Outcomes of valve-sparing surgery in heritable aortic disorders: results from the AVIATOR registry

Abstract

Open in new tabDownload slide

OBJECTIVES

Root reimplantation has been the favoured approach for patients with heritable aortic disorder requiring valve-sparring root replacement. In the past few years, root remodelling with annuloplasty has emerged as an alternative to root reimplantation in the general population. The aim of this study was to examine the late outcomes of patients with heritable aortic disorder undergoing valve-sparring root replacement and compare different techniques.

METHODS

Using the AVIATOR registry (Aortic Valve Insufficiency and ascending aorta Aneurysm InternATiOnal Registry), data were collected from 5 North American and European centres. Patients were divided into 4 groups according to the technique of valve-sparing used (root reimplantation, root remodelling with ring annuloplasty, root remodelling with suture annuloplasty and root remodelling alone). The primary endpoints were freedom from aortic regurgitation (AR) ≥2 and freedom from reintervention on the aortic valve. Secondary endpoints were survival and changes in annular dimensions over time.

RESULTS

A total of 237 patients were included in the study (reimplantation = 100, remodelling + ring annuloplasty = 76, remodelling + suture annuloplasty = 34, remodelling alone = 27). The majority of patients had Marfan syndrome (83%). Preoperative AR ≥2 was present in 41% of the patients. Operative mortality was 0.4% (n = 1). No differences were found between techniques in terms of postoperative AR ≥2 (P = 0.58), reintervention (P = 0.52) and survival (P = 0.59). Changes in aortic annulus dimension were significantly different at 10 years (P < 0.05), a difference that started to emerge 4 years after surgery.

CONCLUSIONS

Overall, valve-sparring root replacement is a safe and durable procedure in patients with heritable aortic disorder. Nevertheless, root remodelling alone is associated with late annular dilatation. The addition of an annuloplasty, however, results in similar freedom from AR, reintervention, survival and changes in annulus size compared to reimplantation.

INTRODUCTION

While total aortic root replacement (Bentall procedure) has been considered the gold standard for the management of aortic root aneurysms in patients with heritable aortic disorders (HADs), valve-sparing root replacement (VSRR) has recently emerged as the preferred procedure whenever feasible. This is based on the notion that the aortic valve is often structurally normal in these patients and can therefore be preserved at the time of surgery [1]. Furthermore, patients with HAD usually present for surgery at a young age. Therefore, avoidance of lifelong anticoagulation with a mechanical prosthesis or early valve degeneration with a bioprosthesis represents significant benefits. Indeed, various studies have highlighted the advantages of VSRR over composite valve grafts in patients with Marfan syndrome [2, 3].

The 2 main VSRR techniques are the remodelling and reimplantation techniques [4]. While remodelling is considered a more physiological reconstruction of the aortic root, reimplantation has been regarded as a more appropriate approach for patients with HAD because it fixes the basal ring, thus preventing annular dilatation. A study from the Toronto group confirmed this notion [5], showing lower rates of recurrent aortic regurgitation (AR) and reoperation after reimplantation. In recent years, the addition of an aortic annuloplasty using an external ring or a circumferential suture technique has been proposed to mitigate potential annular dilatation in patients undergoing remodelling [6]. This addition may provide a more durable operation and allow a more physiological root reconstruction. While there are some published reports of VSRR in patients with HAD, most are single-centre, single-surgeon series. In addition, there is a paucity of data addressing the role of various contemporary VSRR techniques for the management of patients with HAD.

The AVIATOR Registry (Aortic Valve Insufficiency and ascending aorta Aneurysm InternATiOnal Registry) is an international research network established in 2013 within the Heart Valve Society. It includes data on patients with AR, ascending aortic aneurysms and patients undergoing aortic valve repair or replacement. The aim of this study was to compare the role of various VSRR techniques in the surgical management of patients with HAD. The primary endpoints of this study were the cumulative incidence of postoperative AR ≥2 and cumulative incidence of reintervention on the aortic valve. Secondary endpoints were survival and changes in annular dimensions over time.

PATIENTS AND METHODS

Ethics statement

Institutional Review Board approval was obtained at each participating site (IRB #2017-1974—AORTA). For data collection into AVIATOR, individual patient consent was waived.

Study population

Patients operated between 1 January 1996 and 1 January 2018 in 5 different centres were included in the study. Two centres initiated their VSRR program in 1996. The other centres debuted their programs in 2001, 2003 and 2013, respectively (Supplementary Material, Table S1). Patients were subdivided into 4 groups (root reimplantation, root remodelling with ring annuloplasty, root remodelling with suture annuloplasty and root remodelling alone) according to the VSRR technique performed. The type of valve-sparing technique used was based on surgeon preference and expertise.

