Is small cusp size a limitation for aortic valve repair?^†

Abstract

Open in new tabDownload slide

OBJECTIVES

We sought to investigate cusp size limitations for valve repair in patients with aortic regurgitation (AR).

METHODS

Preoperative computed tomography was performed in 105 patients. Cusp geometric height (GH) and annulus size were measured. Mean patient age was 60.7 ± 13.7 years. Mean GH of 3 cusps was used in the analysis. Annulus cusp mismatch was graded using predicted coaptation length. Patients were categorized by mean GH into group S (GH <16 mm; n = 35) or L (GH ≥16 mm; n = 70).

RESULTS

Preoperative mean GH was 17.1 ± 2.3 mm. GH and body height were significantly correlated (r = 0.61). Intraoperative mean GH (18.8 ± 2.2 mm) was larger than preoperative mean GH (P < 0.0001). However, postoperative (17.1 ± 2.0 mm) and preoperative mean GH did not differ. Moderate AR was not present on predischarge echocardiography. Mild AR was observed in 51% and 17% of patients in groups S and L, respectively (P = 0.006). During follow-up, moderate or severe AR was observed in 14% and 10% of patients in groups S and L, respectively (P = 0.74). Two patients in group S required reoperation for a regurgitant valve. Twenty (83%) and 15 (21%) patients in groups S and L, respectively, had severe annulus cusp mismatch before surgery. Annulus cusp mismatch resolved in most patients in group L postoperatively, whereas more than half the patients in group S still had severe mismatch.

CONCLUSIONS

Small cusp size (GH <16 mm) is not necessarily a contraindication in aortic valve repair. However, most patients in this group had annulus cusp mismatch. Root replacement or secure annulus plication is mandatory to correct annulus cusp mismatch.

INTRODUCTION

Deep coaptation is essential for successful valve repair. Aortic regurgitation (AR) relates to dilatation of the aortic root. A large cusp size ensures sufficient coaptation, whereas small cusps may not provide an adequate valve area for coaptation. Cusps may be retracted or merely correspond to small body size. Theoretically, matching an annulus or root according to cusp size is a practical and simple solution. Nonetheless, there is no established matching method. In clinical practice, valve repair is based on average cusp and annulus size of a population. Schäfers et al. [1] proposed that geometric height (GH) measured intraoperatively is useful for evaluation of cusp size, and might serve as the basis for decision-making in aortic valve repair. Determining GH in advance enables accurate operative planning.

In clinical practice, 3-dimensional, high-resolution multidetector computed tomography (MDCT) can be used to assess heart valves. Transcatheter aortic valve implantation (TAVI) is a minimally invasive therapeutic option for patients with aortic valve stenosis. A pre-TAVI evaluation requires precise measurement of the aortic root and aortic valve. However, use of MDCT for valve repair is still not yet common. Successful valve repair is thought to depend on surgical skill as echocardiographic images of valve leaflets or cusps are not always clear.

Here, we evaluated the aortic valves of patients with AR using preoperative CT to identify cusp size limitation for valve repair.

MATERIALS AND METHODS

Since 2011, preoperative CT evaluation of the aortic valve has been routinely performed in our hospital. In the current study, cusp GH and annulus size were measured in 105 patients who underwent aortic valve repair surgery. Patients with a bicuspid valve were excluded. The mean patient age was 60.7 ± 13.7 years (ranged 19–77 years); of them, 79 were men. Mean body height was 164.9 ± 9.9 cm and mean body surface area (BSA) was 1.69 ± 0.20 m². According to Boodhwani’s [2] system, 54 valves were classified as type I (annulus dilatation), while the other 51 valves were type II (prolapse). The diagnosis of prolapse was made when a distinct difference in effective height (EH) was observed between the diseased cusp and the others. The belly of the cusp is usually bent and accompanied by an eccentric regurgitant jet. Making the diagnosis was feasible on high-resolution CT. Cases of type I root dilatation associated with a prolapse lesion were classified as type II. Prolapse was observed in 38 right coronary, 12 left coronary and 11 non-coronary cusps. In 2 patients, all cusps were prolapsed.

