Aortic valve neocuspidalization in paediatric patients with isolated aortic valve disease: early experience

Abstract

OBJECTIVES

There is growing interest in the aortic valve (AV) neocuspidalization technique for the treatment of aortic valve disease (AVD). We report our medium-term results with this procedure performed in a paediatric patient population.

METHODS

Between July 2016 and May 2020, 22 patients with both congenital and acquired isolated AVD were treated with neocuspidalization. The primary outcome was progression of the preoperatively assessed AVD in the immediate postoperative course and at follow-up. Secondary outcome was freedom from reintervention by material used. Potential predictors of failure were analysed in relation to the primary outcome.

RESULTS

The median age at operation was 13.9 (interquartile range, 9.8–16.2) years, and the prevailing AV defect was stenosis in 10 cases (45%) and incompetence in 12 (55%). Pre-treated autologous pericardium was used in 13 patients whereas bovine pericardium in 9. Effective treatment of AV stenosis or regurgitation was achieved and remained stable over a median follow-up of 11.3 (4.7–21) months. Three patients required AV replacement at 4.9, 3.5 and 33 months. At follow-up, an upward trend of both median indexed vena contracta jet widths and aortic peak and mean gradients were recorded, the latter associated with a failure to grow the aortic annulus. Predictor of such outcome turned out to be the use of bovine pericardium. A significant inverse linear correlation between AV peak gradient at follow-up and preoperative aortic annular size (P = 0.008) was also demonstrated.

CONCLUSIONS

The Ozaki procedure is safe and effective in paediatric patients with AV disease. The use of heterologous pericardium should probably be minimized. Moreover, preoperative small aortic annuli should probably be promptly treated by means of an associated ring enlargement procedure.

INTRODUCTION

The optimal management of aortic valve (AV) disease in paediatric age is still challenging [1]. The Ross operation [2] is widely considered the gold standard therapy in children [3–6] despite its documented limitations related to autograft dilatation and durability of both autograft and pulmonary conduit [7, 8]. The aortic valve neocuspidalization (Ozaki operation) (AV Neo), first described by Ozaki et al. [9] in adult patients, has recently been considered as a viable therapeutic alternative in paediatric patients with a wide spectrum of AV disease including aortic stenosis, aortic regurgitation, infective endocarditis, prosthetic valve endocarditis and annulo-aortic ectasia. However, despite the good results in mid-term follow-up with a satisfactory freedom from reoperation reported in adults by Ozaki et al. [10], very limited data are available in the literature for paediatric patients, both in terms of number of cases and length of follow-up [11, 12].

The aim of this study was to evaluate the early- and medium-term results of the Ozaki procedure in paediatric patients with isolated AV disease.

METHODS

Study population

From July 2016 to May 2020, 22 patients aged <18 years with both congenital and acquired isolated aortic valve disease (AVD) underwent surgical treatment at our institution. Patient recruitment and data analysis were retrospectively performed. The Scientific Board of Bambino Gesù Children's Hospital and Research Institute approved the study and waived the need for informed patient consent.

Surgical procedures

Briefly, surgical procedures were performed on moderate hypothermic cardiopulmonary bypass with aortic-bicaval cannulation with myocardial protection achieved by intermittent infusion of cold blood cardioplegia. Valve leaflets were made with either autologous pericardium when available or glutaraldehyde- (CardioCel, Milton, Australia) or photo-fixed (Photofix, Crylife, Kennesaw, GA, USA) decellularized bovine pericardium. First, diseased cusps were excised and inter-commissural edges were measured using the original sizing apparatus (JOMDD, Tokyo, Japan). Then, 3 new cusps were designed and trimmed on a template from autologous pericardium treated with buffered 0.6% glutaraldehyde solution for 5 min, when available. The 3 neo-cusps were similar in size to those excised from three-leaflet valves with the reproduction of native commissures, while they were symmetrical in case of unicusp/bicuspid valves, with the recreation of new commissures. In the very preliminary experience, leaflet sizing was performed according to the modification proposed by Rankin et al. [13]. Suturing technique was always applied as previously published [9]. Finally, in young children with small aortic annulus bilateral enlargement of the AV ring was performed according to a modified Yamaguchi approach [14].

Echocardiography

A single reviewer independently measured preoperative, postoperative and follow-up echocardiogram images with a second reviewer analysing a random sample. The aortic annulus, the aortic root and the peak and mean gradients across the AV and the vena contracta jet width were measured and indexed to body surface area. In particular, aortic incompetence, determined by vena contracta, was measured in parasternal long axis views and categorized according to its indexed value. Aortic annular size was also measured in parasternal long axis view. Finally, leaflet mobility of neo-AV was observed in parasternal long- and short-axis views, using also 3D imaging where applicable. Categorization of aortic stenosis and incompetence is in accordance with Baird et al. [12].

