Midterm durability of valve-sparing root replacement in bicuspid and tricuspid aortic valves

Abstract

Background

Among patients with bicuspid aortic valves (BAV) who are potential candidates for valve-sparing root replacement (VSRR), the long-term durability of this technique is not well understood. This study aimed to compare the clinical and echocardiographic outcomes of VSRR in those with BAV and tricuspid aortic valve (TAV) morphology.

Methods

This was a retrospective analysis of patients who underwent VSRR between 2007 and 2021 at a single center. Kaplan-Meier and log-rank analysis were used to estimate and compare freedom from mortality, progression to >2+ aortic insufficiency (AI), and reoperation between groups (BAV vs TAV). Preoperative and postoperative echocardiographic data were collected and assessed for temporal changes in mixed-effect models.

Results

A total of 185 patients (BAV, n = 52, 28.1%; TAV, n = 133, 71.9%) underwent VSRR. At baseline, BAV patients were younger (42.4 ± 11.6 vs 52.3 ± 12.6 years; P < 0.01) and had more severe AI (47.9% vs 27.0%; P = 0.02). Average cardiopulmonary bypass and cross-clamp times were similar. There were no differences in rates of postoperative complications, intensive care unit or hospital days, or 30-day readmission. TAV patients’ 1-, 5-, and 8-year survival rates were 99.2% [95% CI 97.8–100], 96.7% [93.5–99.9%], and 92.2% [85.6–99.3%], respectively. Overall, there were no differences between groups regarding freedom from mortality (P = 0.18), reoperation (P = 0.51), or recurrent >2+ AI (P = 0.97). 

Conclusions

VSRR can be safely performed on patients with BAV and TAV morphology, yielding similar midterm freedom from mortality, recurrent >2+ AI, and reoperation. This technique should be considered in carefully selected patients with aortic root pathology and BAV anatomy when performed at experienced centers.

Traditionally, the fastest and safest way to surgically treat aortic root pathology with or without concomitant aortic valve pathology is to replace it with a composite graft. This is known as the Bentall procedure, first described in 1968 by Drs. Hugh Bentall and De Bono.¹ However, use of a prosthesis comes with several considerations, including risk of thromboembolism, hemorrhage, infection, need for lifelong anticoagulation in mechanical valves, and the eventual need for reoperation for bioprosthetic degeneration.² The valve-sparing root replacement (VSRR) technique was first described by Dr. Tirone David in 1989 and consists of reimplanting the aortic valve into a conduit.^2,3 This technique preserves patients’ native valve, reinforces the ventricular-aortic junction, and spares patients the risks associated with prosthetic valves.

Bicuspid aortic valve (BAV) and aortic root disease presents earlier in life and is associated with distortion of the aortic root geometry as well as leaflet pathology compared to tricuspid aortic valve (TAV). It is common practice for patients with suboptimal aortic valve and root anatomy to undergo the Bentall procedure. However, centers of excellence have reported outcomes suggesting that VSRR is a reasonable option for patients with BAV morphology, with excellent long-term durability of repair and freedom from recurrent valve dysfunction.^4,5 The overall robustness of data suggesting the widespread applicability and replicability of such results is still lacking. The purpose of this study was to compare the midterm follow-up, durability of repair, and freedom from reoperation in BAV and TAV patients who underwent VSRR at a single institution.

METHODS

Patient selection

This was a single-center, retrospective review of all consecutive patients who underwent a VSRR from January 2007 to December 2021 at the Baylor Scott & White Heart Hospital in Plano, Texas. Patients were identified using the institutional Society of Thoracic Surgeons (STS) Adult Cardiac Surgery Database. This study protocol was approved by the Baylor Scott & White Institutional Review Board under protocol #020-331. Patients were consented over the phone with a scripted consent form prior to obtaining follow-up information.

