Ross Procedure for Aortic Regurgitation versus Stenosis in Adults With and Without Autograft Support

Abstract

OBJECTIVES

The Ross procedure for aortic regurgitation (AR) and abnormal aortic valve morphologies is associated with an increased risk of autograft dilatation. Autograft support may ameliorate this problem. We analysed the results for all haemodynamic lesions and the effect of autograft support.

METHODS

A retrospective analysis was conducted of patients who underwent a Ross procedure at Saarland University Medical Center between December 1995 and December 2023. Three hundred and fifteen patients underwent full-root replacement with or without autograft support. Twenty-three (7%) were younger than 18 years and were excluded. The cohort was divided into 3 groups: patients with aortic stenosis (AS), AR and combined disease (CD). End points included survival, freedom from reoperation and AR and aortic root dimensions; these were compared among the 3 groups. Median follow-up was 3.6 (range 0.01–26.6) years and 95% complete.

RESULTS

Overall, 292 adult patients [male 74%; mean age 39 years (SD: 10)] were analysed with (n = 209) or without autograft support (n = 83). Patients with AS (n = 79; 28%) were compared to those with AR (n = 77; 25%) and those with CD (n = 136; 50%). Valve morphology was unicuspid (n = 141; 48%), bicuspid (n = 109; 38%) or tricuspid (n = 42; 14%). Survival at 15 years was similar across the groups (AR 86%; AS 93%; CD 94%; P = 0.123). Freedom from autograft reoperation was 90% at 10 years (AR 80%; AS 95%; CD 92%; P = 0.009). With autograft support, it was 93% at 10 years (AR 90%; AS 93%; CD 95%; P = 0.179). Neither a unicuspid (hazard ratio 1.072; 95% confidence interval 0.34–3.43; P = 0.907) nor a bicuspid aortic valve (hazard ratio 0.102; 95% confidence interval 0.08–1.26; P = 0.102) was associated with reoperation.

CONCLUSIONS

Patients with AR and an unsupported root replacement do have an increased risk of reintervention, irrespective of aortic valve morphology. With autograft support, however, autograft stability is excellent, irrespective of the underlying lesion. Thus, the Ross procedure in its supported version can be offered to all haemodynamic types and valve morphologies.

CLINICAL REGISTRATION

CEP 203/19.

Open in new tabDownload slide

INTRODUCTION

Aortic valve-sparing operations are increasingly used for aortic regurgitation (AR) in younger adults [1]. When the valve cannot be repaired, it is commonly replaced using conventional valve substitutes. The shortcomings of prosthetic valve replacement in younger adults [2] have led to a renewed interest in the Ross procedure [3] because of low rates of valve-related complications, normal quality of life and excellent long-term survival [4, 5].

Concerns, however, have been raised about the Ross procedure for isolated AR, a bicuspid aortic valve (BAV) and a dilated aortic annulus [6–11]. These features, in particular, annular dilatation, are considered as risk factors for autograft failure [8–10, 12].

In root dilatation due to connective tissue disease [13, 14], common sense precludes a Ross procedure [15]. In conjunction with congenital aortic malformations [BAV, unicuspid aortic valve (UAV)], it is less clear whether the pulmonary artery is prone to histologic and functional changes similar to those of the aneurysmal ascending aorta [14–16], resembling cystic medial necrosis [14, 17]. However, aortic complications are infrequent in patients with BAV and aortic aneurysm [18], rendering autograft replacement a possibility. In fact, studies with autograft support have yielded acceptable durability [19, 20].

In studies comparing aortic stenosis (AS) and AR, root size progression was considerably more pronounced in patients with AR [7, 16, 20]. None of these studies, however, have investigated the impact of autograft stabilization [7–11, 16]. In effect, in most cases, neither root support nor blood pressure control protocols were applied. Despite a higher risk of reintervention in patients with AR, the late survival benefit with the Ross procedure in AS was preserved in patients with AR.

It has recently been shown that the probability of autograft dilatation and reintervention can be mitigated by autograft support [19, 20]. The purpose of this study was to assess mid- and long-term results of the Ross procedure in patients with AR compared to those with AS, also analysing the effect of autograft support.

PATIENTS AND METHODS

Ethics statement

The investigation was approved by the Saarland Regional Ethics Committee (CEP 203/19). Individual patient consent was waived for the analysis and publication in anonymized fashion.

Patients

We conducted a retrospective analysis of patients who underwent a Ross procedure at Saarland University Medical Center between December 1995 and December 2023. Five patients underwent the root-in-root technique and were excluded. We identified 315 patients who underwent full-root replacement with or without autograft support. Of these, 23 patients (7%) were younger than 18 years and were excluded.