Computed tomography measurements

The aortic valve and aortic root were assessed using MDCT with a 128 × 2-slice dual-source scanner (Somatom Definition Flash, Siemens, Germany). Data acquisition was gated to the electrocardiogram to allow retrospective analysis and reconstruction at each of 10% of the RR intervals of the cardiac cycle. The data were analysed using a workstation (Syngo.Via, Siemens). End-diastolic images were usually selected for analysis, as cusp immobility is expected at this phase. The GH was measured from the nadir of the sinus to the centre of the free margin. In cases of a prolapsed cusp, the bent portion was traced with an electronic pen (Fig. 1). The average GH of 3 cusps was included in the analysis. The basal annulus area was measured in a transverse plane crossing the nadirs of the aortic valve. The dimensions of the sinotubular junction and the widest portion of the Valsalva sinuses were also measured. Each diameter was calculated using a mathematical formula (area = π × r²). In this study, the term ‘annulus’ represents the basal annulus diameter measured on CT. Before discharge, 88 patients underwent CT for aortic valve reassessment.

Prediction of coaptation length

Assumptions for the calculation include uniform-sized cusps meeting at the centre of a circle.

Annulus cusp mismatch

Annulus cusp mismatch signifies a relative shortage of cusp size for the aortic root size. In this study, mismatch was graded according to predicted CL. A predicted CL <2 mm indicated severe mismatch, while 2–4 mm indicated mild mismatch. No mismatch was present when the predicted CL was >4 mm.

Surgical management

All operations were performed via a median sternotomy, with standard cardiopulmonary bypass, in mild hypothermia. Myocardial protection was accomplished with antegrade tepid blood cardioplegia and maintained with antegrade direct coronary reinfusion. The ascending aorta was opened transversely 1 cm above the top of the commissures. GH was measured intraoperatively. The commissures were placed under radial tension using stay sutures. The free margin of the cusp to be measured was held with a forceps and placed under tension. The GH was measured using a ruler from the aortic insertion in the nadir to the nodules of Arantius of the respective cusp [1].

The repair technique for functional aortic annulus plication included basal annuloplasty with a polytetrafluoroethylene suture [3] (n = 13) or Teflon band (n = 11), a valve-sparing root replacement procedure with reimplantation (n = 56) or remodelling (n = 5) and subcommissural annuloplasty (n = 23) or annuloplasty at the sinotubular junction (n = 3). For type II lesions, leaflet prolapse was corrected with central leaflet plication (n = 44), free margin resuspension (n = 6) [4] or a pericardial patch to close a large fenestration (n = 4) [5] or cusp extension (n = 1).

In this study, 105 patients were categorized into 2 groups according to preoperative mean GH measured on MDCT. Group S consisted of 35 patients with mean GH <16 mm. Group L consisted of 70 patients with mean GH ≥16 mm. All patients underwent transthoracic echocardiography; 88 underwent CT scan prior to discharge. Echocardiography was performed yearly thereafter. Follow-up was complete in 94% of patients at intervals of 2–76 months (mean 36 ± 23 months).

Statistical analysis

Data analysis was performed using SPSS Statistics 19 (IBM Corporation, Armonk, NY, USA). The normal distribution of GH was analysed using the Shapiro–Wilk test. Correlations were assessed using the Pearson’s correlation coefficient. Categorical variables are presented as percentage and compared using the χ² test or Fisher’s exact test. Continuous variables are reported as mean ± standard deviation or median (interquartile range). Differences were analysed using the Mann–Whitney U-test; P-value <0.05 was considered statistically significant.

RESULTS

Geometric height

The GH of the right coronary, the non-coronary and the left coronary cusp measured on preoperative CT was 16.1 ± 2.6, 18.3 ± 3.2 and 17.0 ± 2.6 mm, respectively (Table 1). The GH of the non-coronary cusp was greater than that of the left and the right coronary cusp (P < 0.0001 and P < 0.0001). The GH of the left coronary cusp was greater than that of the right coronary cusp (P < 0.0001). The preoperative mean GH was 17.1 ± 2.3 mm. Significant correlations between the mean GH and the clinical parameters were found for body height (r = 0.61) (Fig. 3), BSA (r = 0.55), annulus diameter (r = 0.49) and Valsalva sinus diameter (r = 0.58). The intraoperative measured mean GH (18.8 ± 2.2 mm) was larger than the preoperative mean GH measured on CT (P < 0.0001). The postoperative mean GH measured on CT was 17.1 ± 2.0 mm. There was no significant difference between the preoperative and the postoperative mean GH. Prolapse of the right coronary cusp was identified in 38 patients. The GH of prolapsed right coronary cusps was compared to that of non-prolapsed right coronary cusps. The prolapsed GH (15.4 ± 1.7 mm) was smaller than the non-prolapsed GH (16.6 ± 3.0 mm) (P = 0.03). However, intraoperative measurement revealed no significant difference between the prolapsed GH (19.1 ± 2.0 mm) and non-prolapsed GH (18.9 ± 2.6 mm) (P = 0.54).

