Supracommissural replacement of the ascending aorta and the aortic valve via partial versus full sternotomy—a propensity-matched comparison in a high-volume centre

Abstract

 

Open in new tabDownload slide

OBJECTIVES

Full sternotomy (FS) is the common surgical access for patients undergoing open aortic valve replacement (AVR) with concomitant supracommissural replacement of the tubular ascending aorta. Since minimally invasive approaches are being used with increasing frequency in cardiac surgery, the aim of this study was to compare outcomes of patients undergoing AVR with supracommissural replacement of the tubular ascending aorta via FS versus partial upper sternotomy (PS).

METHODS

We included all patients who underwent elective AVR with concomitant supracommissural replacement of the tubular ascending aorta at our institution between 2000 and 2015. Exclusion criteria were emergency surgery, other major concomitant procedures and reoperations. After 2:1 propensity score matching, outcomes of patients with PS and FS were compared.

RESULTS

A total of 652 consecutive patients were included, 117 patients operated via PS and 234 patients operated via FS. Cardiopulmonary bypass time and aortic cross-clamp time of the PS and FS groups were 89 vs 92 min (P = 0.2) and 65 vs 70 min (P = 0.3), respectively. Postoperative morbidity was low and there were no significant differences in postoperative outcomes between patient groups. In-hospital mortality was 1.7% in the PS vs 0.4% in the FS group (P = 0.3). Kaplan–Meier analysis revealed no difference in mid-term survival (P = 0.3). Reoperation rates for valve or aortic complications were very low with no significant difference between groups.

CONCLUSIONS

In a high-volume centre with extensive experience in minimally invasive cardiac surgery, AVR with concomitant supracommissural replacement of the tubular ascending aorta via PS results in similar outcomes with regard to safety and longevity when compared to conventional FS.

INTRODUCTION

In patients with indication for aortic valve replacement (AVR), current guidelines recommend concomitant supracommissural replacement of the ascending aorta (scAAR) once the diameter reaches 45 mm, or earlier if rapid diameter progression is present [1–4]. In addition, concomitant AVR is recommended in patients undergoing surgery for patients with an ascending aorta aneurysm when the aortic valve is moderate-to-severely stenotic or insufficient, even in the absence of valve-related clinical symptoms [2]. In 2019, a combined AVR and scAAR procedure was performed 1372 times in Germany (and ∼10 000 times in Europe), with a mortality rate of 4.2% (i.e. 500 patients/year in Europe) [5].

While partial upper sternotomy (PS) has been established as a less-invasive approach for isolated AVR [6], full sternotomy (FS) remains the standard approach for combined and complex procedures. Numerous studies have demonstrated that partial upper sternotomy approach in patients undergoing isolated AVR shortens intensive care unit (ICU) and hospital stay and overall recovery time, in addition to reducing blood transfusion and pneumonia rates, when compared to FS [6, 7]. Therefore, it has been suggested that minimally invasive approach not only provides a cosmetic and socio-economic benefit but also improves overall acceptance of life-saving cardiac surgery among patients undergoing AVR through a full median sternotomy.

However, there are only small studies available on the topic of minimally invasive AVR with concomitant scAAR, and no study has yet conducted a true comparison of the minimally invasive PS approach with FS in this context. The aim of this study was to therefore compare perioperative and postoperative results of patients who underwent AVR with concomitant scAAR via PS to those of a propensity score-matched patient cohort that received the same operation via FS.

METHODS

Patient selection

The study was approved by the ethics committee of the medical faculty of the University of Leipzig (177/15). We retrospectively reviewed our institutional database and included all patients ≥18 years who underwent elective AVR with concomitant scAAR at our institution between 2000 and 2015. Exclusion criteria were emergency surgery and additional major procedures such as coronary artery bypass grafting (CABG), other heart valve surgery, aortic root and/or arch procedures and reoperations. The decision for surgical intervention was predominantly made by aortic valve dysfunction.

Patient charts were reviewed by 1 examiner and aortic diameters were measured by routine electrocardiogram-gated, contrast-enhanced computed tomography (CT) for aortic root and ascending aorta at the level of the pulmonary artery bifurcation. Aortic valve disease was assessed preoperatively by transthoracic and intraoperatively by transoesophageal echocardiography by a certified echocardiographer in all patients.