All patients who underwent VSRR were identified. Patient demographics, baseline characteristics, operative details, and outcomes were obtained from the STS database and supplemented by manual chart review in cases of missing data. The primary endpoint of the study was progression to >2+ aortic insufficiency (AI) after VSRR and reoperation. Secondary outcomes included all-cause mortality, cardiopulmonary bypass time, cross-clamp time, postoperative stroke, new renal replacement therapy, prolonged ventilator requirements (defined by the STS as >24 hours), blood product transfusion, new pacemaker requirement, incidence of postoperative atrial fibrillation prior to discharge, reoperation unrelated to VSRR, intensive care unit and hospital length of stay, in-hospital mortality, discharge disposition, 30-day readmission, and reoperation. The morphology of BAV as classified by Dr. Hans-H Sievers⁶ was obtained from operative dictations.

As part of our patient selection, we typically exclude patients with aortic stenosis or more than mild leaflet calcification and valves with multiple fenestrations. Ultimately, the decision to proceed with reimplantation is decided by the surgeon intraoperatively after careful inspection of valve characteristics including commissural relationship, leaflet geometric height, and leaflet effective height. Prior to surgery, the mechanical or bioprosthetic alternative option is discussed and decided with the patient should the patient’s valve be irreparable and the patient requires a Bentall procedure.

Long-term survival status was determined via phone call to patients or by review of electronic medical records and online obituary searches according to a previously described and validated protocol.⁷ Patients without a confirmed date of death were censored as “alive” at the time of their last documented clinical contact.

Echocardiographic assessment

Echocardiographic parameters were obtained from available reports at the following time points: within 2 months prior to surgical intervention (preoperative), <3 months postoperatively (early), 6 to 18 months postoperatively (midterm), and >24 months postoperatively (long-term). Left ventricular end-diastolic diameter, aortic valve mean gradient, aortic valve peak gradient, and severity of AI were obtained and evaluated for temporal changes between the two groups over the study period.

Statistical analysis

Analysis on patients undergoing VSRR was performed comparing outcomes of patients with BAV and TAV anatomy. Descriptive statistics including mean ± standard deviation and frequency (%) were used to summarize continuous and categorical variables, respectively. Differences in continuous patient characteristics between groups were assessed with Student’s t test or Wilcoxon rank-sum test, as appropriate. Chi-square or Fisher’s exact tests were used to determine differences in baseline proportions between study groups. Categorical variables were compared using either χ2 or Fisher exact test as appropriate. A Kaplan-Meier plot and log-rank test were used to estimate and compare the long-term survival in both groups. We used mixed effects models to examine the changes in echocardiographic markers over time and whether those changes were affected by the surgery type (BAV or TAV). The latter was investigated by adding an interaction term between surgery type and time. A compound symmetry working correlation matrix was assumed to capture the correlation structure of the repeated outcomes. We then used a multivariate mixed-effects model to compare the changes in each echocardiographic marker between the study groups, having adjusted for baseline characteristics and comorbidities. A backward selection method with a selection criterion of α = 0.05 was used to retain prognostic factors that were associated with the outcome. Due to the clinical importance of age, patient age at the time of surgery was included in all models regardless of its significance level. Statistical analysis was performed in RStudio (Boston, MA).

RESULTS

Patients

There were 185 patients who underwent a VSRR between 2007 and 2021, of whom 52 (28.1%) had BAV and 133 (71.9%) had TAV anatomy. Most patients were male (86%; 159/185). Patients with BAV tended to be younger (42.4 ± 11.6 vs 52.3 ± 12.6 years; P < 0.01) and had a lower incidence of hypertension (48.1% vs 72.2%; P < 0.01). Other comorbidities were similar between the two groups. Overall, 50.3% (93/185) of the entire study cohort had moderate or severe AI, with BAV patients tending to have more severe preoperative AI compared to TAV patients (46.2% vs 23.3%; P = 0.02). Among patients with BAV who underwent VSRR, none had Marfan’s syndrome (BAV 0% vs TAV 16.7%; P < 0.01). There were more patients in the TAV group who had a history of prior aortic surgeries (BAV 0% vs TAV 9.0%; P = 0.02). More patients with BAV morphology underwent aortic valve repair compared with TAV patients at the time of their VSRR (11.3% vs 1.5%; P < 0.01). The demographics and baseline comorbidities of these patients are detailed in Table 1.