The cohort was divided into 3 groups: patients with AS, AR and combined disease (CD) (Fig. 1). Combined aortic disease was defined as ≥grade 2 (mean aortic valve gradient ≥20 mmHg) and AR ≥grade 1 [1]. Patients were categorized according to their original valve lesion. End points included survival, freedom from reoperation and AR and aortic root dimensions; these were compared among the 3 groups.

Patients were categorized as ‘supported’ Ross if they received external sinus stabilization using the remnants of the native aortic wall. If the technique covered less than 70% of the autograft circumference, it was not considered ‘supported’ (i.e. in patients with endocarditis and root abscess).

Surgical technique

The full-root replacement technique was used as described previously [21]. The autograft was implanted in an intra-annular position using a continuous running suture (polypropylene 4–0).

If the sinotubular diameter exceeded 30–35 mm, the tubular ascending aorta was replaced with a Dacron graft [size according to body surface area (BSA); n = 78; 27%]. This approach was used throughout our practice.

In patients with a dilated annulus (>26 mm, n = 180), an external annuloplasty was added using a pericardial strip (n = 75; between 1998 and 2008) and an expanded polytetrafluoroethylene suture (Gore-Tex CV-0; W.L. Gore & Assoc., Munich, Germany, n = 105) thereafter [21]. This step was performed irrespective of the valve lesion. The expanded polytetrafluoroethylene suture was placed at the level of the basal plane [21] and tied around a Hegar dilator (21 or 23 mm according to BSA) placed into the autograft [21].

For autograft support, the 3 tongues of the native aortic root in the commissural area were fixed to the sinotubular suture outside the autograft (n = 209; Fig. 2) [21] by including them in the suture line of the distal autograft anastomosis. This technique has been used since 2001 irrespective of the valve lesion or morphology, unless there was insufficient aortic tissue, e.g. in endocarditis with a large root abscess. This technique results in coverage of more than 70% of the autograft circumference. All anastomoses, including the external support, were completed before resuming coronary circulation (Fig. 2). No synthetic material was used for support.

If the Ross procedure was performed as a reoperation, the same technique was used. In all instances, limited mobilization of the coronary arteries (<1 cm) was sufficient to allow for tension-free reinsertion into the autograft.

Blood pressure monitoring was implemented for 6 months postoperatively to a target systolic pressure <110 mmHg in all patients because the most relevant root progression occurs within the first postoperative months [22]. Patients were seen either in our outpatient clinic or by their referring cardiologists.

Follow-up

All patients were seen regularly by their referring cardiologists or in our clinic. Echocardiograms from our institution and from referring cardiologists were reviewed. All patients were followed prospectively (clinically and echocardiographically at discharge, 3 months, 1 year, and yearly thereafter). Systolic gradients were measured using continuous-wave Doppler. AR was determined using colour Doppler according to European guidelines [1].

Median and mean follow-up examinations were 3.6 (range 0.01–26.6) years and 6.5 years (SD 7.1). Clinical and echocardiographic follow-up examinations were 95% complete (1601 patient-years) at the time of study closure in December 2023.

Statistical analyses

Non-normally distributed variables are presented as median (interquartile range) and continuous variables are presented as mean ± standard deviation. They were compared using analysis of variance with post hoc tests and the Bonferroni correction for normally distributed data and the Kruskal–Wallis test for non-normally distributed data. Categorical variables are expressed as frequencies (%) and were compared using the χ² test or the Fisher test. Time-dependent data were assessed using the Kaplan–Meier method and the log-rank test. Survival and freedom from reoperation were calculated at 1, 5, 10 and 15 years. All tests were 2-sided, and P values of <0.05 were considered statistically significant for all analyses. Missing data were replaced using mean imputation.

Changes in autograft root dimensions were determined through a multivariable linear general estimated equation with repeated measurements and residuals clustered at the patient level. All model iterations included age, BSA, height, gender, support technique and haemodynamic lesion as control variables. Sinus diameters were analysed as absolute diameters. A root diameter >45 mm was considered dilated and an autograft failure. A Cox proportional hazard regression was performed to identify predictors for the development of recurrent AR, reoperation and death. For multivariable analysis, we selected variables based on clinical relevance. All models were adjusted for patient age and sex, and relevant comorbidities were coded as yes/no. The proportional hazards assumption for the regression model fit was tested for all events of interest and were verified using the Schoenfeld residuals.