A total of 237 consecutive patients with HAD were included in the study (reimplantation = 100, remodelling + ring annuloplasty = 76, remodelling + suture annuloplasty = 34, remodelling alone = 27). Most patients had Marfan syndrome (83%). Eighteen patients (8%) were diagnosed with familial thoracic aorta aneurysm and dissection, 14 (6%) had Loeys–Dietz syndrome, 4 patients (2%) had Turner syndrome, 3 (1%) were diagnosed with vascular Elhers–Danlos syndrome and 2 had other HAD. Overall, median follow-up was 5.0 years (interquartile range = 2.6–9.5) and 93% complete (1568 patients-year). Patients in the ‘remodelling alone’ group had a longer median follow-up [remodelling alone = 11.7 (8.9–16.8) years, reimplantation = 5.0 (2.4–9.7) years, remodelling + ring annuloplasty = 4.7 (2.3–7.3), remodelling + suture annuloplasty = 4.3 (2.8–5.6); P < 0.01]. Demographic data are summarized in Table 1. Ninety-seven (41%) patients had AR ≥ 2 preoperatively (reimplantation = 37%, remodelling + ring annuloplasty = 34%, remodelling + suture annuloplasty = 47%, remodelling alone = 67%, P = 0.02). Eleven patients (5%) presented with acute aortic dissection (reimplantation = 7%, remodelling + ring annuloplasty = 1%, remodelling + suture annuloplasty = 0%, remodelling alone = 11%; P = 0.03). During the same period, 134 patients with HAD underwent an aortic valve replacement or a Bentall procedure in the participating institutions (these were mostly performed in 3 centres before the initiation of VSRR programmes) and 1004 patients without HAD underwent VSRR.

Surgical techniques

The principles of aortic root remodelling and reimplantation have previously been described [7–11]. Briefly, remodelling alone (without any support of the basal ring) is performed by detaching the coronary buttons, excising the sinuses of Valsalva and anastomosing a tailored Dacron tube to the aortic ring and commissures to re-create 3 sinuses. The root preparation is similar for the reimplantation technique. However, a Dacron tube is anastomosed at the level of the basal ring and the commissures, as well as the remaining aortic root tissue, are subsequently sutured inside the graft.

External aortic annuloplasty was popularized in the early 2000s by Lansac et al. [6] and sizing charts were introduced in 2008. The ring annuloplasty is positioned using pledgeted sutures placed on a circumferential plane in the basal ring. The subvalvular stitches are passed through a flexible prosthetic aortic ring [Dacron or external aortic ring (Coroneo™)], which is lowered and tied into position. A total of 76 patients were operated using this technique. In 34 patients, a circumferential suture annuloplasty (using a polytetrafluoroethylene (PTFE) suture) was used to stabilize the basal ring [12]. The 4 different techniques are illustrated in Fig. 1.

Residual cusp prolapse was eye-balled until the introduction of a dedicated calliper which was used from 2008 onwards [13]. When the cusp effective height was found to be <9 mm, valve repair was performed using various techniques to correct cusp prolapse. Valve repair was performed in 114 patients (48%) and central cusp plication was the most common technique used (45%). The other techniques are listed in Table 2.

Definitions

HADs were diagnosed based on specific genetic mutation and/or clinical findings. Aortic annulus measurements were obtained from transthoracic or transoesophageal echocardiography (transthoracic echocardiography (TTE) or transesophageal echocardiography (TEE)) using either two-dimensional or three-dimensional methods. The degree of AR was assessed in accordance with the American Society of Echocardiography (ASE).

Statistical analyses

Continuous variables are presented as mean ± SD and compared using ANOVA with post hoc tests and the Bonferroni correction for normally distributed data and Kruskal–Wallis test for non-normally distributed data. Categorical variables are reported as count or percentages and were compared using Chi-Square test or Fisher’s exact test, as appropriate. Survival was estimated using the Kaplan–Meier method. The log-rank test was used for comparison of survival curves. Reinterventions were analysed using competing risk analysis, with mortality as a competing risk. The Gray test was used to compare groups. Linear mixed model analyses were used to compare changes in aortic annulus diameter over time and depicted using spline functions with 5 degrees of freedom. A Cox proportional hazard regression was performed to identify predictors for the development of AR ≥ 2. The proportional hazard assumptions were verified using the Schoenfeld residuals and the cox.zph function. Exploratory analyses were performed to compare temporal outcomes ‘before 2008’ and ‘after 2008’. The year 2008 was chosen because it equally divides the study period and also because it is the year during which centres started to use a dedicated calliper to assess residual prolapse and also corresponds to the introduction of calibrated extra-aortic ring. For all analyses, P < 0.05 was considered statistically significant. All statistical analyses were performed using R version 3.6.0 (R Foundation, Vienna, Austria).