Table 1.

Demographic characteristics and comorbidities of patients undergoing valve-sparing root replacement with bicuspid and tricuspid aortic valve

	All (n = 185)	Bicuspid AV (n = 52)	Tricuspid AV (n = 133)	P value
Age (y)	49.5 ± 13.1	42.4 ± 11.6	52.3 ± 12.6	<0.01
Male	159 (86%)	46 (88.5%)	113 (85%)	0.70
Body mass index (kg/m2)	29.1 ± 5.4	28.3 ± 3.6	29.4 ± 6.0	0.19
Hypertension	121 (65.4%)	25 (48.1%)	96 (72.2%)	<0.01
Diabetes	16 (8.7%)	2 (3.8%)	14 (10.6%)	0.24
Chronic lung disease	8 (4.3%)	1 (1.9%)	7 (5.3%)	0.44
History of smoking	64 (34.6%)	14 (26.9%)	50 (37.6%)	0.23
Arrhythmia				0.22
Continuous/persistent	7 (3.8%)	0 (0%)	7 (5.3%)
Paroxysmal	9 (4.9%)	3 (5.8%)	6 (4.5%)
Remote	6 (3.2%)	2 (1.1%)	4 (3.0%)
History of CVA	10 (5.4%)	0 (0%)	10 (7.6%)	0.06
Previous MI	3 (1.6%)	0 (0%)	3 (2.3%)	0.56
Cirrhosis	1 (0.5%)	0 (0%)	1 (0.8%)	1.00
Aortic insufficiency
Trace/trivial	26 (16%)	5 (10.4%)	21 (18.3%)	0.25
Mild	44 (27%)	10 (20.8%)	34 (29.6%)	0.34
Moderate	39 (23.9%)	10 (20.8%)	29 (25.2%)	0.69
Severe	54 (33.1%)	24 (47.9%)	31 (27.0%)	0.02
Marfan syndrome	22 (12.0%)	0 (0%)	22 (16.7%)	<0.01
Coarctation of the aorta	2 (1.1%)	1 (1.9%)	1 (0.8%)	0.49
Turner syndrome	1 (0.5%)	1 (1.9%)	0 (0%)	0.28
Type A aortic dissection	4 (2.2%)	0 (0%)	12 (9%)	0.02
Prior aortic surgery	12 (6.5%)	0 (0%)	12 (9%)	0.02
Concomitant surgeries
Aortic valve repair	40 (21.5%)	33 (62.2%)	7 (5.3%)	<0.01
CABG	6 (3.2%)	0 (0%)	6 (4.5%)	0.12
MAZE	2 (1.1%)	0 (0%)	2 (1.5%)	0.37
Mitral valve surgery	3 (1.6%)	0 (0%)	3 (2.3%)	0.27

All values expressed either as number (%) or mean ± standard deviation.

AV indicates aortic valve; CABG, coronary artery bypass grafting; CVA, cerebrovascular accident; MI, myocardial infarction.

Operative details

Out of the 53 BAV patients undergoing VSRR, the predominant morphology was Sievers Type 1 left-right (73.1%), followed by Type 0 (13.2%), Type 1 right-noncoronary (9.4%), and least commonly Type 2 (1.9%). The Sievers classification of the aortic valve was not ascertainable on two patients with BAV. Significantly more patients with BAV underwent aortic valve repair compared to TAV (62.3% vs 5.3%; P < 0.01). The most common technique for repair was leaflet plication of the free margin, which was performed in 75.8% (25/33) of repairs. Less frequently performed repair techniques were commissuroplasty (15%) and leaflet repair or resection (9.1%) (Figure 1). The median [interquartile range] cardiopulmonary bypass time (215.5 [194.0, 242.3] vs 215 [198.0, 241.0] min; P = 0.94) and cross-clamp time (187 [168.8, 206.0] vs 188.5 [172.0, 209.0] min; P = 0.10) were similar between groups.

Figure 1.

Bicuspid aortic valve repair by plication of prolapsed leaflet.

