Effects of Preoperative Aortic Insufficiency on Outcome After Aortic Valve–Sparing Surgery
Background— The presence of significant preoperative aortic insufficiency (AI) or the need for cusp repair has been suggested as a risk factor for poorer outcomes after aortic valve (AV)–sparing surgery. We analyzed the influence of these factors on the mid-term outcomes of AV surgery.
Methods and Results— Between 1996 and 2008, 164 consecutive patients underwent elective AV-sparing surgery. Severe preoperative AI (grade ≥3+) was present in 93 patients (57%), and 54 (33%) had a bicuspid valve. Root repair was performed with either the reimplantation (74%) or the remodeling (26%) technique, and cusp repair was performed in 90 patients (55%). Mean clinical follow-up was 57 months. Hospital mortality was 0.6%. Cusp repair was required in 52% of the patients with preoperative AI ≤2+ and in 57% of those with AI ≥3+ (P=0.6). Cusp repair was required more frequently in bicuspid versus tricuspid valves (91% versus 38%, P<0.001). Overall survival at 8 years was 88±8%. Freedom from AV reoperation at 8 years was similar with preoperative AI ≤2+ versus preoperative AI ≥3+ (89±11% versus 90±7%, P=0.7) and with versus without cusp repair (84±17% versus 92±8%, P=0.5). Freedom from recurrent AI (grade ≥3+) at 5 years was also similar between groups (90±10% versus 89±8%, P=0.9, and 90±8% versus 89±9%, P=0.8, respectively). By multivariate analyses, predictors of recurrent AI ≥2+ were preoperative left ventricle end-diastolic diameter and AI >1+ on discharge echocardiography.
Conclusions— With a systematic approach to cusp assessment and repair, AV-sparing surgery for root pathology has an acceptable mid-term outcome, irrespective of preoperative AI or need for cusp repair.
The pioneering work of Yacoub et al1 and David et al2 has demonstrated that aortic valve (AV)–sparing surgery is a reliable alternative to the modified Bentall operation for patients with aortic root pathology. However, many surgeons are hesitant to perform valve-sparing surgery in patients with cusp pathology or in the presence of severe aortic insufficiency (AI). There are several reasons for this attitude. First, AV-sparing root replacement is a technically challenging procedure, and the addition of cusp repair increases the difficulty of this surgery. Second, despite increasing experience with AV repair, the long-term results of cusp repair remain uncertain. Finally, there are limited and conflicting data available to assess the effect of cusp repair on the durability of an AV-sparing procedure. Burkhart et al3 have shown a detrimental effect of cusp repair on durability, but this was not confirmed by the works of David et al2 or Schäfers et al.4
The aim of this study was to compare early and late outcomes of patients with or without severe preoperative AI who were undergoing valve-sparing aortic root replacement. The primary end points were the need for AV reoperation and the recurrence of significant AI (≥3+) at follow-up. We also sought to determine the risk factors for AI recurrence (≥2+) after valve-sparing surgery.
Subjects and Methods
From January 1996 to September 2008, 164 consecutive patients underwent elective AV-sparing root replacement surgery with or without cusp repair at a single center. In this population, we indentified 71 (43%) patients with none to grade 2+ preoperative AI and 93 (57%) with grade 3+ to 4+ preoperative AI (Table 1). Data on root and valve repair techniques were collected prospectively and analyzed retrospectively. Patients with type A aortic dissection were excluded, but patients undergoing concomitant procedures were included.