Cusp size and clinical results

Preoperative characteristics were markedly different between group S and group L (Table 2). Patients in group S were older (66.3 ± 10.5 vs 58.1 ± 14.4 years, P = 0.004) and shorter (157.7 ± 8.4 vs 168.3 ± 8.7 cm, P < 0.0001). The root sizes were smaller in group S. The mean annulus diameter was 24.2 ± 2.6 mm in group S and 27.0 ± 2.9 mm in group L (P < 0.0001). Of the 35 patients in group S, 20 (57%) patients had severe and 14 (40%) patients had moderate AR; of the 70 patients in group L, 55 (79%) patients had severe and 6 (9%) patients had moderate AR preoperatively. Cusp repair for type II lesions was performed in 18 (51%) patients of group S and 33 (47%) patients of group L. More patients in group L underwent root replacement, and basal annuloplasty was preferentially performed in group S. The addition of annulus support combining root replacement or basal annuloplasty was more common in group L (66% vs 89%, P = 0.02).

Postoperatively, the basal annulus decreased significantly in both groups, but the significant inter-group difference persisted (P < 0.0001). Interestingly, the mean GH increased slightly in group S (P = 0.004) and decreased slightly in group L (P = 0.005). Echocardiography before discharge revealed no moderate AR in either group. Mild AR was observed in 51% of patients in group S and 17% of patients in group L (P = 0.006). During follow-up, moderate or severe AR developed in 14% of patients in group S and 10% of patients in group L (P = 0.74). Two patients in group S required reoperation for regurgitant valves (Table 2). Two patients in group L died during the follow-up. The first one developed acute aortic dissection and had undergone emergent replacement of the root and ascending aorta. She left the hospital with mild AR, but died 12 months later of sudden arrhythmia. The other patient underwent replacement of the dilated root but died 3 years later after a hypoglycaemic event. She had moderate AR during the follow-up without any signs of heart failure.

Annulus cusp mismatch

Preoperatively, 44 patients had severe annulus cusp mismatch, 22 had mild mismatch and 40 had no mismatch. Preoperative severe AR was observed in 65% of cases of severe mismatch, 64% of cases of mild mismatch and 50% of cases of no mismatch. Postoperative mild AR was more frequently observed in cases of severe mismatch (51%) than in cases of mild or no mismatch (51% vs 13%; P = 0.03) (Fig. 4).

Twenty (83%) patients in group S had severe annulus cusp mismatch and 3 (8%) patients had mild mismatch before surgery. Only 3 patients (8%) in group S had no mismatch. In group L, 15 (21%) patients had severe mismatch, 19 (27%) patients had mild mismatch and 36 (51%) patients had no mismatch. After surgery, the annulus cusp mismatch was resolved in most patients in group L, whereas more than half the patients in group S still had persistent mismatch (Fig. 5). Both patients who underwent reoperation had severe annulus cusp mismatch pre- and postoperatively.

DISCUSSION

Limited information is available regarding the normal human cusp dimensions. Thubrikar et al. [6] used echocardiography to determine that the mean GH in the normal heart was 14.4 mm. Khelil et al. [7] reported that the mean GH was 19.7 ± 0.3 mm in 20 hearts extracted from freshly frozen deceased donors. Of note, mean age was 79.9 ± 11.8 years. De Kerchove et al. [8] recently reported that the mean GH was 18.9 ± 1.5 mm in 25 fresh aortic homografts. Mean donors age was 52.3 ± 6 years. Schäfers et al. [1] reported that the mean GH measured intraoperatively was around 20.0 mm in patients who underwent aortic valve repair. There was a strong correlation between GH and body height, which was reconfirmed in our study. Considering that German individuals are usually much larger than Japanese individuals, our data were not different from those of previous reports.

Imaging resolution of the latest CT technology is far better than that of an echocardiogram. The aortic cusps are subject to physiological force in vivo. In annuloaortic ectasia, the outward force acting on the commissures stretches the leaflet free margins circumferentially. A smaller coaptation area aggravates the cusp tension. Intraoperative measurement cannot reproduce physiological beating heart conditions. Moreover, the variable radial force on the commissures might interfere with measurement precision. Our data demonstrated that the intraoperative measurements were 1.7 mm larger than the preoperative CT measurements. This difference cannot be ignored, as CL largely depends on GH value. Which measurements are more reliable? In our study, the postoperative CT measurements were equivalent to the preoperative measurements. Although alteration in haemodynamic force acting on leaflets after repair modifies cusp configuration, the coincidence of the 2 sets of data may imply that CT measurements are more reliable.