RESULTS

Operative results

Operative results are summarized in Table 2. There was 1 in-hospital death for an overall in-hospital mortality of 0.4%. Other postoperative complications, including dialysis, myocardial infarction and reintervention for bleeding, did not differ between groups. None of the patients required a new pacemaker implantation. Three patients had AR graded 2 on discharge echocardiography. All other patients had AR < 2. Cross-clamp time (min) differed significantly between techniques (reimplantation = 128 ± 27, remodelling + ring annuloplasty = 144 ± 33, remodelling + suture annuloplasty = 74 ± 17 and remodelling alone = 112 ± 43; P < 0.01).

Primary outcomes

Postoperative AR ≥2

Thirty-six patients developed AR ≥ 2 during follow-up (reimplantation = 14, remodelling + ring annuloplasty = 10, remodelling + suture annuloplasty = 2 and remodelling alone = 10). The cumulative incidence of AR ≥ 2 for the entire cohort was 22.0 ± 4.1% at 10 years. There was no difference in terms of cumulative incidence of AR ≥ 2 (Fig. 2) when comparing all 4 groups (P = 0.58). At 10 years, the cumulative incidence of postoperative AR ≥ 2 was 20.3 ± 6.6% (root reimplantation), 21.6 ± 7.4% (root remodelling + ring annuloplasty), 7.8 ± 5.6% (root remodelling + suture annuloplasty, at 8 years) and 34.2 ± 9.5% (root remodelling alone).

Among patients who developed AR ≥2, 33 had Marfan syndrome, 1 patient had Loeys–Dietz syndrome, 1 patient had Elhers–Danlos syndrome and 1 patient had Turner syndrome (P = 0.24). The median time between the initial operation and occurrence of postoperative AR ≥ 2 was 2.8 years (1.1–8.2). The characteristics of patients who developed postoperative AR ≥ 2 are summarized in Supplementary Material, Table S1.

Of the 36 patients with AR ≥ 2, 12 had AR graded 3 or 4 (reimplantation = 4 patients, remodelling + ring annuloplasty = 4, remodelling + suture annuloplasty = 0 and remodelling alone = 4; P = 0.08). Most of these patients had Marfan syndrome (11/12 patients) and 1 patient (1/12) had Elhers–Danlos syndrome (P = 0.36). The median time between the initial surgery and the occurrence of AR 3 or 4 was 2.0 years (0.3–2.8). All patients who developed AR graded 3 or more underwent reintervention. In univariable analyses (Table 3), only the presence of any degree of AR on immediate postoperative echocardiography was associated with the development of postoperative AR ≥ 2 [hazard ratio  = 3.6 (95% confidence interval 1.4–8.8), P < 0.01].

Reintervention on the aortic valve

There was a total of 18 reinterventions on the aortic valve during the study period (Marfan syndrome = 15 patients, Elhers–Danlos = 1 patient, Loeys–Dietz syndrome = 1 patient, Beals syndrome = 1 patient). The cumulative incidence of reintervention for the entire cohort was 10.9 ± 3.1% at 10 years and did not differ between techniques. At 10 years, the cumulative incidence of reintervention on the aortic valve was 16.2 ± 6.1% for root reimplantation, 6.6 ± 3.3% for root remodelling + ring annuloplasty, 0% for root remodelling + suture annuloplasty (at 8 years) and 11.1 ± 6.2% for root remodelling alone (Fig. 3, P = 0.52).

Most reoperations (14/18) were performed due to AR ≥ 2 and cusp prolapse was the most common mechanism of AR (7/14). Other primary causes of reintervention were: pseudoaneurysm (n = 1), myxoma with repair of the aortic valve (n = 1) and endocarditis (n = 2). Thirteen patients had a valve replacement (biological aortic valve replacement = 5, mechanical aortic valve replacement = 6, homograft = 1 and ranscatheter aortic valve replacement (TAVR) = 1) and 5 patients underwent valve repair. There was no perioperative death among patients who underwent reintervention.

Secondary outcomes

Survival

Eleven patients died during follow-up. The specific causes of death are listed in Supplementary Material, Table S2. The survival for the entire cohort at 10 years was 92.3 ± 2.9%. At 10 years, survival was 89.8 ± 5.1% after reimplantation, 100% after remodelling with ring annuloplasty, 97.1 ± 2.9% (at 8 years) after remodelling with suture annuloplasty and 88.5 ± 6.3% after remodelling alone (Supplementary Material, Fig. S1; P = 0.59).