Antiplatelet/anticoagulant treatment

Seventeen (77%) patients were treated with Aspirin postoperatively. Beginning July 2018, a strategy based on an initial anticoagulant therapy with warfarin (target international normalized ratio: 1.5–2.5) for 3 months, followed by long-term antiplatelet treatment was used. In 1 patient, the target international normalized ratio was raised to 3.5 for temporary blockage of a valve cusp, which resolved spontaneously without sequelae. Finally, the first patient of the series did not receive any anticoagulant or antiplatelet therapy.

Outcome measures

The primary outcome was the progression of the AVD (either stenosis and/or incompetence) throughout hospital course and at follow-up. The secondary outcome was freedom from reintervention.

Statistical analysis

Continuous and categorical variables were reported as median (interquartile range) and number (percentage), respectively. Between-group differences were compared using two-sample t-tests and Wilcoxon rank sum tests. Within-group differences were compared using one-sample t-tests and Wilcoxon signed rank tests to explore whether changes in echocardiographic scores (e.g. preoperative versus postoperative, postoperative versus last follow-up) differed from zero. Relationships between continuous variables were assessed using Spearman correlation test. Rates of freedom from reintervention were estimated using Kaplan–Meier methodology. The level for statistical significance was set at P < 0.05. All analyses were performed using Stata data analysis and statistical software version 11.1 (StataCorp, College Station, TX, USA).

RESULTS

Demographics

Preoperative characteristics are summarized in Table 1. The median age and weight at operation were 13.9 (9.8–16.2) years and 55 (34–73) kg, respectively, and the prevailing AV defect was stenosis in 10 cases (45%) and incompetence in 12 (55%). Functional classification of prevalent aortic valve incompetence (AVI) [15] was type II in 5 cases (41.7%) and type III in 7 (58.3%). Five (23%) patients had tricuspid, 11 (50%) had bicuspid and 5 (23%) had unicusp AVs, whereas 1 (4%) had structural degeneration of a bioprosthesis. Congenital lesions accounted for the majority (82%) of the pathology and were associated with genetic syndromes in 9% of the cases (Turner, 1 case, Alagille, 1 case). Five (23%) patients had 5 previous surgeries and 1 previous percutaneous balloon aortic valvuloplasty, whereas 3 (14%) had 5 previous isolated percutaneous balloon aortic valvuloplasty.

Table 1:

Preoperative characteristics

Variables	Values
Weight at operation (kg), median (IQR)	55 (34–73)
Age at operation (years), median (IQR)	13.9 (9.8–16.2)
Aortic valve peak gradient (mmHg), median (IQR)	69 (30–85)
Aortic valve mean gradient (mmHg), median (IQR)	40 (20–50)
Vena contracta jet width (mm), median (IQR)	5.5 (0–8)
Indexed vena contracta jet width (mm/m2), median (IQR)	3.6 (0–6.7)
Aortic root (mm), median (IQR)	26.5 (23–29)
Aortic root indexed (mm/m2), median (IQR)	19 (15.3–23.9)
Annulus (mm), median (IQR)	20.5 (18–22)
CPB time (min), median (IQR)	142 (117–172)
Cross-clamp time (min), median (IQR)	103 (91–113)
Redo intervention, n (%)	5 (22.7)
Preoperative valvular disease
Prevalent AVS, n (%)	10 (45)
Prevalent AVI, n (%)	12 (55)
Functional classification of prevalent AVI [15]
Type I, n (%)	0 (0)
Type II, n (%)	5 (41.7)
Type III, n (%)	7 (58.3)
Diagnosis
Congenital aortic valve disease, n (%)	15 (68.2)
Previous repair of doubly committed VSD, n (%)	2 (9.1)
Previous AVR, n (%)	1 (4.5)
Rheumatic/endocarditis disease, n (%)	4 (18.2)
Turner, n (%)	1 (4.5)
Alagille, n (%)	1 (4.5)
Aortic valve anatomy
Structural degeneration of bioprosthetic valve, n (%)	1 (4.6)
Bicuspid aortic valve, n (%)	11 (50)
Unicusp aortic valve, n (%)	5 (22.7)
Tricuspid aortic valve, n (%)	5 (22.7)
Previous operations
Surgery ± PBAV, n (%)	5a (22.8)
Isolated PBAV, n (%)	3b (13.6)

AVI: aortic valve incompetence; AVR: aortic valve replacement; AVS: aortic valve stenosis; CPB: cardiopulmonary bypass; IQR: interquartile range; PBAV: percutaneous balloon aortic valvuloplasty; VSD: ventricular septal defect.

