Valve-Sparing Aortic Root Replacement Technique: Valsalva Graft versus Two Straight Tubular Grafts

Background There are many variations in valve-sparing aortic root replacement techniques. Our aim is to determine the impact of the graft on mid-term outcomes: Valsalva graft vs. two straight tubular grafts. Methods From 2004 to 2020, 332 patients underwent valve-sparing aortic root replacement with either a Valsalva graft (Valsalva group: n = 270) or two straight tubular grafts (two-graft group: n = 62). Data were obtained through chart review and the National Death Index. Primary outcomes were mid-term survival and freedom from reoperation. Results The preoperative characteristics of the groups were similar, but the two-graft group had more type A dissections (32% vs. 19%) and emergent operations (26% vs. 15%) and was younger (45 vs. 50 years). Intraoperatively, the groups were similar, but the two-graft group had longer cross-clamp (245 vs. 215 minutes) and cardiopulmonary bypass times (284 vs. 255 minutes). Postoperative complications including reoperation for bleeding, stroke, pacemaker implantation, and renal failure were slightly more frequent in the Valsalva group, but the differences were not significant. Operative mortality was similar between the Valsalva and two-graft groups (0.7% vs. 0%). Five-year survival in the two-graft group was 100% compared to 96% in the Valsalva group (p=0.56). Five-year freedom from reoperation in the two-graft group was 100% compared to 93% in the Valsalva group (p=0.29). Conclusions The Valsalva and two-graft techniques both have excellent short- and mid-term outcomes. The two-graft technique might have slightly better survival and freedom from reoperation, but a larger sample size and longer follow-up are needed to determine if these advantages are significant.


Introduction
For just over 30 years, valve-sparing aortic root replacement (VSARR) has been performed to treat aortic root aneurysms and aortic root dissection [1]. VSARR involves the replacement of all aortic sinuses and reattachment of the coronary arteries to the root prosthesis while preserving the native aortic valve. Tis procedure is particularly attractive to young patients because they can avoid long-term anticoagulant medications associated with mechanical valve replacement [2].
Te original VSARR can be generalized into two procedures-Yacoub remodeling and David reimplantation. Te remodeling techniques orient the root prosthesis on top of the annulus, while the reimplantation techniques (David-I, IV, V, V-Smod) encapsulate the valve [3]. It is unknown which David reimplantation variation ofers the best shortand long-term competence. More specifcally, surgeons must choose between two types of Gelweave grafts (Terumo, Somerset, NJ), a Valsalva graft (David-V-Valsalva, Figure 1(a)), or two straight tubular grafts (David-V-Smod, Figure 1(b)). Te Valsalva graft used in the David reimplantation procedure aims to reproduce the sinuses of Valsalva. Many surgeons have adopted the Valsalva graft and have reported satisfactory early and mid-term results in patients with and without connective tissue disorders [4][5][6][7][8][9][10]. Te David-V-Smod procedure (two-graft technique), developed at Stanford in 2004, utilizes two straight tubular grafts: a large graft to create the neosinus and a small graft to create the sinotubular junction and replace the ascending aorta [11]. Tis technique is able to better accommodate each patient's unique aortic root anatomy and pathology, including the sizing of the neoannulus, pseudosinuses, sinotubular junction, and commissures [1]. Few studies exist assessing the clinical outcomes of this technique [12][13][14].
Tere is currently no literature directly comparing the short and mid-term clinical outcomes of patients undergoing VSARR using the David reimplantation Valsalva graft or two-graft techniques. We therefore present the perioperative and mid-term outcomes of patients undergoing David reimplantation with the Valsalva graft and two-graft techniques to treat aortic root aneurysm or acute type A aortic dissection. We hypothesize that the twostraight tubular grafts technique will provide similar short-and mid-term outcomes to the Valsalva graft technique.