Our surgical approach to AV insufficiency and proximal aorta pathology was guided by the “repair-oriented functional classification of AI” developed in our center and already described elsewhere.5 During the early period of this study, valve sparing was performed by the remodeling technique.1 Some modifications were brought to the initially described technique. First, subcommissural annuloplasty was performed in some patients as previously described6 to reduce the width of the interleaflet triangles, increase cusp coaptation, and stabilize the ventriculoaortic junction (VAJ). Second, in patients with asymmetrical dilatation of Valsalva sinuses, partial remodeling was performed by replacement of the dilated sinuses and preservation of the nondilated one(s). Since the year 2000, the reimplantation technique (“David operation”) was performed and became our procedure of choice because of the annuloplasty and annular stabilization conferred by the reimplantation technique. Our specific technique for the reimplantation procedure has already been detailed elsewhere.7 A standard tubular graft (Gelweave, Vascutek Ltd, a Terumo Company, Renfrewshire, Scotland) or a Valsalva graft (Gelweave Valsalva, Vascutek Ltd) was used for the reimplantation technique without specific indications for 1 or the other graft. The graft size was chosen by estimating the ideal sinotubular junction (STJ) diameter. Therefore, after dissection of the root and resection of the sinuses, the 3 commissures were brought together until an optimal AV coaptation was obtained. In this position, the diameter of the STJ was measured with Hagar dilatators. In the remodeling technique, graft size corresponded to STJ size, and in the reimplantation technique, the graft was oversized by 4 to 5 mm.
Intraoperatively, the AV was systematically assessed before and after the valve-sparing procedure. After aortotomy and ostial cardioplegia, a first valve assessment was performed to confirm the possibility of valve preservation. The findings were corroborated with preoperative and intraoperative echocardiography, which were routinely performed. In the patients with preoperative eccentric AI due to cusp prolapse, the lesion was identified on the basis of the direction of the jet. For example, an eccentric jet in the direction of the anterior mitral leaflet corresponds to a prolapse of the right coronary cusp in a tricuspid valve or to a prolapse of the coronary (anterior) cusp in a bicuspid valve. Intraoperatively, cusp prolapse was identified by applying vertical traction on each commissure and inspecting the different free margin levels. A cusp was considered prolapsing when its free margin was lower in comparison with the adjacent one(s). The goal of prolapse repair was to equalize free margin levels. In cases where all cusps were prolapsing, the goal was to restore the height of coaptation to the mid-height of the sinuses of Valsalva.
The repair techniques in case of cusp prolapse or low coaptation were central cusp plication with prolene 5/0 or 6/0 or cusp resuspension with a running suture of Gore-Tex 7/0 (W.L. Gore and Associates, Inc, Newark, Del) on the free margin. Both techniques have been previously described elsewhere.8 In case of restricted cusp motion, cusp thinning (free margin or small raphe) or decalcification was used to increase the mobility of the cusp. After resection of a bicuspid valve raphe, direct closure was performed if there was enough residual cusp tissue. If not, the defect was closed with an autologous or bovine pericardial patch. In general, lesions inducing restricted valve motion were repaired after dissection of the root and resection of the sinuses but before implantation of the graft because of better access to the cusp at this stage of the procedure. On the other hand, because the reduction of both the VAJ and STJ has a direct impact on the leaflet free margin level, cusp prolapse and coaptation level were corrected typically after graft implantation but before reimplantation of the coronary ostia. When necessary, the different techniques of repair were combined. For example, after resection or repair of a raphe, a prolapsing free margin was corrected by the cusp prolapse repair technique. After reimplantion of the coronary ostia and before distal anastomosis of the graft, cardioplegia was given through the distal graft clamped around the cardioplegia tip. This maneuver allows for assessment of bleeding and tests the valve competence. Residual AI may be suspected in cases of filling of the left ventricle and can occasionally be identified by transesophageal echocardiography at this time.
After weaning from cardiopulmonary bypass, transesophageal echocardiography was performed in all patients to assess the degree of AI, orientation of the regurgitant jet (if present), and the coaptation length and level. Coaptation length of at least 5 mm at the mid-portion of the free margin and a coaptation level above the AV annulus were prerequisites for a successful repair. The presence of an eccentric residual jet or of a residual AI greater than grade 1+ were indications for reexploration of the valve.
All patients were treated with aspirin after the operation and underwent transthoracic echocardiography at discharge.
Clinical follow-up was conducted through outpatient visits or telephone follow-up by a research nurse. In addition to survival status, information on valve-related complications, including thromboembolism, hemorrhage, endocarditis, reoperation, and cardiovascular symptoms, was obtained whenever possible. Transthoracic echocardiography was performed for all patients at regular intervals during the course of follow-up. Echocardiograms not obtained at our institution were interpreted by the referring physician.