Changes in annular dimensions

Changes in annular dimension were different between the reimplantation, remodelling with annuloplasty and remodelling alone techniques (Fig. 4, P < 0.05). Annulus diameters were similar among the various techniques up to 3 years postoperatively. After 3 years of follow-up, the annulus of patients who underwent remodelling alone was significantly larger.

Temporal trend in outcomes

Exploratory analyses were performed to compare outcomes before and after 2008. A total of 67 patients were operated before 2008 and 170 patients had their operation after 2008. The cumulative incidence of AR ≥ 2 (P = 0.02) and reintervention were higher before 2008 (P = 0.01; Supplementary Material, Fig. S2A and B). There was no difference in terms of survival (P = 0.40; Supplementary Material, Fig. S3).

DISCUSSION

The main findings from this study are that (i) VSRR is a safe procedure in patients with HAD; (ii) late outcomes after VSRR are encouraging in terms of durability; (iii) no differences were observed in terms of long-term recurrence of AR or reintervention between different VSRR techniques; (iv) while remodelling alone is associated with progressive enlargement of the aortic annulus over time, this is eliminated by addition of an extra-aortic suture or ring annuloplasty; and (v) VSRR, whenever associated with annular stabilization, is associated with favourable long-term survival in patients with HAD.

This study confirms the notion that remodelling alone is associated with progressive dilatation of the aortic annulus over time in patients with HAD. This has often been proposed as the main limitation of the remodelling procedure in this patient population. Nevertheless, a recent study examining changes in aortic annular diameters in 66 patients with Marfan or Loeys–Dietz syndrome (mean follow-up: 7 years) found no changes in annular diameter in patients with a preoperative aortic annulus <25 mm [14]. This is not inconsistent with previous studies in patients with Marfan syndrome. Indeed, in a study from the Toronto group, only 3 of the 25 patients undergoing remodelling required reintervention over a follow-up period reaching 26 years [15]. Although this represented significantly more reinterventions than the reimplantation group, it suggests that in most patients with HAD undergoing an isolated remodelling procedure, favourable long-term durability can nonetheless be expected. In this multi-centric study, the incidence of reintervention did not differ between techniques, despite the fact that a greater proportion of patients who underwent remodelling alone presented for acute dissection or with AR ≥ 2. Our results should be interpreted in light of these characteristics as they are known risk factors for VSRR failure. Nevertheless, while fixing the annulus may not be necessary in all patients, it is probably best to stabilize the annulus in all patients with HAD given the current state of knowledge.

While composite valve graft was once the favoured operation in patients with HAD, an increasing number of studies highlight the favourable outcomes associated with VSRR [16]. Our study provides novel information in that it demonstrates that different VSRR techniques are associated with favourable late outcomes in patients with HAD. In this study, the 10-year incidence of reintervention demonstrates the durability of VSRR. Finally, the longitudinal echocardiographic follow-up of this study provides a high level of certainty that the incidence of reintervention truly reflects the incidence of VSRR failure.

Though there was no statistical difference in the cumulative incidence of AR ≥ 2 in this study, this may be in large part due to the limited number of patients in the remodelling alone group. Nevertheless, it is sobering to notice that the cumulative incidence of AR ≥ 2 for the entire cohort was 22% at 10 years, which requires some pause. Similarly, in a recent multicentre study of VSRR in Marfan patients from a number of highly selected US centres, the 1-year and 3-year incidence of AR ≥ 2 was 7% and 10%, respectively [17]. The centres included in our study all have vast expertise in aortic root reconstructive surgery, which underscores the fact that while these operations are very safe, even in expert hands, there is a degree of challenge to ensure optimal long-term durability. This also highlights the need to ensure long-term follow-up of these patients as there is a persistent risk of failure in this patient population. However, it is comforting that only a minority of patients who developed AR had grade 3–4 AR during follow-up. Importantly, there was a clear temporal trend towards improved durability of valve function when comparing results from before and after 2008. This is due to several factors related to technical nuances of each VSRR technique including graft size selection and commissural height, but importantly to more precise management of induced or residual cusp prolapse.