^aFive patients had 5 surgeries and 1 PBAV.

^bThree patients had 5 PBAV.

Autologous and bovine pericardium were used in 13 (59%) and 9 (41%) patients, respectively (Cardiocel^® in 8, Photofix^® in 1) to reconstruct 3 equal leaflets in 10 cases and unequal leaflets [2 equal and 1 either larger (n = 7) or smaller (n = 5)] in 12 cases (Table 2). Leaflet sizing was performed according to Rankin in the first 4 patients of the series, whereas original Ozaki sizers and templates were used in the remaining 18. Finally, 1 patient had bilateral enlargement of the AV ring according to a modified Yamaguchi approach.

Table 2:

Intraoperative characteristics

Leaflet material
Autologous pericardium, n (%)	13 (59)
Glutaraldeyde-fixed bovine pericardium, n (%)	8 (36.4)
Photo-fixed bovine pericardium, n (%)	1 (4.6)
Leaflet sizing
3 equal leaflets, n (%)	10 (45.5)
Unequal leaflets 2 equal +
1 larger, n (%)	7 (31.8)
1 smaller, n (%)	5 (22.7)

Early results

The median hospital stay was 8 (7–11) days. One patient required drainage of a pericardial serous effusion on the 4th postoperative day, whereas 1 patient required intense anticoagulation for temporary blockage of a cusp that occurred on the 7th postoperative day, with subsequent complete recovery of the cusp mobility and valve function.

Follow-up

All patients were monitored and followed up prior to discharge with an echocardiogram, repeated periodically every 6 months following surgery. The first 10 of the series underwent a 4D cardiac magnetic resonance (CMR) at a median interval of 4 months postoperatively. The median follow-up for the whole cohort defined by the last available echocardiogram was 11.3 (4.7–21) months and was 100% complete.

Assessment of AVD by echocardiogram and 4D CMR

Table 3 shows progression of aortic stenosis and aortic regurgitation detected by echocardiography.

Table 3:

Progression of aortic stenosis, aortic regurgitation, aortic ring and aortic root diameter preoperatively, immediately postoperative and at follow-up

Variables	Preoperative, median (IQR)	Postoperative, median (IQR)	Follow-up, median (IQR)	P-value
Aortic valve peak gradient (mmHg)	69 (30–85)	12.7 (8.4–16.3)	19 (15.3–30)	<0.001 *
Aortic valve mean gradient (mmHg)	40 (20–50)	6.9 (4.4–10.3)	11 (8.5–15)	<0.001 *
Vena contracta jet width (mm)	5.5 (0–8)	0 (0–1.5)	1.8 (0–2.8)	<0.001 *
Indexed vena contracta jet width (mm/m2)	3.6 (0–6.7)	0 (0–1.1)	1.1 (0–2.3)	<0.001 *
Aortic root (mm)	26.5 (23–29)	–	27 (23–31)	0.03
Aortic root indexed (mm/m2)	19 (15.3–23.9)	–	17.3 (15–23)	0.94
Aortic ring (mm)	20.5 (18–22)	–	19 (17–22)	0.15
Aortic ring indexed (mm/m2)	14.2 (12.2–18.3)	–	12.2 (11.4–16.2)	0.08

^*P for comparison between preoperative and immediately postoperative period. The bold values indicates level for statistical significance was set at P < 0.05.

Aortic stenosis

Compared with the preoperative condition, a significant reduction in the aortic transvalvular gradient was recorded postoperatively (Fig. 1). In particular, a significant decrease in aortic peak and mean gradients from 69 (30–85) to 12.7 (8.4–16.3) mmHg and from 40 (20–50) to 6.9 (4.4–10.3) mmHg, respectively, was achieved (P < 0.001 for both comparisons). Nonetheless, at follow-up aortic peak and mean gradients showed an upward trend [19 (15.3–30) and 11 (8.5–15) mmHg, respectively].

Figure 1:

Progression of aortic stenosis.

Aortic regurgitation

The indexed vena contracta jet width in the postoperative period significantly decreased from 3.6 (0–6.7) to 0 (0–1.5) mm/m² (P < 0.001) and then increased at follow-up [1.1 (0–2.3) mm/m²] (Fig. 2). At last echocardiogram, 14 (64%) patients had none-trivial AVI, 7 (32%) patients had mild AVI and 1 (4%) had moderate AVI.

Figure 2:

Progression of aortic regurgitation. BSA: body surface area.

Physiological laminar flow through the ascending aorta was found in all 10 patients who underwent 4D CMR.