Continuous data are presented as mean±SD or median (interquartile range) for nonparametric data. Continuous variables were compared with the Student t test for normally distributed variables and the Wilcoxon rank-sum test for nonnormally distributed variables. Categorical variables were compared with the χ2 test unless the cell count was <5 cells, in which case Fisher exact test was used. Failure time data for survival, reoperation, and recurrent AI are presented on Kaplan-Meier survival curves. Exploratory univariate comparisons between groups for failure time data were performed with the log-rank test. Results were considered statistically significant at the P≤0.05 level. Statistical analyses were performed with SAS version 9.1 software. Graphs were constructed with GraphPad Prism 4.0 (GraphPad Software, Inc, San Diego, Calif).
Of the 164 patients undergoing AV-sparing surgery, 71 (43%) patients had none to grade 2+ preoperative AI (group AI ≤2+), and 93 (57%) had severe preoperative AI (group AI ≥3+). Table 1 depicts the preoperative characteristics of the study patients. In the AI ≥3+ group, patients were more symptomatic (New York Heart Association functional classification, P=0.001), had greater left ventricular end-systolic and end-diastolic diameters (P=0.001), and had greater impairments in left ventricular function (P=0.06). In the AI ≤2+ group, there were more patients with Marfan syndrome (P=0.06). The mean aortic root diameter and the presence of a bicuspid valve were similar in both groups.
Table 2 depicts the surgical data. AV reimplantation was performed in 74% (n=121) of the patients and remodeling in 26% (n=43), with 60% of those (n=26) undergoing complete root remodeling. The rates of the different valve-sparing procedures were similar in both AI groups, as was the rate of Valsalva graft use. The mean graft diameter was slightly larger in the AI ≤2+ group undergoing reimplantation, whereas for those undergoing remodelling, the mean graft size was similar in both groups. Subcommissural annuloplasty was performed in 40% (n=17) of patients undergoing the remodelling procedure and had similar rates in both groups.
The rates as well as techniques used for cusp repair were similar in both groups. The need for cusp repair, however, was significantly higher for bicuspid valves compared with tricuspid valves (P<0.001). For patients with bicuspid valves, cusp repair was performed at similar rates regardless of the degree of AI (83% of cusp repair in AI 0 to 1+, 83% in AI 2+, and 97% in AI ≥3+). In tricuspid AVs, however, the repair rate was lower in patients with AI 0 to 1+ in comparison with those with greater AI (30% of cusp repair in AI 0 to 1+, 50% in AI 2+, and 54% in AI ≥3+).
Second cardiopulmonary bypass run for residual AI was slightly more frequent in the AI ≥3+ group, but the mean cardiopulmonary bypass and aortic cross-clamp times were similar in both groups. Intraoperative echocardiographic results are shown in Table 3.
During the early postoperative period, 1 (0.6%) patient from the AI ≥3+ group died from aspiration pneumonia, and 2 patients from the AI ≥3+ group needed early AV reoperation. Rates of reoperation for bleeding (AI ≤2+, 7% versus AI ≥3+, 6.5%; P=0.9) and permanent pacemaker implantation (AI ≤2+, 2.8% versus AI ≥3+, 3.2%; P=0.8) were similar between the 2 groups.