Statistical analyses were performed using SPSS 28.0 (IBM Corp. Released 2021. IBM SPSS Statistics for Macintosh, Version 28.0; Armonk, NY, USA).

RESULTS

Patients

Overall, 292 adult patients were studied [74% male, mean age 39 years (SD: 10), Table 1]. The procedure was performed with (n = 209) or without autograft support (n = 83; Table 2). The cohort was divided into 3 groups: 77 (26%) patients underwent the procedure for AS; 79 (27%), for AR; and 136 (47%), for CD (Fig. 1). Valve morphology was unicuspid (n = 141; 48%), bicuspid (n = 109; 38%) or tricuspid (n = 42; 14%; Table 1).

Ninety-five (33%) patients had undergone ≥1 previous cardiac operation, mostly aortic valve repair (n = 78; 27%; Table 1). The distribution of UAV and BAV was similar among the groups (P = 0.578). Twenty-one patients had a tricuspid AV (TAV) and AR (7%); of these, 15 (71%) had active endocarditis. Group characteristics are summarized in Table 1.

Early results

A concomitant procedure was performed in 116 patients (40%), mostly ascending aortic replacement (n = 78, 27%; Table 2).

There were 2 early deaths (AR n = 1; AS n = 1). One patient underwent surgery following 3 previous root replacements and died from a sudden haemorrhage after 2 weeks. The second patient had a reduced ejection fraction preoperatively; he developed low cardiac output postoperatively with non-occlusive mesenteric ischaemia and died of multiorgan failure 5 days postoperatively.

There were no myocardial infarctions. Intraoperatively, 1 patient (CD) developed right ventricular dysfunction, which was treated by coronary artery bypass and a temporary ventricular assist device. He developed a thromboembolic stroke but recovered neurologically. Another patient (AS) required temporary left ventricular support for ventricular dysfunction. He was weaned successfully on postoperative day 5 and continues to do well.

Four patients (2%; AR n = 4) required a permanent pacemaker implant. None of these had conduction abnormalities preoperatively; active endocarditis with perivalvular root abscess was the indication for surgery in all (Table 2).

There were no early reoperations. Two patients had AR 2 at discharge. There was no difference among the 3 groups regarding perioperative complications (Table 2).

Late results

Ten patients died during the follow-up period (AR, n = 5; AS, n = 3; CD, n = 2); 5 died of cardiac causes (8 months–4 years postoperatively), 1 of which remains unknown (CD n = 1). The causes included pulmonary homograft endocarditis with intravenous drug abuse (n = 1), cardiac arrhythmia (n = 2) and death of uncontrollable coagulopathy (n = 1; presumed heparin-induced thrombocytopaenia) at autograft reoperation 8 months after the index operation (AR n = 1, unsupported).

Five patients (AS n = 2, CD n = 1, AR n = 2) died of non-cardiac causes (6 months–11 years postoperatively). These included sepsis from urinary tract infection (n = 1), pneumonia (n = 1), intravenous drug abuse (n = 2) and cancer (n = 1). Overall survival was 91% at 15 years (AS: 93%, AR: 86%, CD: 94%; P = 0.123; Fig. 3). Freedom from cardiac death was 96% at 15 years. Without endocarditis, survival at 15 years was 93% (AS: 92%, AR: 94%, CD: 94%; P = 0.94).

By Cox regression analysis, active endocarditis (P = 0.037) and the lack of autograft support (P = 0.038) were associated with death. Neither BAV (P = 0.111), UAV (P = 0.192) nor the haemodynamic lesion (P = 0.839) was associated with death. Data are provided in Supplementary Material, Table S1.

Reoperation

Any reoperation was necessary in 26 patients (right ventricular conduit, n = 12). Overall freedom from reoperation was 83% at 10 years and 80% at 15 years.

Fourteen patients (4.8%) underwent autograft reoperation between 0.3 and 21 years postoperatively (median: 7.8 years) as a replacement (n = 2; 14%) or a valve-sparing procedure (n = 12; 86%).The indications were dilatation +/− relevant AR (n = 7, 50%; AR n = 4, AS n = 1, CD n = 2), isolated relevant AR (n = 5,36%; AR n = 2, AS n = 1, CD n = 2) and active endocarditis (AR n = 2,14%). Freedom from autograft reoperation was 90% at 10 years and 86% at 15 years (AR: 71% vs AS: 95% vs CD 89%; P < 0.001; Fig. 4A). In patients with endocarditis, it was 78% at 10 years and 66% at 15 years. Excluding patients with endocarditis, it was 92% at 10 years and 88% at 15 years (AR 74%, AS 95%, CD 92%; P < 0.001).

