-rank test (SPSS 11.0 for Windows, SPSS, Inc). The Cox model was used to assess the consistency of treatment effect by testing for interactions between the type of surgery (technique and presence of autograft reinforcement) and prespecified baseline characteristics. To identify predictive variables for shorter time to autograft reoperation, we first performed a univariate analyses by using the Cox proportional hazard regression model. Multivariable Cox proportional hazard models were used to confirm whether differences between the operative groups persisted in the presence of preoperative variables. The presence of interactions and the proportionality of hazards assumption was checked for the final model, including operative group and significant preoperative variables. The following factors were analyzed as potential risk factors for autograft reoperation attributable to structural and nonstructural failure (infective endocarditis as a reoperation indication in 8 patients was excluded): age, sex, year of surgery, predominant aortic hemodynamics, hypertension, previous aortic valve intervention, presence of bicuspid aortic valve, technique, and presence of reinforcement procedures.
In accordance with the new guidelines,15 autograft performance over time was analyzed and reported with the use of longitudinal modeling as previously described.3,15,17 The echocardiographic data were analyzed by using a multi-level linear model (MLWin 2.0, Centre for Multilevel Modeling). Various regression models were tested on the study’s dataset, and the linear model provided an appropriate fit for the study. This model provides a linear regression line with an intercept and slope for each individual patient, and it estimates the mean intercept and slope across patients. The intercept and slope are assumed to vary randomly for the different patients. The intercept (±SE) corresponds to the notional value at the time of surgery, the slope (±SE) represents the annual progression of these measurements. Because this is a multicenter study, it reflects the daily practice of the Registry sites, nevertheless the uniformity of the preoperative data are not warranted and may have an influence in the statistical evaluation of the results. In an attempt to neutralize this center-specific influence, we integrated a center variable, allowing for the effect which the different centers may have to the results of this study, in all statistical analyses performed. This model was applied to analyze AR and aortic root dimensions over time, as well as AR as a function of aortic root dimensions for the surgical subgroups Root and SC and subgroups with and without R. The intercept and slopes provided represent the mean values across the population or subgroups throughout the period of the study and should not be extrapolated beyond this. For the small subgroup of 30 patients with root inclusion technique, separate estimation of the AR development and AR dimensions over time was also performed.
Clinical Outcome (>30 Days) and Autograft Reoperations
Table 3 provides information on the observed mortality and reoperations. Overall cumulative survival was 94.6% (95% CI 92.8 to 96.4%) at 10 years, freedom from autograft reoperation (with the exclusion of 8 patients operated for infective endocarditis15) was 96.8% (95% CI 95.5 to 99.0%) at 5 and 89.6% (95% CI 86.1 to 93.0%) at 10 years. When allowing for technique and presence of R, the SC and the Root+R revealed a significantly better freedom from reoperation at 10 years in comparison with Root−R (94.2% [95% CI 90.4 to 97.9%] and 93.2% [95% CI 88.2 to 98.2%] versus 88.3% [95% CI 76.5 to 90.1%], respectively, P=0.001; Figure 1). Autograft reoperation rates for structural (26.9% of all reoperations) and nonstructural valve failure (61.2% of all reoperations) were significantly higher in Root−R in comparison with Root+R, SC-R, SC+R (12.5% versus 2.3%, 2.6%, 2.5%, respectively, P<0.0001). The Root−R group accounted for 55.9% of all reoperations (Table 3). The instantaneous hazard for reoperation for all subgroups is displayed in Figure 2. The multivariable Cox proportional hazard model of the 4 operative groups showed strong evidence that patients operated with the Root technique without the use of reinforcement techniques tend to have shorter times to reoperation (Table 4). The effect of including a number of preoperative variables in the model is shown in Table 5. The differences between the operative groups persist in the presence of these variables. Details of the final model including operative group and the only significant preoperative variable, aortic regurgitation, are given in Table 6. There was no significant evidence of an interaction between the variables in the final model or that the proportionality assumptions were violated.
Development of Aortic Regurgitation Over Time
AR grade was found to develop approximately linearly with follow-up time. Based on 3803 measurements, the mean initial AR grade was 0.531 (±0.094) with an average increase of AR grade of 0.032 (±0.005) per year. There is significant evidence that AR increases with time (P<0.0001), but the amount of this increase is clinically not substantial.