Operative technique

All operations were performed via either full median sternotomy or partial upper J- or T-shaped sternotomy at the level of the third or fourth intercostal space. Cardiopulmonary bypass (CPB) was usually established via distal ascending aorta and right atrium cannulation in the FS group. In the PS group, venous access was gained via either direct cannulation of the atrial appendage or via the femoral vein. Arterial access was either performed direct via the distal ascending aorta or, in select cases, via femoral cannulation. A left ventricular vent was used in all operations, usually via the right superior pulmonary vein. Exposure in cases of PS can be challenging but placement of pericardial stay sutures provides optimal exposure. Vent placement can be facilitated by deconnecting ventilation and increasing volume drawout on CPB. If the pulmonary vein is still difficult to reach, insertion via pulmonary artery is a reasonable option. Antegrade application of crystalloid or blood cardioplegia was conducted in all cases. In patients with moderate or more aortic insufficiency, cardioplegia was administered directly into the coronary ostia using mushroom- or olive-tipped catheters. According to patient preference and individual risk profile, standard biological and mechanical prostheses were used for AVR, and tubular Dacron prostheses for scAAR (e.g. Hemashield Platinum™ woven double velour collagen impregnated polyester graft, Getinge, Gothenburg, Sweden). Depending on the daily logistics and/or perioperative risk profile, patients were either fast tracked through our perioperative anaesthesia care unit or admitted to ICU prior to transfer on intermediate care unit.

Follow-up

Follow-up (FU) was performed according to the institutional database supplemented by individual patient records. FU data for long-term survival or adverse outcomes were routinely recorded via direct telephone interviews with the patient, a close relative or the referring physician and was closed on 21 July 2017. In patients with aortic aneurysm limited to the ascending aorta, no regular CT FU was done.

Statistical analysis

A 2:1 propensity score matching was performed, resulting in 2 patient groups with comparable preoperative characteristics. Propensity score matching is a method to reduce selection bias in a retrospective study setting by creating a score for each individual that represents the probability of being assigned to the treatment and not the control group, using baseline variables with potential influence on outcomes in a logistic regression model. The variables included were age, sex, body mass index, arterial hypertension, diabetes, pulmonary hypertension, smoking status, chronic pulmonary disease, urgency of surgery, presence of a bicuspid aortic valve and ascending aortic diameter. Preoperative data were compared using Welsh’s two-sided t-test (due to 2:1 matching, no paired testing of matched groups was feasible) or Pearson’s χ² test and, after the assessment of continuous variables for normal distribution, reported as mean (standard deviation) or number (percentage). Perioperative and postoperative data were analysed using Pearson’s χ² test or Fisher’s exact test, as appropriate, for categorical data, Wilcoxon rank sum test for continuous data and Kaplan–Meier analysis with log-rank test for survival. Categorical variables are presented as total numbers (percentages), and odds ratios with 95% confidence intervals were used to express their association with the event of interest. Continuous variables are presented as median with interquartile range. Statistical analysis was performed using SPSS version 25.0 [8] and R version 3.6.1 [9] with the tidyverse [10] and survival [11] packages. Statistical significance is indicated by a P-value of <0.05.

RESULTS

Patient cohort

A total of 652 patients met the inclusion and exclusion criteria, of which 123 underwent AVR and scAAR via PS and 529 via FS. Proportions of conventional and minimally invasive operations between 2000 and 2015 are displayed in Fig. 1. Unmatched groups showed significant differences in the rates of pulmonary hypertension, presence of bicuspid aortic valve and mean ascending aorta diameter. Aortic valve stenosis was the predominant valve dysfunction in both groups (PS 78%, FS 83%, P = 0.3). A total of 117 patients of the PS group (mPS) were successfully matched to 234 patients of the FS group (mFS) by 2:1 propensity score matching. Distribution of propensity scores in the 2 unmatched cohorts is visualized in Fig. 2. The 2 matched groups had no significant differences in any of the preoperative variables and the change in standardized mean differences between the unmatched and matched cohorts is depicted in Fig. 3. Characteristics of the unmatched and matched groups, including standardized mean differences, are displayed in Table 1.