Postoperative outcomes

There was no difference in the rate of postoperative stroke, renal replacement therapy for renal failure, prolonged ventilator requirements, blood product transfusion, or new pacemaker implantation between the two groups. Excluding 4 BAV and 13 TAV patients with preexisting paroxysmal or continuous/persistent atrial fibrillation, the incidence of new-onset postoperative atrial fibrillation was 17.3% (9/53) in patients with BAV and 25.8% (34/133) in patients with TAV (P = 0.25). There were 3 (5.8%) BAV patients who had a reoperation for bleeding vs 1 (0.8%) TAV patient (P = 0.04) during the perioperative period. There was one in-hospital death in a TAV patient on postoperative day 10 secondary to significant hemorrhage. No perioperative mortalities were observed in the BAV group. There were no differences in intensive care unit and hospital length of stay or 30-day readmission between the two groups. There were 11 (6%) patients who were readmitted within 30 days of discharge. The reasons for readmission were the following: 6 for arrhythmias, 1 for fever, 1 for cerebrovascular accident, 1 for pericardial effusion, 1 for hematuria in the setting of renal cell cancer history, and 1 with both atrial flutter and a cerebrovascular accident (Table 2).

Table 2.

Outcomes after valve-sparing root replacement

Outcome	All (n = 186)	Bicuspid AV (n = 53)	Tricuspid AV (n = 133)	P value
CPB time (min)	214.5 [201.5, 229.5]	215.5 [194.0, 242.3]	215 [198.0, 241.0]	0.94
Cross-clamp time (min)	190 [179.8, 213.5]	187 [168.8, 206.0]	188.5 [172.0, 209.0]	0.10
Stroke	2 (1.1%)	1 (1.9%)	1 (0.8%)	0.49
Renal replacement therapy	1 (0.5%)	0 (0%)	1 (0.8%)	1.0
Prolonged ventilator requirements	7 (3.8%)	3 (5.8%)	4 (3.0%)	0.41
Blood product transfusion	30 (16.2%)	8 (15.4%)	22 (16.5%)	1.0
New pacemaker requirement	2 (1.1%)	1 (1.9%)	1 (0.8%)	0.50
Atrial fibrillation	43 (23.4%)	9 (17.3%)	34 (25.8%)	0.25
Reoperation	5 (2.7%)	4 (7.7%)	1 (0.8%)	0.02
Bleeding	4 (2.2%)	3 (5.8%)	1 (0.8%)	0.07
Valve dysfunction	1 (0.5%)	1 (1.9%)	0 (0%)	0.28
ICU LOS (h)	43.0 [29.5,56.0]	41.6 [29.6,55.5]	48.8 [37.2,64.5]	0.14
Hospital LOS (days)	5 [4.0,6.3]	6 [5.0,7.0]	5 [5.0, 7.0]	<0.01
In-hospital mortality	1 (0.5%)	0 (0%)	1 (0.8%)	1.0
Discharged home	180 (97.8%)	52 (100%)	128 (97%)	0.48
30-day readmission	11 (6.0%)	5 (9.6%)	6 (4.5%)	0.30

Categorical variables expressed as number (%). Continuous variables expressed as mean ± standard deviation or median [interquartile range] as appropriate.

CPB indicates cardiopulmonary bypass; ICU, intensive care unit; LOS, length of stay.

In our cohort, the Kaplan-Meier estimated survival of patients undergoing VSRR with TAV morphology was 96.7% [interquartile range (IQR) 93.5–99.9%] at 5 years, and there have been no deaths up to 5 years in patients with BAV morphology. The overall median length of follow up was 5.4 [IQR 3.0,7.9] years (Figure 2a).

Figure 2.

Kaplan-Meier estimates of patients with bicuspid aortic valve (BAV) and tricuspid aortic valve (TAV) undergoing valve-sparing root replacement looking at (a) long-term survival, (b) freedom from reoperation, and (c) freedom from recurrence of >2+ aortic insufficiency. These variables were similar between the two aortic valve morphologies in the long term. Shaded areas represent 95% confidence intervals.