We found that GH measurements were unfeasible in deformed and prolapsed cusp. The GH measured intraoperatively was larger than that measured on CT. Interestingly, the GH of a prolapsed cusp was similar to that of a non-prolapsed cusp. Therefore, the GH of only non-prolapsed cusps may be used for calculating the mean GH. A new aortic cusp sizer to assess the aortic valve in the operating room was recently developed by Izzat et al. [9]. The aortic cusp sizer was used as a template to identify existing cusp and root alterations. These sizers are based on normal cusp dimensions. Using this method, a cusp can be categorized as either prolapsed or restricted if it stretched out or retracted, according to the selected sizer, respectively. Nevertheless, this modality requires intraoperative decision-making.

Mild AR was observed in 51% of patients with smaller cusps immediately after surgery; more than half of the patients still had a severe annulus cusp mismatch. Two patients who required reoperation also had had a severe annulus cusp mismatch postoperatively. During the early period of this study, we did not fully understand the concept of annulus cusp mismatch and the basal annulus was not systematically addressed. The clinical results might improve if we could establish an effective annulus management strategy. Using this model, we constructed a diagram to represent the relationship among mean GH, annulus diameter and predicted CL (Fig. 6). In clinical practice, this diagram seems to be useful for ascertaining a target annulus size according to mean GH. In a valve with a mean GH of 15 mm, the annulus should be reduced to <19 mm to achieve a target CL >4 mm. In a valve with mean GH of 14 mm, the annulus must be constricted to <18 mm. In reality, a mean GH of 14 mm seems to be the lower limitation for aortic valve repair.

The model for prediction of CL requires several assumptions. Each cusp should be of the same size and the cusps should meet at the centre of a circle. In the normal heart, GH, CL and free margin length are similar among cusps [8]; therefore, the model approaches reality. The straight portion of the cusp belly may be loosely curved towards the left ventricle. Therefore, the calculation may somewhat overestimate the CL. Bierbach et al. [10] suggested that the EH should be ≥9 mm for successful aortic valve repair. The average BSA of the study cohort was 1.87 m², whereas the average BSA in our cohort was 1.69 m². According to Bierbach, the average EH was 8 mm for a BSA of 1.6 m² in healthy individuals. The aim of the study was to clarify reparability of small cusp. Considering differences in physique, we preset the EH at 8 mm.

In mitral valve repair, supranormal coaptation depth is easily accomplished with a commercialized annuloplasty device. Considering that the CL of the normal heart is 3.3 ± 0.8 mm [8], the safety margin of the aortic valve to ensure sufficient coaptation may be 4–6 mm, i.e. narrower than that of the mitral valve. Consequently, aortic annulus management demands more sophisticated tools. Many ideas have been proposed [3, 11, 12], but an aortic annuloplasty system has yet to be established. Another limitation of annulus repair is the level gap between the basal annulus and aortoventricular junction, notably at the right coronary sinus [7]. This anatomical characteristic may preclude adequate annulus plication. Stable circumferential annuloplasty and especially root reimplantation are recommended whenever the annulus diameter is ≥28 mm [13]. We also preferentially selected root reimplantation with full external dissection for robust annular stabilization.

Three patients in group S had no annulus cusp mismatch before surgery. The cusp GH was 15.5 mm and the annulus diameters were 18.6, 19.9 and 20.5 mm. All patients were women with heights of 141, 146 and 157 cm. Postoperative AR was trivial in 2 and mild in the other. Therefore, a GH of 16 mm may not always be a contraindication for valve repair. Nevertheless, when annulus cusp mismatch exists, cautious evaluation is mandatory and valve replacement should be reconsidered. Annulus cusp mismatch is also present in patients with GH >18 mm. Particular attention to an appropriate approach to the basal annulus is required in these circumstances.

CONCLUSION

Our findings suggest that a small cusp size (GH <16 mm) is not necessarily a contraindication to aortic valve repair. However, most patients in this group have annulus cusp mismatch. Theoretically, a tight annuloplasty may regulate regurgitation, but the existence of severe mismatch may preclude such a small valve from becoming a good candidate for valve repair. Annulus cusp mismatch is also present in patients with larger cusps. Root replacement or secure annulus plication to correct annulus cusp mismatch is mandatory.

Conflict of interest: none declared.

Footnotes

REFERENCES