In the current analysis, and allowing for a random center effect, no difference between the Root and SC groups could be observed in terms of initial AR grade (0.519±0.10 versus 0.543±0.101; P=0.71) and annual progression rate (0.035±0.007 versus 0.029±0.006; P=0.49). Allowing for annulus reinforcement, in the SC groups, SC+R had higher initial AR grade compared to SC-R (0.667±0.112 versus 0.491±0.100; P=0.0045), whereas no difference could be observed in the annual progression of AR (P=0.57, Figure 3a). In the Root subgroups, a higher initial AR grade in Root+R was observed (0.678±0.125 versus 0.471±0.116; P=0.031), however in the presence of annulus R, AR remained stable for the first decade, in contrast to Root−R, in which AR increased at 6-fold rate compared to Root+R (Root−R: 0.067±0.010 AR grade/yr versus Root+R: −0.013±0.012 AR grade/yr; P<0.001, Figure 3b). No significant differences between the SC and root inclusion technique could be observed in terms of initial AR grade and the annual increase of it. All models remained robust after adjusting for cofounding preoperative variables.
Changes of Autograft Dimensions Over Time
An appropriate regression model to study diameter changes at the level of the aortic annulus, sinus, and sinotubular junction with time was a linear model:
Diameter (time)=(Initial diameter±SE)+(Annual increase of diameter±SE)×time (yr).
Initial dimensions were comparable between Root and SC (25.18±1.05 mm versus 24.49±1.06 mm; P=0.26). The Root group dilated 3 times faster in the first decade (0.316±0.046 mm/yr versus 0.103±0.039 mm/yr, P<0.0004). Taking the presence of annulus R into consideration, no difference could be observed between the Root subgroups. In SC, the presence of R led to higher initial annulus diameter (24.30±0.657 mm versus 24.94±0.837 mm, P=0.008), the rate of annulus dilatation did not differ (Figure 4a).
Sinuses of Valsalva
Initial dimensions were comparable between Root and SC (33.81±1.09 mm versus 32.43±1.050 mm; P=0.10). During the first decade Root dilated 4 times faster than SC (0.259±0.063 mm versus 0.064±0.039 mm/yr respectively, P=0.008). No differences of the annual diameter increase within the subgroups of each technique could be observed when allowing for the presence of R (Figure 4b).
A tendency toward lower initial diameters were observed in SC (29.54±1.58 versus 31.11±1.59 mm; P=0.067), in the Root group the sinotubular junction (STJ) tended to dilate almost 3-fold faster than in SC during the first decade (0.602±0.058 mm/yr versus 0.219±0.047 mm/yr, P<0.0001). When allowing for STJ R, no differences between the subgroups of each technique (with or without STJ reinforcement) could be observed.
No significant differences between SC and root inclusion technique could be observed in terms of autograft dilatation over time. All models remained robust after adjusting for cofounding preoperative variables.
Change of Dimensions and Autograft Regurgitation
The Root technique resulted in wider range of annulus and STJ diameters and a trend toward a higher slope of AR development with increasing diameters compared to the SC technique. AR development with increasing annulus or STJ diameters was lower in SC, which, together with the narrower range and lower slope of diameters in SC, makes this technique more robust against AR development with increasing annulus or STJ diameters. (Figure 5a and 5b).
The Ross procedure can be performed as an attractive alternative in selected patients, however various groups present mid- to long-term results indicating that the autograft function may deteriorate over time with the hazard of eventually mandating a reoperation.1–3,6
Significant research has been conducted regarding the mode of autograft failure after the Ross procedure. Early autograft failure is often attributed to technical errors, as was the case with the technically demanding and difficult to reproduce subcoronary technique.18 The introduction of the root replacement technique19 seems to ameliorate this early autograft failure, however reports of progressive autograft dilatation3,7,20,21 and subsequent late autograft failure have recently emerged.2,7
Understanding the modes of autograft failure after the Ross procedure, many groups have used modified techniques or R to correct abnormalities in the aortic root area and thus prevent anatomic mismatch,21,22 or to stabilize parts of the aortic annulus prone to dilatation.8–11 The long-term impact of R on the autograft function and durability remains largely unknown. Thus main focus of the present study was to unveil the effect of such R on autograft function. The presence of 2 different techniques in the Ross Registry (SC and Root) presents a challenge for this analysis, mainly because the evolution of the native aortic root pathology hosting a subcoronary implant is very different than the evolution of the freestanding pulmonary autograft root technique, and as such, reinforcement techniques might play different roles and serve different purposes in each of these techniques.