Assessment of aortic ring and root dimension by echocardiogram

Table 3 shows the progression of both aortic ring and aortic root diameter. Compared to the postoperative period, a non-significant decrease at follow-up of both absolute and indexed measurements of the aortic annulus from 20.5 (18–22) to 19 (17–22) mm (P = 0.15) and from 14.2 (12.2–18.3) to 12.2 (11.4–16.2) mm/m² (P = 0.08), respectively, was found. On the other hand, a significant increase in aortic root absolute diameter was detected at follow-up [from 26.5 (23–29) to 27 (23–31) mm (P = 0.03)]. Indexed measurements did not confirm such trend [from 19 (15.3–23.9) to 17.3 (15–23) mm/m² (P = 0.94)].

Predictors of AV performance

Table 4 shows factors significantly associated with peak AV gradient at follow-up. Smaller preoperative aortic annulus dimension and the use of bovine pericardium (compared to autologous pericardium) were significantly associated with higher peak AV gradient. A trend towards higher AV peak gradient was also found in patients with prevalent preoperative AV stenosis defect. Preoperative aortic root dimension and the Rankin modification for leaflet sizing were also tested but did not reach statistical significance. When the mean AV gradient at-follow-up was analysed, the same predictors of AV performance were identified (data not shown). Finally, a significant inverse linear correlation between AV peak gradient at follow-up and preoperative aortic annular size (P = 0.008) was demonstrated (Fig. 3).

Table 4:

Predictors of aortic valve performance

Variables	Peak AV gradient at follow-up (mmHg)	P-value
Preoperative valvular disease		0.07
Prevalent AVS	24.8 (16.8–31.5)
Prevalent AVI	16.4 (8.4–22.1)
Aortic annulus dimension (mm)	–	0.008
Material used		0.01
Bovine pericardium	30 (17.5–33)
Autologous pericardium	15 (8.4–22.1)

AV: aortic valve; AVI: aortic valve incompetence; AVS: aortic valve stenosis. The bold values indicates level for statistical significance was set at P < 0.05.

Figure 3:

Relationship between aortic valve peak gradient at follow-up and preoperative aortic annulus dimension. AV: aortic valve.

Survival/reoperations

Two (9%) of the initial patients who underwent the Ozaki procedure required early conversion to valve replacement. The first was an 8-year-old boy with stenotic bicuspid AV, who developed intermittent myocardial ischaemia due to redundancy of the left coronary cusp (Cardiocel^® heterologous pericardium) obstructing at times the left coronary ostium. The patient was successfully converted to a Ross procedure 4.9 months after Ozaki operation and is in New York Heart Association (NYHA) class I 31 months postoperatively. The second patient was a 17-year-old young boy who had already undergone combined mitral and AV repair for rheumatic disease who subsequently underwent the Ozaki procedure using Cardiocel^® heterologous pericardium as a rescue treatment after a failed attempt of mechanical AV replacement. This patient experienced an aseptic aortic root disruption with lung congestion 3.5 months after surgery. He was then successfully converted to homograft aortic root replacement and is in NYHA class I 30 months postoperatively.

A third patient, a 9.5-year-old child with repaired truncus arteriosus, previous AV repair, previous AV replacement with bioprosthesis and previous Ozaki operation (first patient of our series), developed severe stenosis 33 months postoperatively and required mechanical AV replacement. Of note is the fact that the child was not receiving any anticoagulant or antiplatelet therapy. At operation all 3 cusps originally made with heterologous pericardium (Cardiocel^®) appeared uniformly severely calcified. Interestingly, the aortic ring diameter at reoperation was 2 mm larger than previously measured at the time of the Ozaki procedure. This child is in NYHA class I and leads a normal life 14 months after the procedure.

Freedom from reintervention was greater when autologous pericardium was used (Fig. 4). No endocarditis episodes were detected in our series.

Figure 4:

Freedom from reintervention by material used (no deaths were observed during the study period).

DISCUSSION

The ideal surgical treatment of AV disease in children is still debated, as different approaches have specific advantages and drawbacks. Although AV repair appears as the ideal approach in paediatric patients, outcomes are limited and difficult to ascertain, with some series suggesting 46% reoperation at 7.5 years in older children [16]. The use of prosthetic material is universally related with a shorter longevity of the repair [17, 18] and with an increased risk of mortality, mainly related to coronary ostia occlusion [19]. Suboptimal results are also reported for the leaflet extension technique used for type III AV insufficiency in children, as a result of limited leaflet mobility and persistent gradients [17]. Therefore, Ross operation has been regarded as the gold standard for paediatric AV disease despite its technical difficulty and the non-negligible peri-operative risks [1–6, 20]. However, the Ross procedure is not free from late complications related to autograft dilatation and durability of both autograft and pulmonary conduit [7, 8, 21].

