Surgical outcomes of aortic valve replacement in children with Ross and Ozaki procedure

Abstract

OBJECTIVES

To evaluate the midterm outcomes of the Ross and Ozaki procedures as alternatives to mechanical valve replacement in children with aortic valve disease.

METHODS

All patients undergoing the Ross or Ozaki procedure between January 2017 and December 2023 were included. Primary outcomes were moderate or greater aortic stenosis (AS) or aortic regurgitation (AR) after surgery. Secondary outcomes included reoperations.

RESULTS

The cohort comprised 54 patients, with 35 patients (65%) undergoing the Ross procedure and 19 undergoing (35%) the Ozaki procedure. The mean age at surgery was 8.5 years (standard deviation: 4.0). Patients undergoing the Ozaki procedure had more cases of predominant AR and large aortic valve annuli compared to the Ross group. The median follow-up time was 3.8 years (interquartile range, 2.0–5.3 years). Freedom from moderate or greater AS/AR was 92% at 2 years and 88% at 4 years in the Ross group, versus 59% at 2 years and 30% at 4 years in the Ozaki group (P < 0.001). Freedom from neoaortic valve reoperation was 100% at 2 years and 94% at 4 years in the Ross group, compared to 92% at 2 years and 71% at 4 years in the Ozaki group (P = 0.002). Overall freedom from any reoperation was similar between groups (P = 0.16).

CONCLUSIONS

The Ross procedure appears to provide a more durable neoaortic valve, particularly for patients with predominant AS or mixed lesions. The Ozaki procedure, primarily performed in patients with AR in our cohort, showed suboptimal midterm outcomes. Further studies with larger cohorts are needed to validate these findings.

Open in new tabDownload slide

INTRODUCTION

The optimal surgical approach for aortic valve disease in children remains a subject of debate. While valve repair is generally preferred when feasible [1], certain cases involve valves deemed unrepairable or unlikely to provide long-term durability, necessitating valve replacement [2]. Mechanical valve replacement, which requires lifelong anticoagulation therapy, is typically considered a ‘last resort’ due to its associated higher mortality and morbidity [2, 3]. The Ross procedure, which replaces the aortic valve with a pulmonary autograft, has experienced a renaissance in recent decades but remains technically demanding [4, 5]. The Ozaki procedure (AVNeo, or aortic valve neocuspidization) has shown promising results in adult populations [6, 7]. However, despite growing enthusiasm in the congenital field, its long-term outcomes in paediatric patients are still underexplored [8]. Since 2017, our centre has utilized both the Ross and Ozaki procedures as alternatives to mechanical prostheses for aortic valve replacement. This study aims to evaluate the midterm surgical outcomes of these approaches.

PATIENTS AND METHODS

Ethics statement

This study was approved by the Institutional Review Board of Shanghai Children’s Medical Center (SCMCIRB-K2024164-1, 7 August 2024). The need for individual consent was waived due to the retrospective nature of the study.

Patients

Between January 2017 and December 2023, 54 patients underwent Ross procedure (n = 35, 65%) or Ozaki procedure (n = 19, 35%) for aortic valve replacement and were included in the study. The annual case volume for both procedures are shown in Supplementary Material, Fig. S1. During the study period, the Ross procedure was performed by 7 surgeons, with Dr Haibo Zhang accounting for 25 cases (71%). The Ozaki procedure was performed by 4 surgeons, with Dr Zhang performing 10 cases (53%).

Surgical techniques

The decision to perform either the Ross or Ozaki procedure was made during the preoperative multidisciplinary team meeting and finalized based on intraoperative assessment (Fig. 1). If durable aortic valve repair was deemed unachievable, valve replacement with either the Ross or Ozaki procedure was considered. The Ross procedure was generally favoured for patients with predominant aortic stenosis (AS) or mixed lesions, provided there were no contraindications such as connective tissue disorders, autoimmune disorders, coronary artery anomalies or significant size discrepancies between the aortic and pulmonary valve annuli, and the pulmonary valve demonstrated good functionality. The Ozaki procedure was preferred for patients presenting with predominant aortic regurgitation (AR) and a large aortic valve annulus.