Freedom from moderate/severe aortic insufficiency and echocardiographic findings

All patients had intraoperative evaluation of their valve with transesophageal echocardiography before and after valve reimplantation. Temporal trends of echocardiographic variables are shown in Figure 3. It is notable that although the overall mean and peak gradients are higher in the BAV group, the median mean gradient for BAV patients was 6 [IQR 4–10.5] mm Hg and the peak gradient was 10 [IQR 8–18] mm Hg in the long term after VSRR in the cohort of patients.

Figure 3.

Temporal trends in echocardiographic variables in patients who underwent valve-sparing root replacement with bicuspid aortic valve (BAV) and tricuspid aortic valve (TAV) morphology. Patients who underwent reoperation were not included. Preoperative is defined as echocardiograms obtained within 2 months prior to surgery; early is within 3 months postoperatively; midterm is 6 to 18 months postoperatively; and long-term is >24 months postoperatively. (a) The number of patients with varying grades of AI from the preoperative to postoperative period. (b–d) The median values and interquartile ranges of the aortic valve (AV) (b) mean gradient, (c) peak gradient, and (d) left ventricular end diastolic diameters are represented by box and whiskers plots. Although the median, mean, and peak gradient in BAV patients looks slightly higher overall, the differences are not clinically significant.

A total of 9 (4.9%) patients in the cohort developed recurrent >2+ AI, 7 of which have undergone reintervention (2 BAV, 7 TAV). There were 6 (3.2%) patients in the TAV group who had a reintervention related to the VSRR. One of these patients had an AV repair at their index operation. The time to reintervention ranged from 0.7 to 9.1 years. All six patients had severe AI and underwent bioprosthetic aortic valve replacement. One who underwent reoperation <2 years after initial VSRR had poor coaptation due to leaflet tethering. Two patients who developed recurrent AI >4 years after their initial surgery were noted to have fenestrated leaflets on their reoperation, and three patients were found to have leaflet prolapse on reoperation. The patient with BAV who underwent reintervention 4.5 years from the index operation had a Sievers Type 0 morphology that underwent leaflet plication at the index operation and was noted to have calcifications on his leaflets inhibiting coaptation on the reoperation. Two patients had echocardiographic evidence of >2+ AI but have not undergone reoperation. One patient was young, with a Sievers 1 left-right BAV, and underwent VSRR for acute aortic dissection, and the other patient had a TAV with some leaflet fenestrations that were repaired at the time of the index surgery. Both of these patients are symptom free. The Kaplan-Meier estimated freedom from recurrence of recurrent >2+ AI (Figure 2b, P = 0.97) and reoperation (Figure 2c, P = 0.51) was similar between patients with BAV and TAV.

DISCUSSION

The patient’s native aortic valve anatomy influences surgical management of aortic root pathology. The VSRR affords patients the benefits of preserving their native aortic valve while avoiding the risks of prosthetic implantation. Conventionally, this procedure is offered to patients with normal tricuspid morphology. Aortic root pathology and valvular disease typically present earlier in BAV patients; therefore, the long-term benefits of the VSRR technique for BAV patients are particularly relevant. In this study, we found that patients with BAV undergoing VSRR have equivalent midterm survival, freedom from reoperation, and freedom from >2+ recurrent AI compared to patients with TAV.

Early on, a study by Aicher et al showed that VSRR could be performed in BAV patients without compromising circulatory arrest or ischemia times.⁸ More recent studies looking at midterm follow up after VSRR in these patients have demonstrating promising results, although there is variability in these results. A study by Bavaria et al⁹ propensity matching patients with BAV and TAV undergoing VSRR showed that although patients with BAV morphology presented with worse preoperative AI and more frequently required leaflet repair, the incidence of postoperative mortality, stroke, pacemaker implantation, and reoperation was similar to patients with TAV morphology. In addition, there was evidence of left ventricular remodeling in BAV patients on echocardiographic follow-up. The rate of perioperative complications and low mortality was also similar between BAV and TAV patients. In a similar study, Ouzanian et al published the results of their series of 333 patients and reported that although the probability of AI increased over time in both cohorts, there was no significant difference in mortality, reoperation, and valve-related events.⁴ In contradistinction, a study by Mokashi et al showed that in a propensity-matched study of BAV and TAV patients, BAV patients had greater progression of AI and less freedom from reoperation after 5 years.¹⁰ In our series, patients with BAV also had a higher incidence of moderate to severe preoperative AI at baseline, and the midterm survival, reoperation, and recurrent moderate to severe AI was similar to that of TAV patients. It is notable that our BAV cohort had a higher median aortic valve mean and peak gradient across the time period of the study; however, even in the long term, the degree of stenosis after VSRR was mild and did not suggest clinically significant changes in the patients. Further studies are needed to evaluate these findings and their progression in the long term.