Freedom from autograft reoperation at 10 years was 94% with and 81% without support; at 15 years, it was 94% with and 73% without support (P < 0.001; Fig. 5). Freedom from autograft reoperation at 10 years in patients with autograft support was 90% with AR, 93% with AS and 95% with CD (P = 0.179; Fig. 4B). At 10 years, it was 92% without and 89% with ascending aortic replacement (P = 0.182), and 89% without and 86% with annuloplasty (P = 0.718). Interestingly, 11 out of 32 patients with endocarditis were treated with annuloplasty; in these, no reoperation was required.

By Cox regression analysis, only preoperative isolated AR was associated with reoperation (P = 0.001). Neither UAV (P = 0.907), BAV (P = 0.102) or a dilated annulus was associated with reoperation (P = 0.373). Autograft support (P = 0.002) was a protective factor. With autograft support, preoperative isolated AR was not associated with reoperation (P = 0.088) whereas endocarditis remained associated with reoperation (P = 0.015). Data are provided in Supplementary Material, Table S1.

Autograft regurgitation

Twenty-three patients developed AR ≥ 2 postoperatively (AR n = 9; 39%, AS n = 6; 26%, CD n = 8; 35%; P = 0.568). Of these, 2 had AR 2 at discharge (AR n = 1, CD n = 1); in 1, AR decreased and is now trivial 20 years postoperatively. The second patient showed progression to severe AR and required a reoperation 8 months postoperatively. Of the remaining 21 patients, AR remained stable in 11. It progressed to relevant AR in 10 instances (6 with autograft dilatation) and required a reoperation (AR n = 7, CD n = 3).

Freedom from AR ≥ 2 was 83% at 15 years (AR: 64%, AS: 92%, CD: 88%; P = 0.161). Freedom from AR ≥ 2 at 15 years was 87% with and 80% without autograft support (P = 0.233). Excluding patients with endocarditis, freedom from AR ≥ 2 at 15 years was 92% with and 82% without autograft support (P = 0.529).

By Cox regression analysis, AR > 1 at discharge was associated with recurrent AR (P = 0.044). The use of autograft support (P < 0.001) was a protective factor against AR. With support, AR > 1 at discharge was not associated with AR (P = 0.691). Neither valve morphology (P = 0.739) nor endocarditis (P = 0.062) was associated with recurrent AR. (Data are provided in Supplementary Material, Table S1.

Autograft dilatation

Dilatation of sinus diameter >45 mm was observed in 9 patients (AR n = 3, AS n = 3, CD n = 3; Table 2). Two patients developed sinus dilatation >50 mm after 10 and 14 years (AR n = 2, unsupported). At 10 years, freedom from autograft dilatation (>45 mm) was 94% (AR 92%; AS 95%; CD 94%; P = 0.525). It was 95% with and 91% without autograft support (P = 0.253) and 93% without and 98% with ascending aortic replacement (P = 0.755).

Patients without autograft support had higher progression rates over time (Fig. 6) and showed a significant root size progression over time in all groups (AR: 0.991 mm/(year*m), P = 0.038; AS: 0.688 mm/(year*m), P = 0.041; CD: 0.877 mm/(year*m), P = 0.001). Patients with autograft support had lower progression rates in all groups (AR: 0.091 mm/(year*m), P = 0.862; AS: 0.019 mm/(year*m), P = 0.962; CD: 0.077 mm/(year*m), P = 0.707).

By Cox regression analysis, neither BAV (P = 0.808) nor UAV (P = 0.233) were associated with autograft dilatation (>45 mm). Data are provided in Supplementary Material, Table S1.

DISCUSSION

In recent years, the interest in the Ross procedure has increased as the replacement option with the fewest long-term complications [3]. Its use in patients with AR, however, is controversial because of concerns over autograft failure. Irrespective of durability, the late survival benefit observed with the Ross procedure in AS is preserved in AR [20]. Several studies have found a higher risk of reintervention in the presence of preoperative AR and a dilated aortic annulus [7, 8, 10, 11, 16]. In those studies, the mechanism of autograft failure was not always clear. It has been postulated that the presence of AR and a dilated aortic annulus may present a genetic disposition with aneurysmatic degeneration of the pulmonary autograft [17], leading to failure [23]. We can partially confirm this hypothesis because patients with AR ‘without’ autograft support had a higher risk of failure even when excluding patients with endocarditis. With support, however, autograft stability was independent of the underlying haemodynamic lesion, confirming investigations in which patients with preoperative AR and AS had similar reoperation rates [7, 22, 24]. Overall, autograft support had the most significant impact on durability in patients with AR.