Materials and Methods
Te Institutional Review Board (IRB) or equivalent ethics committee of the University of Michigan, Michigan Medicine (Ann Arbor, MI) (HUM00155922, 12/2/2020), approved the study protocol and publication of data. Patient written consent for the publication of the study data was waived by the IRB due to minimal risk to patients.

Subjects.
Between September 2004 and September 2020, 332 consecutive patients with aortic root aneurysm or acute type A aortic dissection underwent valve-sparing aortic root replacement with a Valsalva graft (n � 270) or two straight tubular grafts (n � 62). Data were obtained through the Society of Toracic Surgery Data Warehouse to identify the relevant cohort and determine the perioperative, operative, and postoperative variables. Tese data were supplemented through medical chart review including echocardiographic details for specifed variables. All derived data collected that support the fndings of this study are available by request from the corresponding author. Survival and reoperation data were collected by medical record review and National Death Index data through June 30th, 2021 [15].

Operative Technique.
Te David procedure has been performed at our institution since late 2004 and consistently since 2005. Te operative strategy at our institution for VSARR has been previously described [16]. Two modifcations of the David procedure were used based on surgeon preference: Valsalva graft or two straight tubular grafts. One surgeon exclusively used the two-graft technique, and the other surgeons exclusively used the Valsalva graft technique. Te David procedures were performed from 2004 to 2020, with 59% of these procedures performed from 2013 to 2020. Te size of the Valsalva graft is determined by the leftnoncommissural height. Six to twelve subannular stitches are placed. For the two-graft technique, the graft size of the neosinus is determined by the height of the cusps of the aortic valve, and the small graft for the neosinotubular junction reconstruction is determined by the neoannulus (basal ring of the aortic root). Eleven to thirteen subannular stitches are placed to prevent further dilation of the aortic root. Te rims of dissected aortic sinus wall are anastomosed to the inside of the new sinus graft with full-thickness bites. Two coronary buttons are reimplanted. Hemiarch or arch replacement with reimplantation of one-to four-head vessels is performed based on the arch pathology. All procedures utilize cold crystalloid or blood cardioplegia and standard cardiopulmonary bypass. Antegrade cerebral perfusion or retrograde cerebral perfusion is used for cerebral protection during hypothermia circulatory arrest. One surgeon exclusively used antegrade cerebral perfusion whereas the other two surgeons used retrograde cerebral perfusion and profound hypothermia circulatory arrest for hemiarch and antegrade cerebral perfusion for more complex arch reconstruction.

Statistical Analysis.
Continuous data were presented as median (interquartile range, 25%, 75%) and categorical data as n (%). Univariable comparisons between groups were performed using chi-square tests for categorical data. Wilcoxon rank-sum tests were performed for continuous data. Survival curves and long-term freedom from aortic insufciency were estimated using the Kaplan-Meier method with the log-rank test. Cumulative incidence function curves were adjusted for death as a competing risk using the Fine and Gray subdistribution method to assess the incidence of reoperation due to severe aortic insufciency over time. Te Gray test was used to test the diference in the cumulative incidence function curves between groups. Cox proportional hazards model was used to evaluate risk factors for late mortality adjusting for VSARR technique, age, sex, bicuspid aortic valve, prior cerebrovascular accident, prior cardiac surgery, and primary indication. Follow-up rates were determined by using Clark's completeness index. p values less than 0.05 were considered statistically signifcant. Statistical calculations were performed with SAS (SAS Institute, Cary, NC).  (Table 3).