Clinical follow-up was complete for 99% of patients, totaling 778 patient-years, with a median follow-up time of 52 months (interquartile range, 23 to 84 months). Follow-up was longer in the AI ≥3+ group (median follow-up time of 58 versus 32 months). During this period, there were 7 deaths in each group, all of which were from cardiac causes in the AI ≤2+ group and 5 were cardiac deaths in the AI ≥3+ group. Overall survival was 88±8%, and freedom from cardiac death was 91±7% at 8 years. Freedom from cardiac death was similar between both groups (AI ≤2+, 89±11% versus AI ≥3+, 91±9% at 8 years; P=0.13; Figure 1). Late AV reoperation was necessary in 8 patients for recurrence of severe AI, with no in-hospital mortality. At 8 years, overall freedom from AV reoperation was 90±7%, and freedom from AV replacement was 93±5%, as 2 patients had late AV rerepair. Freedom from AV reoperation was similar between the 2 preoperative AI groups (AI ≤2+, 89±11% versus AI ≥3+, 90±7% at 8 years; P=0.7). There were also no differences in AV reoperation between patients having an AV-sparing operation with or without cusp repair (AV-sparing alone was 92±8% versus AV-sparing+cusp repair was 84±17% at 8 years; P=0.5; Figure 2) and between patients having tricuspid versus bicuspid AV (tricuspid valve 88±8% versus bicuspid valve 91±9%; P=0.3) procedures. During follow-up, 3 patients had strokes and 1 had AV endocarditis, producing a thromboembolic, hemorrhagic, or endocarditis rate of 0.5%/patient-year. At the last follow-up, of 139 survivors with no cardiac reoperation, 119 (85%) were in New York Heart Association functional class 1, 17 (12%) were in class 2, and 3 (2%) were in class 3, with no differences between groups (P=0.5).
At discharge echocardiography, residual AI 2+ or greater was slightly but not significantly higher than AI ≥3+ (Table 3). Echocardiographic follow-up data were available for 155 patients (94%) after a median of 36 months (interquartile range, 13 to 68 months). Echocardiographic follow-up was longer in the AI ≥3+ group (median follow-up time of 44 versus 24 months). The overall freedom from AI ≥3+ was 89±7% at 5 years and 83±7% at 8 years with no difference between the 2 preoperative AI groups (AI ≤2+, 90±10% versus AI ≥3+, 89±8% at 5 years; P=0.9). No differences were observed between patients having an AV-sparing operation, with or without cusp repair (AV-sparing alone 89±9% versus AV sparing+cusp repair 90±8% at 5 years; P=0.8), and between patients having tricuspid versus bicuspid AV (tricuspid valve 85±9% versus bicuspid valve 98±2% at 5 years, P=0.3; Figure 3) The identified causes of recurrent AI in early and late AV reoperations were low coaptation or deficit of central coaptation (n=5), cusp prolapse (n=2), a technical error in valve reimplantation (n=1), endocarditis (n=1), and not indentified (n=1).
Predictors of Recurrent AI (≥2+)
An exploratory univariate analysis was performed to identify predictors of recurrent AI ≥2+ (Table 4). Preoperative left ventricular end-diastolic or systolic diameter was the strongest independent predictor of recurrent AI, followed by the presence of AI >1+ on discharge echocardiography and an ejection fraction >50% preoperatively.
Valve-sparing surgery for dilated aortic root without AV stenosis has become a reliable alternative to the Bentall operation, as both reimplantation and remodeling techniques have demonstrated durable native valve function inside the neoaortic root.1,2 The avoidance of oral anticoagulation and the prevention of prosthetic valve–related complications are the principal advantages of this technique. However, valve-sparing surgery remains a complex procedure, and concerns remain regarding postoperative valve dysfunction, particularly in patients with severe preoperative AI or an anatomically abnormal AV.
In this study, we analyzed the clinical and echocardiographic outcomes of patients with or without significant preoperative AI who were undergoing valve-sparing surgery. We found that with the use of a systematic approach to cusp assessment and repair, AV-sparing surgery had comparable mid-term outcomes in patients with none/mild versus severe preoperative AI. The outcomes were also comparable in patients undergoing AV-sparing surgery alone versus AV-sparing surgery combined with cusp repair.
The overall survival of 88% at 8 years is comparable to that reported by David et al2 and Schäfers and associates (Aicher et al9) in a large series of valve-sparing operations. The freedom from AV reoperation of 90% at 8 and 10 years is comparable to the 95% at 10 years reported by David et al2 and the 96% at 10 years reported by Schäfers and coworkers (Aicher et al9). With respect to the freedom from recurrent severe AI, the 83% at 8 years is comparable to the 85% at 10 years reported by David et al,2 and the rates of 85% for the tricuspid valve and of 98% for the bicuspid valve at 5 years are similar to the 88% and 96% at 5 years reported by Schäfers and coworkers (Aicher et al9).