R in this study were performed to correct anatomic abnormalities or mismatch, or prophylactically to stabilize the aortic root and prevent postoperative autograft dilatation. The intention of the operating surgeon and the indication for performing R was determined by the operating surgeon at each institution.
Our main observation was that a significant proportion of the Root−R subgroup required reoperation in comparison to the 3 other subgroups. The Root−R group contributed 57% of all reoperations observed in the registry. The leading cause of reoperation in the Root group was nonstructural valve deterioration15 (87% of all reoperations in this group) presenting in the form of autograft dilatation, whereas in the SC group 63% of all reoperations were attributed to structural valve deterioration,15 mainly as cusp prolapse. The addition of R seemed to decrease reoperation rates because of nonstructural valve deterioration in the Root+R group leading to reoperation rates similar to the SC technique. In the SC group, we could not show a significant impact of R on reoperation rates attributable to either structural or nonstructural valve deterioration.
Autograft failure appears after the first 6 to 8 years in the Root−R group with an exponentially rising instantaneous hazard rate, while remaining stable throughout the observational period in the SC group. For the time period studied, the Root+R subgroup had similar reoperation risk rates as the SC group. This finding is in concordance with previous studies.23 From our data it could be hypothesized that the larger the preoperative annulus dimensions, the lower the ability of the autograft to provide adequate leaflet coaptation with progressive root dilatation. In the Root group, R leads to smaller annulus diameters, and thus they may act prophylactically. It is however unknown whether in the long term reinforcement procedures prevent or postpone autograft function deterioration.
In this study we could not observe a significant difference with regards to AR development over time between the SC and the Root as overall groups. This is in contrast to a previous report of ours,3 where we found a significantly increased initial AR in the Root patients. This difference, however, could be attributed to the additional 321 patients (32% more than the 2006 report3) added to the registry in the last 2 years. Moreover, because of the nonuniformity of the preoperative data, in this analysis we allowed for a center effect to mitigate any systematic reporting error between the registry sites. The presence of R effectively prevented AR development over time in the Root group, whereas in the SC R had no effect on the AR increase, albeit for a greater initial AR. R in SC was implemented only in large annulus observed at operation to reduce or reshape the effective annular size to improve cusp coaptation. Here, the indication for R was not prophylactically but therapeutic.
A consistent finding in this study is the larger initial postoperative aortic root diameters in patients undergoing R in both groups (Figure 4a and 4b), although this does not always reach the level of statistical significance. Given, however, that R are most likely to reduce the aortic root diameters, one can argue that R are often performed to treat underlying abnormalities, thus reducing aortic root dimension to restore the ideal anatomic relations. Although in our series R had a positive effect in terms of autograft durability, there are in the literature notable series of patients operated with the Root technique without R with excellent long-term outcome.1,5 The effects of proper patient selection bias, however, cannot be ruled out because an often intraoperative selection of the appropriate patient pathology is common in the setting of the Ross procedure. It may well be that in patients with an ideal aortic root pathology, a Root−R technique may have excellent outcome. In addition, it cannot be ruled out that special modifications on individual patient basis in very experienced hands can prevent postoperative dilatation without the need for R with synthetic material.5
The SC technique was associated with significantly reduced rates of aortic root dilatation at all levels of the aortic root. In the Root technique, R did not influence the progressive autograft dilatation over time for the time period of the study. In this study we could observe an increased AR in patients with increased annulus and STJ diameters. Although a causal effect could not be established, this could be explained by the observational nature of this study and the increased reserve of the aortic root in terms of dilatation24 that would lead to AR, and as such the time frame of this study could be insufficient to show that the AR increase is solely caused by root dilatation.