Selection criteria for minimally invasive access included pulmonary comorbidities (e.g. chronic obstructive pulmonary disease), osteoporosis and frailty of the patient. Other soft criteria were deep located aortic valve, horizontal orientation of aortic root, elongated aorta, mediastinal shift and experience with femoral cannulation.

Operative details of the matched cohort

The study revealed no differences regarding implantation rates of biological and mechanical valve prostheses between the matched groups; details are displayed in Table 2. Additional morrow resection was performed in 7.3% of the patients in the mPS group and 7.7% in the mFS group (P = 1.0). There was 1 conversion to FS in the mPS group. No significant differences in median operation time, CPB time and aortic cross-clamp time were detected in the analysis (see Table 2).Postoperative outcome of the matched cohort

Total ventilation time was comparable in both cohorts with a slightly but not significantly increased rate of reintubation in the mFS group. The median ICU stay and the median intermediate care unit stay did not differ significantly between the 2 groups (for details see Table 2).

Stroke rates, re-exploration rates for acute postoperative bleeding and pericardial effusion were comparable. Sternal instability and sternal wound healing disorders were of similar rates in the 2 groups. The pacemaker implantation rate in the PS group was numerically 2.5 times higher but did not reach statistical significance. Details on postoperative outcome are found in Table 3.

Survival of the matched cohort

There were 2 cases of in-hospital deaths due to low-cardiac output in each group. With a median FU of 234 days in the PS group and 466 days in the FS group, the Kaplan–Meier analysis with log-rank test (Fig. 4) did not show any statistical difference in overall survival between the 2 groups (P = 0.9).

Aortic events and re-do surgery

As far as available FU data entails, no aortic event occurred in either of the groups and no patient needed a reoperation for recurrent aortic aneurysm disease during the FU period. A total of 2.6% of the mPS patients and 3% of the mFS patients required reoperation for valve-related disease. The reasons were prosthetic endocarditis in 1.7% of each group, paravalvular leak in 0.9% of the ‘mPS’ and 0.4% of the ‘mFS’ group and prosthetic valve degeneration in 0.4% of the ‘mFS’ group with none in the ‘mPS’ group.

DISCUSSION

Minimally invasive access surgery for isolated cardiac procedures is progressively on the rise and has become routine in specialized centres. Aortic and mitral valve surgeries via PS or small lateral thoracotomy incisions, respectively, have been shown to provide equivalent or even superior surgical outcomes to those following FS, without increase in perioperative risk or complication rates [12, 13]. Furthermore, the use of sutureless and rapid deployment valves in isolated minimally invasive AVR has increased over the past years as recently published by Berretta et al. [14]. However, in our cohort, we implanted a sutureless valve in only one patient.

Main concerns when introducing less-invasive incisions are prolonged procedural times and bleeding complications, due to restricted access and vision of the surgical field [15]. Thus, regarding less-invasive approaches, Kaneko et al. [16] conducted a retrospective analysis on patients undergoing aortic valve and concomitant aortic surgery via PS or FS, demonstrating no difference in outcomes in patients undergoing AVR with concomitant aortic surgery despite longer CPB and aortic cross-clamp times. Although CPB and aortic cross-clamp times appeared remarkably prolonged in the PS group (PS: 152 vs FS: 108 min), overall low morbidity and mortality rates were found [16]. Staromłyński et al. [17] revealed similar outcomes following isolated scAAR, AVR + scAAR and aortic root replacement via PS. The median CPB and aortic cross-clamp times in the AVR + scAAR group were within the range previously described by Kaneko (PS: 161 vs FS: 108 min, respectively). Of the AVR + scAAR patients, 7% needed surgical revision for bleeding and in-house mortality was 0% [17]. Another small series of 20 consecutive patients receiving AVR + AAR via right anterior mini-thoracotomy [18] showed satisfactory outcomes with no operative mortality. However, the median CPB and aortic cross-clamp times of 291 and 163 min, indicating that this approach is technically more challenging and could potentially increase the risk of CPB-related complications. As described by Modi and Chitwood [19] peripheral venous cannulation can be performed without any additional risk. Aortic calcification and soft plaques of the descending aorta and the aortic arch can be examined in preoperative contrast-enhanced CT. This is routinely performed at our institution not only for calcification but also for anatomical features, i.e. position of the aortic annulus.