The Ross procedure was performed using the full-root technique. Briefly, after inspecting the aortic valve, the main pulmonary artery was transected at the bifurcation, and the pulmonary autograft was harvested. The autograft was then implanted in the intra-annular position with coronary transfer. The loose jacket technique was then employed to reinforce the autograft with the native noncoronary sinus, the left-right commissure pillar and the native aortic tissue above the coronary button. The right ventricle-to-pulmonary continuity was established either with a bovine jugular vein conduit (BalMedic, Beijing, China; all before May 2019; n = 7, 20%) or a handmade tri-leaflet expanded polytetrafluoroethylene (ePTFE; Preclude, W. L. Gore & Associates, Inc, AZ, USA; n = 28, 80%) valved conduit.

The Ozaki procedure was described previously [9]. Briefly, the autologous pericardium (n = 13, 68%) was harvested and was treated with 0.6% glutaraldehyde solution for 10 min and then rinsed 3 times in saline for 6 min. In redo cases where there were no sufficient native pericardium available, bovine pericardium (BalMedic, Beijing, China; n = 6, 32%) was used instead. Intercommissural distance were measured using the Ozaki sizers (JOMDD, Tokyo, Japan), and the cusps were cut using the template. Modified Ozaki procedure with single leaflet reconstruction was performed in 8 patients (42%).

Anticoagulation

For patients undergoing the Ross procedure with handmade ePTFE-valved conduits, we typically prescribed both warfarin and aspirin for the first 6 months, followed by aspirin alone for a total duration of 2 years. Patients who underwent the Ozaki procedure were treated with aspirin for 6 months.

Study outcomes

The primary outcomes of our study were the occurrence of moderate or greater AS or AR after surgery, defined by a peak gradient ≥36 mmHg or vena contracta width >4 mm on echocardiography [9, 10]. Secondary outcomes included reoperations after initial surgery, specifically those on the neoaortic valve, as well as any related reoperations. Postoperative complications were defined as any adverse events occurring within 30 days of surgery, including but not limited to major bleeding, arrhythmias necessitating medical intervention, infections (e.g. mediastinitis, endocarditis) and neurological events.

Statistical analysis

Data were analysed using SPSS software version 22.0 (IBM-SPSS Inc, Armonk, NY, USA). Continuous variables were presented as mean and standard deviation (SD) for normally distributed data or as median [interquartile range (IQR)] for skewed distributions. The normality of continuous data was evaluated using the Shapiro–Wilk test. Parametric tests (Student’s t-test) were employed for comparisons of normally distributed continuous variables between groups, while non-parametric tests (Wilcoxon rank-sum test) were used for skewed data, accounting for the presence of outliers or distribution asymmetry. Categorical variables were summarized as frequencies and percentages and compared using chi-squared test, or Fisher’s exact test when the sample size was <40 or when >20% of expected cell counts were <5. For survival analysis, freedom from moderate or greater AS/AR, as well as freedom from reoperation, were estimated using Kaplan–Meier curves with 95% confidence intervals (CIs). Competing risk analysis was considered; however, due to the absence of failure events in some groups and the temporal sequence of failure and competing events, Kaplan–Meier methods were deemed suitable for risk evaluation in this study. The log-rank test was applied to assess differences between groups. To ensure the appropriateness of the log-rank test, the proportional hazards assumption was evaluated by inspecting the log-minus-log survival plots. The reverse Kaplan–Meier method was used to estimate the median follow-up time. There were minimal missing data in our study; therefore, a complete case analysis was conducted.