In this current study, the rate of aortic valve repair was significantly higher in patients with BAV who underwent VSRR compared to patients with TAV. Other studies have reported similar findings, with about 60% to 70% of BAV patients requiring concomitant repair.^4,11 The most commonly used technique for repair was plication of the leaflet-free margin. A study by El-Khoury et al described the importance of root deconstruction with annular stabilization, attention to leaflet geometric and effective heights, and commissural anatomy during reimplantation. In their described technique, the goal in BAV morphology is to increase leaflet mobility by trimming or resecting the raphe and promoting “180°/180°” alignment of the commissures by tailoring the annulus with selective annuloplasty.^11,12 Consensus among experienced surgeons during discussion of a recent study from the Cleveland Clinic was also that patients with Sievers Type 0 BAV anatomy had higher chances of successful and durable reimplantation.⁵ In our series, the majority of BAV patients had Sievers Type 1 left-right morphology, with only 13% of BAV patients having Sievers Type 0 valves. The technique at this institution involves spacing commissures to evenly distribute leaflet tension through leaflet repair rather than aiming for perfect commissural symmetry. Our results suggest that at least in the midterm, recurrence of AI in BAV patients remains low with this strategy.

Limitations

This study is subject to limitations of a retrospective analysis from a single center. Another limitation of this study is the median follow-up of 5 years. Indeed, we intend to pursue longer-term follow up in our own series. However, our follow-up is similar to that reported by other similar series.^5,9 VSRR procedures were performed by a group of six surgeons whose intraoperative technique varied slightly. The primary difference was between graft sizing for the root reconstruction. Some surgeons prefer to use a Valsalva graft to reimplant the valve, while others prefer a method of oversizing the conduit for the root and buttressing the graft to create the geometry of the neosinuses. The authors believe that this particular variation in technique influences the long-term outcomes of the VSRR, as overarching principles of root stabilization and leaflet geometric and effective height in valve reimplantation were consistent. In addition, standard-of-care echocardiography was performed and interpreted by different cardiologists and not overread by a core laboratory. Verification of the measurements was not possible in the first decade of our experience due to lack of access to raw echocardiographic images. This limits the comparison and interpretation of echocardiographic variables, particularly in the long-term. Follow-up information for patients referred from out-of-state institutions was not entirely complete due to lack of access to electronic medical records or outdated contact information.

CONCLUSION

The consideration of VSRR in BAV patients remains relatively novel and arguably controversial. This study adds to the growing evidence that VSRR in patients with BAV morphology yields outcomes comparable to those of VSRR for TAV in the early and midterm. Our findings suggest that this technique is applicable to well-selected patients, sparing patients the respective long-term risks associated with replacing the aortic valve with a mechanical or biologic prosthesis. Further studies from centers of excellence providing data in the immediate, midterm, and long-term periods are necessary to establish the viability of the VSRR as an option for BAV patients that achieves long-term durability.

Funding Statement

Data acquisition efforts were funded by a philanthropic gift from Satish and Yasmin Gupta to Baylor Scott & White The Heart Hospital, Plano.

DISCLOSURE STATEMENT

The authors have no conflicts of interest to disclose. Data acquisition efforts were funded by a philanthropic gift from Satish and Yasmin Gupta to Baylor Scott & White The Heart Hospital, Plano.