In our experience, cusp repair was performed in 55% of patients, with a significantly higher repair rate in the bicuspid (90%) versus the tricuspid (47%) valve. Cusp repair was not predicted by the severity of preoperative AI, as shown by the 30% rate of repair in the group of patients with a tricuspid AV and no to grade 1+ AI preoperatively. Such frequent use of cusp repair seems related not only to our own surgical approach but also to that of David et al2 and Schäfers and coworkers (Aicher et al9), who reported equivalent or slightly greater rates. We observed a higher cusp repair rate in the reimplantation versus remodeling technique (63% versus 30%, P<0.001). This difference was also reported by David et al2 and Svensson et al10 but not by Erasmi et al,11 who reported similar rates of cusp repair in both techniques, or by Schäfers and colleagues (Aicher et al9), who reported a 60% rate of cusp prolapse repair with the remodelling technique. The differences in the rate of cusp repair between techniques may also represent an increasing use of cusp repair during the study period (1996 to 2002, 33% versus 2003 to 2008, 67%).
The predominant role of cusp repair in AV-sparing surgery is explained by the close relation between the aortic cusps and what has been termed the “functional aortic annulus” (FAA). The FAA is a 3-dimensional structure consisting of the VAJ and the STJ, which are linked together by the anatomic, crown-shaped AV annulus. This structure supports the cusps and directly influences their coaptation.5 Candidates for AV-sparing surgery usually present with central AI due to FAA dilatation, causing outward displacement of the cusp and a deficit of central coaptation. Chronic overstretching of the cusps induces their elongation,12 which, in combination with decreased coaptation, is probably the mechanism favoring the development of cusp prolapse. In AV-sparing procedures, reduction of the dilated FAA diameter has the beneficial effect of bringing the leaflets inward and increasing their coaptation, but it can also lower the level of coaptation. This level, or height, of coaptation may drop excessively, especially if the cusp free margins are elongated or if the STJ reduction is exaggerated.13,14 The resulting low coaptation has been correlated with an increased risk of recurrent AI by Pethig et al15 in a prospective study of patients having AV-sparing surgery with the reimplantation technique.
In patients with root dilatation and associated cusp prolapse, the FAA reduction usually accentuates the prolapse. Sometimes new prolapse can appear after graft implantation. One reason may be accentuation of a preexisting small and undetected prolapse that is unmasked by the sparing procedure. Another reason may be modification of the valve geometry when commissures are reattached to the graft.
The first important step to ensure good cusp coaptation is the choice of prosthesis size. No consensus exists on the method of prosthesis sizing for both reimplantation and remodelling techniques. The different methods used are based on an estimation of the STJ to give an optimal cusp coaptation,1,2,11 on the VAJ diameter,9 or on the height of the cusps.2,15 In our opinion, whatever the sizing method used, downsizing of the graft is probably worse than oversizing, especially in patients with large cusps. Effectively, downsizing results in an abnormal valve configuration with cusp billowing in the outflow tract and low coaptation.
After graft implantation, cusp repair is performed when necessary to equalize the free margin levels and optimize cusp coaptation. Our objective is to achieve a coaptation length of at least 5 mm and a coaptation height that reaches the midpoint of the sinuses of Valsalva. Central cusp plication is an easy and quick technique to repair cusp prolapse; it leaves a minimal amount of suture material on the cusp, and adjustments can easily be made with the addition or removal of plication stitches. Free margin resuspension with polytetrafluoroethylene takes more time and leaves more suture material on the cusp. However, this technique can be used to close stress fenestrations and to reinforce a free margin weakened by resection and sutures. Triangular resection and pericardial patch, essentially used for bicuspid valve repair, are susceptible to failure by suture dehiscence and patch fibrosis or calcification.16 However, satisfactory mid-term results have been reported with these techniques.17–19 The similar outcomes for bicuspid and tricuspid AVs in sparing surgery reported by Shäfers and colleagues (Aicher et al20) and confirmed in this series (Table 4) support the approach of bicuspid valve preservation, even when cusp repair is more frequent and complex than in the tricuspid AV.