RESULTS

Patient characteristics

A total of 54 patients were included in the study, with 35 (65%) undergoing the Ross procedure and 19 (35%) receiving the Ozaki procedure. The mean age at repair was 8.5 years (SD: 4.0). The majority of cases were due to congenital lesions (n = 47, 87%), followed by infective endocarditis (n = 5, 9%), with 2 additional patients presenting with significant AR following a previous arterial switch operation. Among the patients, 13 (24%) had AS, 22 (41%) had AR and 19 (35%) had mixed lesions. Valve morphology included bicuspid (n = 28, 52%), tricuspid (n = 25, 46%) and 1 case of quadricuspid aortic valve (2%). A total of 12 patients (22%) had a history of prior aortic valve intervention, with 5 having undergone balloon aortic valvuloplasty and 7 surgical aortic valvuloplasty. The median aortic valve annulus diameter was 1.8 cm (IQR, 1.4–2.2), with a mean Z-score of +2.3 (SD: 3.0).

There was no significant difference in age at repair between the Ross and Ozaki groups. Patients undergoing the Ozaki procedure predominately presented with AR or mixed lesions (n = 18, 95%) and had a significantly larger aortic valve annulus compared to the Ross group [2.2 cm (IQR, 1.9–2.4) vs 1.5 cm (IQR, 1.4–1.9), P < 0.001] (Table 1).

Intraoperative results

Concomitant procedures included Konno procedures in 3 patients and mitral valvuloplasty in 3 patients in the Ross group, while in the Ozaki group, 1 patient underwent concomitant mitral valvuloplasty and 1 patient had ventricular septal defect closure. Four patients (21%) undergoing the Ozaki procedure required a second bypass run to address regurgitant neoaortic valve, whereas none was required in the Ross group (P = 0.012). There were no significant differences in cardiopulmonary bypass time [158 min (IQR, 138–209) vs 155 min (IQR, 137–167), P = 0.16] or aortic cross-clamp time [109 min (IQR, 99–129) vs 116 min (IQR, 102–127), P = 0.70] between the 2 groups. Post-bypass transoesophageal echocardiography revealed that mild or less AS/AR was achieved in 34 patients (97%) in the Ross group, compared to 14 patients (74%) in the Ozaki group (P = 0.017). The remaining 5 patients in the Ozaki group left the operating room with mild-to-moderate AS/AR (Table 2).

Early outcomes

One patient (3%) who underwent the Ross procedure died on the operative day. This was a 12-year-old male with AS who had a prior history of surgical aortic valvuloplasty. Despite having patent proximal coronary artery flow, the patient could not be weaned from cardiopulmonary bypass due to severe left ventricular systolic dysfunction, and the family declined the recommendation of extracorporeal membrane oxygenation support. There were no early mortalities in the Ozaki group.

Patients in the Ross group appeared to experience more postoperative complications compared to the Ozaki group (14% vs 0, P = 0.15). In the Ross group, 1 patient required cardiopulmonary resuscitation for sudden cardiac arrest on postoperative day 1, another underwent resternotomy for bleeding, a 3rd for pericardial effusion and 1 experienced gastrointestinal bleeding related to anticoagulation therapy, in addition to 1 in-hospital mortality as aforementioned. In contrast, no postoperative complications were observed in the Ozaki group. The duration of mechanical ventilation [45 h (IQR, 23–72) vs 24 h (IQR, 21–44), P = 0.020] and intensive care unit stay [4 days (IQR, 3–5) vs 4 days (IQR, 3–4), P = 0.049] were significantly longer in the Ross group compared to the Ozaki group; however, the overall postoperative hospital length of stay was similar between the 2 groups [15 days (IQR, 11–20) vs 10 days (IQR, 8–16), P = 0.51] (Table 2).