The present study is a retrospective nonrandomized study. The intention of the surgeon when performing R in SC was primarily to treat an underlying pathology, whereas in Root, R was mainly applied routinely as a part of root replacement. Early postoperative z values are provided only for the aortic annulus, mainly because of the fact that large databases of normal values in the adult population do not exist for the other counterparts of the aortic root components. No technique to support the sinus of Valsalva was performed in this patient population. A further post-hoc subgroup analysis to identify the most appropriate type of reinforcement material and or specific operative techniques was regarded statistically inappropriate because of the retrospective nature of this study. We believe that this should be performed in the setting of a prospective randomized trial. A small surgical subgroup operated with the root inclusion technique was included in the SC group. Subanalysis of key items (AR, autograft dimensions, reoperations) did not reveal any difference between the SC and the root inclusion technique group. A possible limitation may be the different follow-up times of the various study groups, with R in the Root, being mostly implemented in the last 8 years, having the shortest follow-up time. However the differences observed in outcome and autograft function were statistically and clinically significant for the time period studied.
We can conclude that in patients undergoing the Ross procedure, autograft reinforcement procedures performed either prophylactically to prevent autograft dilatation, or therapeutically to correct an underlying a suboptimal anatomy, lead to lower development of AR over time. Surgical autograft reinforcement is able to reduce reoperation rates for autograft failure because of nonstructural valve deterioration in the root replacement Ross procedure for the time period of this study. These procedures appear to be safe, present with good long-term outcome, and should strongly be taken into consideration.
The authors thank Katrin Meyer for her excellent data management and secretarial support at the Registry Site in the Department of Cardiac and Thoracic Vascular Surgery, University Clinics Schleswig-Holstein, Campus Lübeck. We thank Petra Lingens and Ilse Beyer for their secretarial support as well as Jana Engelmann and Anja Paap for their documentation support at the Department of Cardiac and Thoracic Vascular Surgery, University Clinics Schleswig-Holstein, Campus Lübeck.
C.A.B. serves as medical director of the South Africa Homograft Bank.
E.I.C. and T.H. contributed equally to this study.
Presented in part at American Heart Association Scientific Sessions 2008, November 8–12, 2008, New Orleans, La.
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.
Hanke T, Stierle U, Boehm JO, Botha CA, Matthias Bechtel JF, Erasmi A, Misfeld M, Hemmer W, Rein JG, Robinson DR, Lange R, Horer J, Moritz A, Ozaslan F, Wahlers T, Franke UF, Hetzer R, Hubler M, Ziemer G, Graf B, Ross DN, Sievers HH. Autograft regurgitation and aortic root dimensions after the Ross procedure: the German Ross Registry experience. Circulation. 2007; 116: I251–I258.
Klieverik LM, Takkenberg JJ, Bekkers JA, Roos-Hesselink JW, Witsenburg M, Bogers AJ. The Ross operation: a Trojan horse? Eur Heart J. 2007; 28: 1993–2000.
Akins CW, Miller DC, Turina MI, Kouchoukos NT, Blackstone EH, Grunkemeier GL, Takkenberg JJ, David TE, Butchart EG, Adams DH, Shahian DM, Hagl S, Mayer JE, Lytle BW. Guidelines for reporting mortality and morbidity after cardiac valve interventions. J Thorac Cardiovasc Surg. 2008; 135: 732–738.
Bonow RO, Carabello BA, Kanu C, 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, Smith SC Jr, Jacobs AK, Adams CD, Anderson JL, Antman EM, Fuster V, Halperin JL, Hiratzka LF, Hunt SA, Nishimura R, Page RL, Riegel B. ACC/AHA 2006 guidelines for the management of patients with valvular heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (writing committee to revise the 1998 Guidelines for the Management of Patients With Valvular Heart Disease): developed in collaboration with the Society of Cardiovascular Anesthesiologists: endorsed by the Society for Cardiovascular Angiography and Interventions and the Society of Thoracic Surgeons. Circulation. 2006; 114: e84–e231.
Svensson G, Aljassim O, Svensson SE, Bech-Hanssen O, Kjellman U. Anatomical mismatch of the pulmonary autograft in the aortic root may be the cause of early aortic insufficiency after the Ross procedure. Eur J Cardiothorac Surg. 2002; 21: 1049–1054.
Notzold A, Scharfschwerdt M, Thiede L, Huppe M, Sievers HH. In-vitro study on the relationship between progressive sinotubular junction dilatation and aortic regurgitation for several stentless aortic valve substitutes. Eur J Cardiothorac Surg. 2005; 27: 90–93.