Mid-Term
Outcomes. Te median follow-up time for mid-term survival was 6.4 years (3.2 years two-graft and 7.0 years Valsalva), and the follow-up rate for survival was 100% in both groups. Te 5-year survival rate of the whole cohort was 97% (95% confdence interval: 94%, 98%). Te 5year survival rate in the two-graft group was 100%, and in the Valsalva group was 96% ( Figure 2). Cox proportional hazards model adjusting for VSARR strategy, age, sex, bicuspid aortic valve, prior cerebrovascular accident, prior cardiac surgery, and a primary indication of operation showed that bicuspid aortic valve, prior cerebrovascular accident, and prior cardiac surgery were risk factors for late mortality (Table 4). Te median follow-up time for mid-term reoperation data was 3.3 years (1.3 years two-graft and 4.0 years Valsalva), and the follow-up rate for reoperation was 65% in the two-graft group and 79% in the Valsalva group for the study period. Te Valsalva group had a total of 13 reoperations due to severe aortic insufciency, one reoperation due to acute type A aortic dissection, and one reoperation due to aortic root prosthetic graft infection while the two-graft group had no reoperations. Te 5-year freedom from reoperation rate due to severe aortic insufciency, adjusting for death as a competing factor, was 100% in the two-graft group and 93% in the Valsalva group ( Figure 3).
Te median follow-up time for echocardiographic data was 2.5 years (1.4 years two-graft and 3.5 years Valsalva), and the follow-up rate for echocardiographic data was 65% in the two-graft group and 79% in the Valsalva group for the study period. Te 5-year freedom from aortic insufciency grade moderate or higher, which was either detected clinically and confrmed by follow-up echocardiography, or found on routine yearly echocardiographic follow-up, was 96% in the two-graft group and was 91% in the Valsalva group ( Figure 4).

Discussion
Our study is the frst to directly compare the early and midterm clinical outcomes of VSARR with David reimplantation using either the Valsalva graft or two-graft techniques. We report excellent outcomes for both techniques; there were no diferences in postoperative complications in addition to no diferences in operative, in-hospital, and 30-day mortalities between the two-graft and Valsalva groups ( Table 1). Tere were also no diferences in 5-year survival (two-graft: 100%, Valsalva: 96%), 5-year freedom from reoperation due to severe aortic insufciency adjusting death as a competing factor (two-graft: 0%, Valsalva: 93%), and 5year freedom from moderate or greater aortic insufciency (two-graft: 96%, Valsalva: 91%) (Figures 1-3).
In 2000, De Paulis and colleagues introduced a new Dacron graft-the Valsalva graft-for the David reimplantation technique. Te Valsalva graft skirt stretches in the horizontal plane to better recreate the sinotubular junction for a more natural aortic root anatomy. De Paulis et al. and colleagues reported exceptional early postoperative results and no operative deaths [17], and our study showed similar perioperative results (Table 3). In 2002, they utilized the Valsalva graft in the Bentall, reimplantation, and remodeling procedures and noted the Valsalva reimplantation technique decreased the risk of bleeding and stabilized the aortic wall and annulus [18]. Since De Paulis and colleagues' pioneering work, the majority of surgeons have adopted this VSARR technique to treat aortic root disease if the native aortic valve is salvageable. Te Valsalva graft sinus distensibility has been shown to be appropriately preserved, and its valve-opening characteristics were consistent with normal valve function [19]. Excellent early outcomes in patients with and without Marfan syndrome have been reported [6][7][8][9][10]. Te midterm outcomes of VSARR with a Valsalva graft in our study were similar to those found in other large centers, such as the fndings of a multicenter Italian study from 2006 which showed a 5-year freedom from moderate or higher aortic insufciency of 89% [6], compared to 91% in our study.
An additional David reimplantation technique, although far less commonly used due to its steep learning curve and unforgiving nature, is the David-V-Smod. Te Stanford group reasoned that this technique allows the aortic valve to be more easily reimplanted within the larger graft, which   recreates the neosinuses. Subsequently, the distal end of the smaller graft, which makes up the sinotubular junction and recreates the proximal ascending aorta, can be better matched to the diameter of the aorta than in the Valsalva reimplantation technique [10]. Te advantage of the twograft technique is that each implantation of the aortic root is tailored to each individual patient's specifc anatomy instead of a one-size fts all approach. For example, the heights of the various commissural posts can be diferent, and sometimes, one commissural post can be as much as 1-2 cm higher than the other two (Figures 5(a) and 5(b)). It is easier to utilize the two-graft technique to accommodate this uneven commissural height (    A common determinant for reoperation of VSARR procedures is the postoperative progression of aortic insufciency after the initial VSARR surgery. Tus, we found that the David-V may ofer better hemodynamics as we report a 5-year freedom from moderate or severe aortic insufciency of 96% and 91% in the two-graft and Valsalva groups, respectively, although this diference was not signifcant ( Figure 4). Te Valsalva graft has been efective in patients with and without Marfan syndrome; the reported 10-year freedom from moderate or severe aortic insufciency in patients with Marfan syndrome is 93.5% and without Marfan syndrome is 87.1% [5]. Tese data are lacking for the David-V-Smod procedure.
Reoperation between both techniques was also not diferent in our study, confrming previous reports. Similar to our results for the Valsalva group, studies have reported the 5-year freedom from reoperation due to aortic insufciency ranges from 89-91%, and De Paulis et al. reported    an overall 10-year freedom from reoperation of 91% [4][5][6][7] ( Figure 3). It seems that the David-V-Smod may ofer slightly better reoperation rates, however, as the Hospital Universitaria 12 Octubre group reported a 5-year freedom from reoperation of 96% [12], and Kvitting et al. reported a 10-year freedom from reoperation for all causes of 92.2% [13]. Te 5-year freedom from reoperation rates due to severe aortic insufciency for the David-V-Smod group was 100% in our study ( Figure 3). Overall, our fndings are consistent with the literature and demonstrate the efcacy of both procedures and their successful implementation at diferent institutions.
On average, patients undergoing the David-V-Smod were on cardiopulmonary bypass and cross-clamped 29 and 30 minutes longer, respectively, than those who received the Valsalva reimplantation. Tis was due to more complex operations in the two-graft group, including more acute type A dissection repair and more aggressive arch replacement. Although short-and mid-term results are excellent for both the Valsalva reimplantation and David-V-Smod, there is far less experimental and clinical evidence for the two-graft technique; however, our study demonstrated that the twograft technique may have exhibited mildly better survival, reoperation, and lower incidence of moderate or severe aortic insufciency. Tus, a longer follow-up is required to determine whether this diference is signifcant in the long term.