In the context of AR and annular dilatation, a BAV has been a risk factor for autograft failure [6, 11, 24]. Interestingly, we did not observe such a relationship; only a UAV was associated with a lower freedom from autograft reoperation [5, 22]. Autograft failure was rarely found in patients ‘with autograft support’ (1/137; 1%) [22, 25]. In a few of our patients with unicuspid aortic valves, the remnants of root tissue were insufficient for circumferential stabilization after excision of the coronary ostia. Our findings regarding UAV are new, which may be related to the suboptimal recognition of the UAV [26].

Progression of root size over time is frequently observed with all morphologies and underlying haemodynamic valve dysfunction. In our series, relevant autograft dilatation was observed only in patients ‘without’ support. Patients ‘with’ support showed stable root size irrespective of haemodynamic lesion or valve morphology. Thus, the determinant of autograft failure appears to be the surgical technique [20, 22]. This observation is supported by other publications showing excellent results using autograft support, such as inclusion with the native aortic wall [19, 24], a Dacron prosthesis [9] or stabilization of the sinus and the sinotubular junction (STJ) portion [16, 20]. Although the use of Dacron as support is limited to adults, concerns exist that this technique may result in coronary artery distortion, inflammatory changes of the autograft and a higher risk of infection [27, 28].

The higher rate of reoperations observed with endocarditis may be explained by the fact that autograft support was limited in these instances, and an annuloplasty was not added because of fear of persistent infection with extensive perivalvular abscess. None of the patients with endocarditis and full support, including annuloplasty, underwent a reoperation. The autograft support was also consistently used in reoperations; direct reimplantation of the coronary arteries was always feasible.

Ascending aortic dilatation has been similarly associated with autograft failure [29], and STJ dilatation has been assumed as a mechanism of failure [29]. Consequently, sinotubular stabilization has been advocated [16, 20]. It is currently uncertain whether this approach will suffice without additional root support. In our study, 78 patients underwent ascending aortic replacement as sinotubular stabilization. Some of these patients required reoperation for autograft dilatation, generally without complete autograft support. Thus, STJ stabilization ‘alone’ seems insufficient to prevent dilatation completely. A combination of annular and STJ stabilization has been proposed by others [16, 20] with limited follow-up. Based on our experience, a combination of annular, sinus and STJ stabilization may be the best approach for durability [22], both in AR and AS.

Limitations

The main limitation of this study is its observational design. Although data of consecutive procedures were obtained prospectively, the analysis was performed retrospectively, and treatment allocation was not randomized. The reproducibility of our findings may be limited to experienced surgeons in a high-volume centre performing the procedures. Despite these limitations, this study is one of only a few long-term studies comparing the results of patients with AR and AS with and without autograft support.

CONCLUSION

The Ross procedure is a good option with excellent survival for patients with not only AS but also with AR, if the surgical technique addresses the risk factors for failure. Patients with AR seem to have an increased risk of reintervention only, ‘without’ autograft support. External support mitigates autograft failure and dilatation in all valve morphologies and surgical indications, particularly in patients with AR.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

FUNDING

No funding was received for this study.

Conflicts of interest: none declared.

DATA AVAILABILITY

Data underlying this article will be shared on request to the corresponding author.

ACKNOWLEDGEMENTS

We thank Pavel Zacek for participation in creating the illustrations.

Author contributions

Karen Beatrix Abeln: Conceptualization; Data curation; Formal analysis; Methodology; Visualization; Writing—original draft. Lennart Froede: Data curation; Investigation; Software; Validation. Tristan Ehrlich: Investigation; Methodology. Idriss Souko: Project administration; Supervision; Validation. Hans-Joachim Schäfers: Conceptualization; Project administration; Resources; Supervision; Validation; Visualization; Writing—review & editing.

Reviewer information

The European Journal of Cardio-Thoracic Surgery thanks Emre Belli and the other anonymous reviewers for their contributions to the peer review process of this article.

Presented at the European Association for Cardio-Thoracic Surgery Aortic Forum 2023, Brussels, Belgium

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• AR

  Aortic regurgitation

 
• AS

  Aortic stenosis

 
• BAV

  Bicuspid aortic valve

 
• BSA

  Body surface area

 
• CD

  Combined disease

 
• STJ

  Sinotubular junction

 
• UAV

  Unicuspid aortic valve