Recently, AV Neo according to Ozaki showed good short- and mid-term results with no early mortality and adequate postoperative haemodynamic parameters in paediatric patients [11, 12]. In fact, AV Neo preserves the anatomy of the aortic root and its physiological cardiodynamics, is applicable with any valve anatomy even after AV replacement, is technically reproducible and theoretically does not preclude native annular growth.

As for the cusp extension technique, the AV Neo might theoretically induce coronary ischaemia as a result of the presence of high patches. However, the height of the commissures in AV Neo never exceeds the sinu-tubular junction. Coronary obstruction is therefore unlikely provided that excessive redundancy of 1 or more leaflets is avoided. Moreover, the large coaptation area does not create turbulence in the ascending aorta, as demonstrated by postoperative 4D CMR studies. Conversely, large coaptation surfaces of the reconstructed pericardial leaflets may theoretically ensure valve competence during annular growth. In addition, AV Neo enables annular enlargement and leaves the pulmonary valve untouched. It is therefore suitable even in patients with small aortic rings and allows subsequent completion of a Ross operation whenever required.

Despite potential advantages, medium- and long-term results of the AV Neo technique are still lacking, especially with regard to the impact of the patch material used for leaflet construction as well as postoperative annular growth. The aim of this study was to evaluate the efficacy of the Ozaki procedure in paediatric patients and the identification of potential predictors of AV performance.

Our results suggest that the AV Neo represents a safe technique that allows satisfactory resolution of AV disease in children in the early postoperative period as well as at follow-up. Nonetheless, a trend towards an increase in transvalvular gradient at follow-up was also evident. The lack of growth of the aortic annulus might probably explain the increase in AV gradients at follow-up. This is in contrast with Baird et al. [12], who showed significant growth of the aortic ring in all patients at follow-up, primarily in patients who received aortic annular enlargement. A lower percentage of annular enlargement performed in our series (4.5% vs 14% compared to Baird et al. [12]) might probably account for the lower annular growth rate in our study. We also found an inverse correlation between the preoperative dimension of the aortic annulus and at follow-up AV gradients. This might indicate that a more aggressive treatment of annular hypoplasia in patients with prevalent or exclusive stenosis might be necessary at the time of the Ozaki. We could not find a cut-off point in our series. Nonetheless, the proposed completion of annular enlargement in patients with annular Z-scores <2 and in patients whose native annulus measures <15 mm seems reasonable [12].

Higher AV gradients were documented when bovine instead of autologous pericardium was utilized, thus suggesting that the use of heterologous pericardium should be avoided whenever possible.

We also observed a modest increase in AV regurgitation at follow-up. The increment of the median vena contracta jet width values did not translate into clinically relevant postoperative AV defects, with <5% of patients being classified as having moderate AV regurgitation. Longer follow-up periods are needed to evaluate whether clinically significant AV regurgitation will eventually occur.

Limitations

This study has limitations. First of all, this is a single-centre retrospective study with a limited sample size. The limited follow-up time precludes any definitive conclusion. The measurement of the aortic annular size was performed in parasternal long axis view only, thus possibly allowing underestimation of such measurements. The analysis of aortic root dilatation at follow-up might have been overestimated, due to bilateral enlargement of the aortic annulus performed in 1 patient. Eight out of 9 patients received the same type of bovine pericardium, thus precluding the analysis of the impact of different types of heterologous pericardium on outcomes [22].

CONCLUSION

The Ozaki procedure is safe and effective in paediatric patients with AV disease. The use of heterologous pericardium should probably be minimized. Moreover, preoperative small aortic annuli should probably be promptly treated by means of an associated ring enlargement procedure.

ABBREVIATIONS

AV

Aortic valve

AV Neo

Aortic valve neocuspidalization (Ozaki operation)

AVD

Aortic valve disease

AVI

Aortic valve incompetence

CMR

Cardiac magnetic resonance

NYHA

New York Heart Association

 

Conflict of interest: none declared.

Author contributions

Angelo Polito: Data curation; Investigation. Sonia B. Albanese: Supervision; Validation. Enrico Cetrano: Investigation. Marianna Cicenia: Data curation; Investigation. Gabriele Rinelli: Supervision; Validation. Adriano Carotti: Writing—original draft; Writing—review & editing.

Reviewer information

Interactive CardioVascular and Thoracic Surgery thanks Jose I. Aramendi, Masamichi Matsumori and Ludwig C. Müller for their contribution to the peer review process of this article.