Overall, very limited information on the role of minimally invasive surgery in concomitant treatment of aortic valve and aortic pathologies is available. The literature only provides retrospectively collected outcome data of small cohorts, and no study has yet compared PS and FS for combined AVR and scAAR in a propensity-matched analysis. The Mini-Stern randomized-controlled trial, even though underpowered, is the only prospective clinical study comparing AVR through a minimally invasive (118 patients) versus routine sternotomy (104 patients) approach, which had a recruiting period >5 years with a 1-year FU [20]. Interestingly, the authors of the Mini-Stern trial concluded that PS is not superior to FS with regard to outcome and cost effectiveness for AVR.

Our retrospective analysis detected non-inferiority of the less-invasive PS approach for combined AVR and scAAR. Postoperative morbidity and mortality were very low in both treatment groups with no significant differences in procedural safety between the groups. The minimally invasive approach did not prolong procedural times, as was observed in various studies mentioned above and was even shorter than procedures performed through FS reported by Sioris et al. [21]. This is most likely due to the large number of patients operated by experienced surgeons and utilization of innovative techniques, such as rapid deployment valves and CorKnot suturing systems, in selected patients at our institution. Furthermore, we have extended this minimally invasive approach to other complex aortic surgeries, like aortic root replacement and valve-sparing operations. It is imperative to target shorter CPB and cross-clamp times even in minimally invasive procedures because a large multicentre study with 27 125 patients demonstrated aortic cross-clamp time to be an independent predictor of mortality in patients with preserved ventricular function [22].

It should be taken into account that there are potential advantages of PS we did not examine in this study, e.g. improved quality of life and pain reduction, which should be addressed in prospective trials. In Mini-Stern, no significant difference was detected regarding quality of life or pain after isolated AVR via PS or FS [20], but larger studies are needed and from clinical experience, there is reason to assume substantial benefit for PS patients. A number of advantages of PS versus FS for isolated AVR have been shown by a large number of retrospective studies including shorter hospital stay, less transfusions and lower pneumonia rates [6, 7].

Although propensity score matching was performed, some results of this study may have still been affected by selection bias. For example, the similar rates of observed sternal complications between groups may have been due to the fact that surgeons chose a less-invasive approach if they suspected increased risk of impaired sternal healing. Osteopenia or osteoporosis might not have been diagnosed at early stages, but the surgeon’s decision could have been influenced by clinical experience. This is supported by the fact that before matching, the PS group had a significantly higher rate of pulmonary hypertension. These patients were most likely selected for minimally invasive surgery to reduce the risk of pulmonary complications. The largely higher rate of bicuspid valves in the PS group before matching also suggests an increased amount of younger and otherwise healthier patients (although not reflected by median age or other known preoperative variables), who received a less-invasive operation mainly for cosmetic reasons/patient preference. Rather small mean diameters of the mid-ascending aorta are due to the design of our study. Patients with larger aortic diameters commonly present with concomitant enlargement of aortic root and/or arch and were thus not included in this analysis.

Doenst and Lamelas [12] discussed the evolution of minimally invasive cardiac surgery procedures and the rationale to learn, teach and perform these procedures in a comprehensive critical review. They argued that there is a surprising divergence between scientific evidence and existing (frequently cosmetic) concerns among patients and referring physicians. According to their article, a ‘fear of sternotomy’ leads patients to be willing to accept disadvantages, more specifically increased rates of complications or inferior procedural outcomes and worse long-term prognosis. Cardiac surgeons are progressively confronted with this issue in various fields: despite proven long-term superiority of CABG over percutaneous coronary interventions (PCI) [23, 24], the relation of CABG and PCI has been continuously changing in favour of PCI over the last decade in Germany [5, 25]. Furthermore, although superior results of transcatheter AVR (TAVR) compared to conventional AVR in the short term were found in the PARTNER trials [26, 27], survival and cardiac event curves appear to cross in long-term observations [28]. A recent meta-analysis by Sayed et al. [29] compared outcomes of TAVR with those of minimally invasive surgical AVR by examining 11 cohort studies. As expected, TAVR led to shorter hospital length of stay. But despite equality with regard to in-hospital mortality and rates of stroke and bleeding complications, risks of mid-term mortality and paravalvular leakage were significantly higher in TAVR than minimally invasive surgical AVR [29]. Nevertheless, numbers of TAVR keep constantly increasing, while surgical AVR is decreasing consistently [25]. The majority of patients are afraid of undergoing (full) sternotomy, restraining them from undergoing necessary, life-prolonging procedures [12]. Johnston and Roselli [30] even found patients to postpone clearly indicated AVR operations due to ‘fear of sternotomy’.