Aortic stenosis/aortic regurgitation at follow-up

Three patients in the Ross group were lost to follow-up after discharge. The median follow-up time was 3.8 years (IQR, 2.0–5.3 years). In the Ross group, freedom from moderate or greater AS was 100% at both 2 and 4 years after surgery, while in the Ozaki group, it was 88% at 2 years (95% CI 59–97%) and 60% at 4 years (95% CI 21–85%) (P < 0.001) (Fig. 2A). Freedom from moderate or greater AR was 92% at 2 years (95% CI 73–98%) and 88% at 4 years (95% CI 67–96%) after surgery in the Ross group, compared to 66% at 2 years (95% CI 39–83%) and 38% at 4 years (95% CI 13–63%) in the Ozaki group (P < 0.001) (Fig. 2B). Overall, freedom from moderate or greater AS/AR was 92% at 2 years (95% CI 73–98%) and 88% at 4 years (95% CI 67–96%) in the Ross group, and was 59% at 2 years (95% CI 33–78%) and 30% at 4 years (95% CI 8–56%) in the Ozaki group (P < 0.001) (Fig. 2C).

In the Ross group, patients with predominant AR had significantly lower freedom from moderate or greater AS/AR after surgery compared to those with predominant AS or mixed lesions (P = 0.003) (Fig. 3A). In the Ozaki group, neither the type of leaflet material (autologous versus bovine pericardium, P = 0.52) (Fig. 3B) nor the number of neocusps reconstructed (three-leaflets versus single-leaflet reconstruction, P = 0.78) (Fig. 3C) affected the progression to moderate or greater AS/AR after surgery. Additionally, no significant differences in freedom from moderate or greater AS/AR were observed between cases performed by the most experienced surgeon and those by other surgeons (P = 0.14) or between the first 10 cases and the last 9 cases (P = 0.84).

Reoperations

Freedom from reoperation on the neoaortic valve was 100% at 2 years and 94% at 4 years (95% CI 67–99%) after surgery in the Ross group, compared to 92% at 2 years (95% CI 57–99%) and 71% at 4 years (95% CI 33–90%) the Ozaki group (P = 0.002) (Fig. 4A). However, due to the need for right ventricular outflow tract conduit replacement in the Ross group, overall freedom from any reoperation was similar between the Ross and Ozaki groups (P = 0.16) (Fig. 4B).

In the Ross group, 1 patient required aortic valvuloplasty with right ventricular outflow tract conduit replacement due to moderate AR and conduit stenosis 2.8 years after the initial surgery. Additionally, 4 patients required conduit replacement alone.

In the Ozaki group, 6 patients required reoperations on the neoaortic valve due to moderate-severe or severe AR, with 1 patient also accompanied with moderate-severe AS (Video 1). Of the 3 patients who had neoleaflets reconstructed with autologous pericardium, the leaflets were found to be calcified or thickened. Two of the 3 patients who had neoleaflets reconstructed with bovine pericardium required reoperation due to endocarditis. The reoperations were performed via mechanical valve replacement in 5 patients and the Ross procedure in 1 patient (this patient was not included in the Ross group). Details of the reoperations in the Ozaki group are listed in Table 3.

DISCUSSION

The ideal goal of aortic valve surgery in children is to provide patients with a well-functioning valve that offers long-term durability and growth potential while minimizing the need for future reinterventions [2, 11]. The resurgence of the Ross procedure and the introduction of the Ozaki procedure have broadened our surgical options available for paediatric patients, particularly in cases where optimal repair is challenging and suit for different scenarios [11]. In our study, the Ross procedure appeared to provide favourable midterm durability, especially in patients with predominant AS or mixed lesions. In contrast, the Ozaki procedure, which was primarily performed in patients with AR and large annuli in our cohort, was associated with a higher rate of valve dysfunction and reintervention during midterm follow-up. However, these findings should be interpreted cautiously due to the limited sample size, retrospective nature of the study and potential confounding factors.