The remodelling and reimplantation techniques had comparable outcomes in this study (Table 4), with the limitation that those techniques were applied during different periods and for different indications (ie, partial remodeling). Despite the excellent long-term results shown with the remodeling technique,1,9 some authors have reported a slight advantage in favor of the reimplantation technique.2,21 The reason for this may be the absence of VAJ annuloplasty in the remodeling technique, especially in patients with VAJ dilatation. With regard to preoperative AI, there is no convincing evidence to support the choice of 1 root replacement technique over another, especially when annuloplasty is added to the remodeling procedure. A well-performed valve-sparing procedure that optimizes cusp coaptation with the appropriate use of cusp repair techniques is likely more important for a successful and durable result than the actual valve-sparing technique.
The major limitation of this clinical study is that it is a retrospective, single-center study in which aortic cusp repair techniques have evolved during the past decade. Also, the presence of confounding factors, like different techniques of valves sparing, and the differences in the length of follow-up might have affected our analyses. Due to its retrospective nature, some echocardiographic parameters, like orientation of the regurgitant jet (central or eccentric), quantification of AI (eg, regurgitant volume or regurgitant orifice), and length and height of coaptation, were not systematically available on preoperative or postoperative examinations. Further echocardiographic studies may be helpful to further delineate these issues to allow better tailoring of the procedure to each patient.
In this large, single-center series, we demonstrate that with the use of a systematic approach to valve assessment and repair, AV-sparing surgery has good mid-term results, independent of the presence of preoperative AI. Late outcome is also similar in patients with or without cusp repair. At 8 years, freedom from cardiac death and freedom from AV reoperation were 91% and 90%, respectively, for the entire population. The necessity of cusp repair in AV-sparing surgery is more related to valve morphology than to preoperative AI. Therefore, the severity of preoperative AI must not be considered a contraindication for AV-sparing surgery, and cusp repair should be considered an integral component of AV-sparing surgery.
Presented in part at American Heart Association Scientific Sessions 2008, November 8–12, 2008, New Orleans, La.
El Khoury G, Vanoverschelde JL, Glineur D, Poncelet A, Verhelst R, Astarci P, Underwood MJ, Noirhomme P. Repair of aortic valve prolapse: experience with 44 patients. Eur J Cardiothorac Surg. 2004; 26: 628–633.
Boodhwani M, de Kerchove L, El Khoury G. Aortic root replacement using the reimplantation technique: tips and tricks. Interact Cardiovasc Thorac Surg. 2009; 8: 584–586.
de Kerchove L, Glineur D, Poncelet A, Boodhwani M, Rubay J, Dhoore W, Noirhomme P, El Khoury G. Repair of aortic leaflet prolapse: a ten-year experience. Eur J Cardiothorac Surg. 2008; 34: 785–791.
Thubrikar MJ, Labrosse MR, Zehr KJ, Robicsek F, Gong GG, Fowler BL. Aortic root dilatation may alter the dimensions of the valve leaflets. Eur J Cardiothorac Surg. 2005; 28: 850–855.
Maselli D, De Paulis R, Scaffa R, Weltert L, Bellisario A, Salica A, Ricci A. Sinotubular junction size affects aortic root geometry and aortic valve function in the aortic valve reimplantation procedure: an in vitro study using the Valsalva graft. Ann Thorac Surg. 2007; 84: 1214–1218.
Carr JA, Savage EB. Aortic valve repair for aortic insufficiency in adults: a contemporary review and comparison with replacement techniques. Eur J Cardiothorac Surg. 2004; 25: 6–15.
El Khoury G, Vanoverschelde JL, Glineur D, Pierard F, Verhelst RR, Rubay J, Funken JC, Watremez C, Astarci P, Lacroix V, Poncelet A, Noirhomme P. Repair of bicuspid aortic valves in patients with aortic regurgitation. Circulation. 2006; 114 (suppl I): I-610–I-616.
Casselman FP, Gillinov AM, Akhrass R, Kasirajan V, Blackstone EH, Cosgrove DM. Intermediate-term durability of bicuspid aortic valve repair for prolapsing leaflet. Eur J Cardiothorac Surg. 1999; 15: 302–308.