Limitations.
Tis study is limited as a retrospective cohort study-the choice of Valsalva graft versus two straight tubular grafts was by surgeon preference and not randomized, creating some diferences between the groups. Because of the relatively low sample size of the two-graft group, it was difcult to control for some important preoperative factors including age and incidence of acute type A aortic dissection as the primary indication. Tat being said, the fact that the two-graft group had signifcantly more aortic dissection patients than the Valsalva group, while still having similar midterm outcomes, shows that at the very least, the two-graft technique can be safely used in a more high-risk set of patients with comparable results to the Valsalva graft. Te National Death Index database was combined with a manual chart review to assess for long-term mortality, but this method may not catch 100% of deaths.
Our institution aims to conduct follow-up yearly echocardiograms in aortic valve replacement patients, but we are not able to get full echocardiographic follow-up data for all patients. Lastly, although our institution is one of the few tertiary health centers in its area and we therefore expect most of our patients requiring reoperation would return to our institution, some patients may have had reoperations at outside hospitals, which could have caused us to underestimate reoperation rates.

Conclusion
Te mid-term clinical outcomes of both techniques of valvesparing aortic root replacement with David reimplantation are excellent, including survival and freedom from reoperation. Long-term results are needed to determine how the newer two-graft technique compares to the Valsalva graft technique.

Data Availability
Raw data were generated at Michigan Medicine, University of Michigan, Ann Arbor, MI. Derived data supporting the fndings of this study are available from the corresponding author on request.

Disclosure
Alexander Makkinejad and Bailey Brown are the co-frst authors.

Conflicts of Interest
Te authors declare that they have no conficts of interest.