The Ozaki procedure, 1st introduced by Dr Shigeyuki Ozaki in 2007 for adult populations [12], has shown excellent midterm results in adults, with outcomes comparable to bioprostheses in terms of freedom of aortic valve re-replacement, and significantly lower peak gradients [6]. In paediatric populations, the Ozaki procedure has also demonstrated promising short-term outcomes in several pioneering centres. Studies from Boston group reported that 96% and 91% of patients had less than moderate AR and AS, respectively, at the last follow-up, with a median follow-up of 8.1 months [9]. Additionally, freedom from moderate or greater AR and AS was 76% and 86% at 2 years in patients who underwent single leaflet reconstruction [13]. Another study from Great Ormand Street Hospital in London confirmed its acceptable short-term haemodynamic results, with freedom from reoperation or moderate or greater AR at 1, 2 and 3 years being 94, 85 and 79%, respectively, with a median follow-up of 14.1 months [14].

Despite these promising early outcomes, recent data on the midterm outcomes of the Ozaki procedure in children have raised concerns [8], which were hinted at in the aforementioned publications. In the Boston report, 12 of the 57 patients (21%) showed decreased leaflet mobility on follow-up echocardiography, and it was suggested that aggressive anticoagulation therapy with aspirin and warfarin might help reduce this risk. They also advocated for shorter tanning time for pericardial fixation, although the effectiveness of this modification has not been confirmed [9]. In the Great Ormand Street report, 6 reoperations (10%) were performed, with 3 cases due to endocarditis [14]. In our study, freedom from moderate or greater AS/AR was only 59% at 2 years and 30% at 4 years, with endocarditis or calcification observed during reoperations. Notably, the mean age at surgery in our Ozaki group (9.9 years) was younger than that reported in some other paediatric cohorts [9, 14], which may have contributed to the reduced durability of outcomes. However, only 13 of the 776 adult patients required reoperation due to endocarditis [6]. Given the well-known negative impact and unpredictable nature of pericardial material in the aortic valve position in children [15], it is not surprising that the midterm outcomes of the Ozaki procedure have fallen somewhat short of expectations. Although some suggest that the Ozaki procedure could still serve as a palliative option [16], we discontinued its use in children after 2021 (Supplementary Material, Fig. S1).

The Ross procedure offers patients with excellent haemodynamics, with minimal peak gradient observed during the follow-up period, and it has shown to be highly reproducible. However, this procedure is not without its challenges. It is technically demanding with a long learning curve [17]. As demonstrated in our study, postoperative complications were more frequent in the Ross group compared to the Ozaki group, though this difference did not reach statistical significance. In addition, long-term aortic root dilation is a well-known issue [18, 19], and the eventual need for right ventricular outflow tract conduit replacement is almost inevitable [20]. Therefore, careful patient selection is crucial to achieving optimal outcomes [21]. Our study suggests that patients with predominant AS and mixed lesions fare the best following the Ross procedure. Some advocate for delaying the Ross procedure until adolescence or young adulthood, when the autograft could be externally reinforced, and the need for pulmonary conduit replacement might be reduced [22]. In our cohort, due to the limited availability of homografts in China, we utilized handmade ePTFE-valved conduits, which required early anticoagulation, although the duration of the therapy was short. The long-term outcomes of using this type of conduit in the Ross procedure are still to be determined [23].

The objective of our study was not to compare the Ross and Ozaki procedures directly, as the patient demographics were not comparable, with the Ozaki procedure being primarily performed in patients with predominant AR and a large annulus in our cohort. However, there are studies that have compared the 2 techniques in patients with more balanced preoperative characteristics. Polito et al. [24] from Rome reported their experience with 16 patients who underwent the Ross procedure and 22 who underwent the Ozaki procedure. The median age at operation was 12.4 years, with a median annulus diameter of 20 mm. Half of the patients presented with predominant AS and the other half with AR. During a median follow-up of 18.2 months, patients who underwent the Ozaki procedure had a significantly higher aortic peak gradient than those in the Ross group, although the vena contracta jet width was similar between the 2 groups. Freedom from reintervention or death also showed no significant difference. The same group conducted a 4D flow cardiovascular magnetic resonance study to assess the haemodynamic characteristics after the Ross and Ozaki procedures in children [25]. Their results indicated no significant differences in wall shear stress values at the level of the proximal ascending aorta and no clear association between eccentric flow and the type of procedure performed, although the Ozaki group did show a slight increase in peak velocity. Taken together, they concluded that the Ozaki procedure is a viable alternative to the Ross procedure in children. However, it should be noted that their study was short-term, and they acknowledged the need for long-term follow-up in patients undergoing the Ozaki procedure.