In the interest of our patients, time has come, to actively build heart (and aortic) teams to intensify the interdisciplinary dialogue and communicate the optimal personalized therapeutic options to our patients—univocally, and unconditionally with regard to secondary economic interests. Aortic diameter remains the strongest, widely accepted predictor for the prevention of aortic events such as rupture or dissection [3] and by optimizing surgical strategies patients’ concerns regarding the operation can potentially be reduced. Accordingly, non-inferiority could become a sufficient reason to shift from conventional to minimally invasive strategies in a routine setting. Overall, it seems that smaller incisions increase the likeliness of a patient to opt for surgery, independent of proven safety criteria.

Only sufficiently powered, randomized-controlled clinical trials will provide the data required to move this complex field foreword. Mini-Stern was a promising initiative; however, surgical trials are naturally complex in design and cumbersome in execution. Nevertheless, there is no alternative and our community has to acquaint itself and initiate more large multicentric clinical trials to serve our patients adequately.

Strengths and limitations

Its retrospective design constitutes the main limitation of this study. Propensity score matching can only correct for potential selection bias in respect to available data. Thus, remaining bias due to unknown factors could have influenced the outcome comparison of our study cohort. Even though a large number of patients from an experienced high-volume centre were included and subsequently matched, due to the low event rate, some important aspects can only be answered by a multicentric approach, or ideally a large randomized controlled trial. As the surgical approach via PS was applied in more recent year, the FU data in these patients are limited. Caution is mandatory as to the general applicability of the herein presented results, in particular when adopting our strategy to low-volume centres.

However, we provide important information about the safety of a less-invasive method for AVR with concomitant scAAR applicable for experienced centres, as reflected by short CPB and cross-clamp times of both groups compared to previously published data on minimally invasive combined AVR and scAAR [16–18] and even conventional AVR with concomitant scAAR [21]. Our institution is highly specialized, with large experience in aortic and minimally invasive valve surgery. In particular, with respect to the growing popularity of catheter-based valve interventions and the rise of minimally invasive surgery in most surgical fields, clinical decision-making has to be based on evidence rather than emotionally motivated patient’s preference and economic considerations of the healthcare providers. Surgeons who wish to offer reduced invasiveness for a wider spectrum of cardiac operations should consider specializing in high-volume centres to ensure patient safety.

CONCLUSION

We demonstrated that PS is not inferior to FS regarding safety and surgical outcome of AVR with concomitant scAAR in a large volume aortic centre.

Conflict of interest: none declared.

Author contributions

Josephina Haunschild: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Visualization; Writing—original draft; Writing—review & editing. Antonia van Kampen: Methodology; Software; Visualization; Writing—original draft; Writing—review & editing. Konstantin von Aspern: Conceptualization; Formal analysis; Software; Writing—review& editing. Martin Misfeld: Supervision; Writing—review & editing. Piroze Davierwala: Conceptualization; Supervision; Writing—review & editing. Diyar Saeed: Writing—review & editing. Michael A. Borger: Conceptualization; Resources; Supervision; Writing—review & editing. Christian D. Etz: Conceptualization; Investigation; Project administration; Resources; Software; Supervision; Writing—review & editing.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Roman Gottardi, Davide Pacini, Konstantinos Tsagakis and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

Presented at the 34th Annual Meeting of the European Association for Cardio-Thoracic Surgery, Barcelona, Spain, 8–10 October 2020.

REFERENCES

ABBREVIATIONS

 
• AVR

  Aortic valve replacement

 
• CABG

  Coronary artery bypass grafting

 
• CPB

  Cardiopulmonary bypass

 
• CT

  Computed tomography

 
• FS

  Full sternotomy

 
• FU

  Follow-up

 
• ICU

  Intensive care unit

 
• PCI

  Percutaneous coronary interventions

 
• PS

  Partial upper sternotomy

 
• scAAR

  Supracommissural replacement of the ascending aorta

 
• TAVR

  Transcatheter AVR

Author notes