Recurrent AR after VSRR in patients with HAD can be caused by several mechanisms. In fact, disease progression at the level of the aortic leaflets can lead to leaflet prolapse and clinically significant AR. Alternatively, induced cusp prolapse is related to the operative technique itself. Indeed, one of the main principles of VSRR is to reduce the diameter of the sinotubular junction. The resulting decrease in inter-commissural distance can induce leaflet prolapse, which is a known predictor of recurrent AR after VSRR [18]. Furthermore, in patients with HAD, there is often elongation of aortic valve leaflets, both in their radial and circumferential axes [19]. This can also result in prolapse of the cusps and should be corrected at the time of surgery. While ‘eye-balling’ was the main method of assessing cusp prolapse intraoperatively in the past, specific intraoperative tools have contributed considerably to standardize aortic cusp management at the time of VSRR [13, 20]. This may in part explain the temporal trend towards better results after 2008 as observed in our study, and as confirmed by others [21]. Given that the presence of any degree of postoperative AR was the only predictor of AR ≥ 2, our study confirms the importance of obtaining a perfect intraoperative result to ensure long-term durability.

Based on the results of this study, reimplantation and remodelling with annuloplasty in patients with HAD translate into similar late clinical outcomes. To date, there is a paucity of data directly comparing these 2 approaches. Considering that patients with HAD would be most expected to fail in the setting of a remodelling procedure, the fact that no differences are observed suggests that the same would hold in different patient populations. This is in line with a recent single-centre study comparing reimplantation to remodelling with ring annuloplasty in all comers, demonstrating similar survival, freedom from AR and reintervention [22]. Nevertheless, our results should be interpreted with caution because of the relatively low number of patients in each group. Furthermore, the follow-up period may not be long enough to elicit significant differences. Given the young age of these patients, follow-up into the second and third decade is of utmost importance. Altogether, the cumulative incidence of reintervention was higher than that reported by the Toronto group in patients with Marfan syndrome [5]. However, this series differ significantly from ours with only 10% of patients presenting with severe preoperative AR (versus 20% with AR 3–4 and 41% with AR ≥ 2 in our study). Moreover, this study encompasses the experience of several different centres and multiple surgeons. It also captures the outcomes of 4 different VSRR techniques. While there is room for improvement, these results possibly constitute a more representative picture of global outcomes in this patient population, taking into account the learning curve from each centre and perhaps encompassing a wider patient selection.

Limitations

This study has a number of limitations. Follow-up was significantly different between the 4 surgical techniques, perhaps favouring techniques that have a shorter follow-up. Nevertheless, it is the largest study focusing on VSRR in patients with HAD. Because of the multicentre nature of the data, there could be variability in some of the echocardiographic measures. Nevertheless, the international and multi-centric design of this study provides a strong basis for external validity and reproducibility of these results. All data before 2013 were entered retrospectively, which introduces inherent recall and reporting biases. Finally, the follow-up is still relatively short considering the young age of the patients included in this study.

CONCLUSION

In patients with HAD, VSRR is a safe procedure when performed in high-volume centres. All techniques showed similar results in terms of cumulative incidence of postoperative AR ≥ 2, cumulative incidence of reintervention and survival. The addition of an annuloplasty to the remodelling procedure provides better annulus stability when compared with remodelling alone. Nevertheless, there is a significant incidence of recurrent AR in the first decade after surgery. Though it does not immediately translate into high rates of reintervention, longer follow-up is necessary to determine the durability and confirm the proposed benefits of VSRR in this patient population.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

Acknowledgement

Pavel Zacek for his help with the illustrations.

Funding

None.

Conflict of interest: Emmanuel Lansac reports consulting fees from HeartWare, Syncardia, and Medtronic, and a consultant agreement with CORONEO, Inc; Vincent Chauvette is a recipient of the Vanier Scholarship.

Data availability statement

The data underlying this article will be shared on reasonable request to the corresponding author.

Author contributions

Vincent Chauvette: Conceptualization; Data curation; Formal analysis; Methodology; Project administration; Supervision; Validation; Writing—original draft. Jolanda Kluin: Conceptualization; Investigation; Methodology; Writing—original draft. Laurent de Kerchove: Conceptualization; Investigation; Methodology; Writing—original draft. Gebrine El Khoury: Conceptualization; Investigation; Writing—original draft. Hans-Joachim Schäfers: Conceptualization; Investigation; Methodology; Writing—original draft. Emmanuel Lansac: Conceptualization; Investigation; Writing—original draft. Ismail El-Hamamsy: Conceptualization; Investigation; Methodology; Project administration; Supervision; Writing—original draft.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Ruggero De Paulis, Yutaka Okita, Clarence Pienteu Pingpoh and the other anonymous reviewer(s) for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• AR

  Aortic regurgitation

 
• HAD

  Heritable aortic disorder

 
• VSRR

  Valve-sparring root replacement