Limitations

This study has several limitations that warrant cautious interpretation of the findings. First, its retrospective design and small sample size inherently limit the robustness of our conclusions. The absence of adjustments for potential confounders further restricts direct comparisons between the Ross and Ozaki groups. While propensity score matching could reduce these differences, our limited sample size precluded meaningful application of such methods. The subgroup analyses should also be considered exploratory rather than definitive, as they are based on a small sample without confounder adjustment.

Second, selection bias may be present, as our cohort does not represent the full spectrum of paediatric aortic valve disease. Patients who underwent mechanical valve replacement following failed aortic valve repairs were excluded, which may have influenced the observed outcomes. Additionally, the Ross and Ozaki groups had distinct baseline characteristics, particularly in lesion type and annular size, limiting direct comparability between the 2 procedures. This study does not aim to establish superiority of 1 procedure over the other but rather to describe midterm outcomes based on current clinical decision-making practices.

Third, the introduction of the Ozaki and Ross procedures at our centre in 2017 and 2018, respectively, suggests that a learning curve may have influenced the outcomes. The impact of surgical experience on midterm durability should be considered, especially in comparison to results from high-volume centres with more extensive expertise [26].

Finally, as a single-centre study, our findings may not be fully generalizable to other institutions or broader patient populations. Larger, multicentre, prospective studies with longer follow-up are necessary to validate our findings and better define optimal surgical strategies for paediatric aortic valve disease.

CONCLUSIONS

The Ross and Ozaki procedures have expanded our armamentarium of treating aortic valve disease in children. Our exploratory findings suggest that the Ross procedure may provide a more durable neoaortic valve, particularly in patients with predominant AS or mixed lesions. In contrast, the Ozaki procedure, primarily performed in patients with predominant AR and large annuli in our cohort, may be associated with suboptimal midterm outcomes. However, due to the small sample size and retrospective nature of this study, these findings should be interpreted with caution. Larger, multicentre, prospective studies are needed to validate these results.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

FUNDING

This work was supported by the Natural Science Foundation of China [82370307, 82070322] and the Innovation Project of Distinguished Medical Team in Ningbo [2022020405].

Conflict of interest: none declared.

DATA AVAILABILITY

The data underlying this article will be shared on reasonable request to the corresponding author.

Author contributions

Wen Zhang: Conceptualization; Data curation; Formal analysis; Writing—original draft; Writing—review & editing. Qi Jiang: Conceptualization; Data curation; Formal analysis; Writing—original draft; Writing—review & editing. Yiman Liu: Data curation; Formal analysis; Writing—review & editing. Yifan Zhu: Formal analysis; Software. Renjie Hu: Formal analysis; Methodology; Software. Yuqi Zhang: Formal analysis; Methodology; Supervision; Writing—review & editing. Wei Dong: Conceptualization; Methodology; Supervision; Writing—review & editing. Haibo Zhang: Conceptualization; Methodology; Supervision; Writing—review & editing.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Victor T. Tsang, Rajesh Sharma and the other, anonymous reviewers for their contribution to the peer review process of this article.

Presented at the 38th EACTS Annual Meeting, Lisbon, Portugal, 11 October 2024.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• AS

  Aortic stenosis

 
• AR

  Aortic regurgitation

 
• CI

  Confidence interval

 
• ePTFE

  Expanded polytetrafluoroethylene

 
• IQR

  Interquartile range

 
• SD

  Standard deviation

Author notes